Electrical Engineering Dictionary - Shariflcsl/useful/Electrical Engineering...Electrical...

“Book/Definitions”Electrical Engineering Dictionary.Ed. Phillip A. LaplanteBoca Raton: CRC Press LLC, 2000

SpecialSymbols

α-level set a crisp set of elements belong-ing to a fuzzy setA at least to a degreeα

Aα = {x ∈ X | µA(x) ≥ α}

See alsocrisp set, fuzzy set.

1f common symbol for bandwidth, inhertz.

εrGaAs common symbol for gallium ar-senide relative dielectric constant.εrGaAs =12.8.

εrSi common symbol for silicon relativedielectric constant.εrSi = 11.8.

ε0 symbol for permitivity of free space.ε0 = 8.849× 10−12 farad/meter.

εr common symbol for relative dielectricconstant.

ηDC common symbol for DC to RF con-version efficiency. Expressed as a percent-age.

ηa common symbol for power added ef-ficiency. Expressed as a percentage.

ηt common symbol for total or true effi-ciency. Expressed as a percentage.

0opt common symbol for source reflec-tion coefficient for optimum noise perfor-mance.

µ0 common symbol for permeability offree space constant.µ0 = 1.257× 10−16

henrys/meter.

µr common symbol for relative perme-ability.

ω common symbol for radian frequencyin radians/second.ω = 2 · π · frequency.

θ+ common symbol for positive transitionangle in degrees.

θ− common symbol for negative transi-tion angle in degrees.

θcond common symbol for conduction an-gle in degrees.

θsat common symbol for saturation anglein degrees.

θCC common symbol for FET channel-to-case thermal resistance in◦C/watt.

θJC common symbol for bipolar junction-to-case thermal resistance in◦C/watt.

A∗ common symbol for Richardson’sconstant.A∗ = 8.7 amperes· cm/◦K

BVGD See gate-to-drain breakdownvoltage.

BVGS See gate-to-source breakdownvoltage.

dv/dt rate of change of voltage with-stand capability without spurious turn-on ofthe device.

Hci Seeintrinsic coercive force.

ne common symbol for excess noise inwatts.

nsh common symbol for shot noise inwatts.

c©2000 by CRC Press LLC

nt common symbol for thermal noise inwatts.

10base2 a type of coaxial cable used toconnect nodes on an Ethernet network. The10 refers to the transfer rate used on standardEthernet, 10 megabits per second. The basemeans that the network uses baseband com-munication rather than broadband communi-cations, and the 2 stands for the maximumlength of cable segment, 185 meters (almost200). This type of cable is also called “thin”Ethernet, because it is a smaller diameter ca-ble than the 10base5 cables.

10base5 a type of coaxial cable used toconnect nodes on an Ethernet network. The10 refers to the transfer rate used on stan-dard Ethernet, 10 megabits per second. Thebase means that the network uses basebandcommunication rather than broadband com-munications, and the 5 stands for the max-imum length of cable segment of approxi-mately 500 meters. This type of cable is alsocalled “thick” Ethernet, because it is a largerdiameter cable than the 10base2 cables.

10baseT a type of coaxial cable used toconnect nodes on an Ethernet network. The10 refers to the transfer rate used on standardEthernet, 10 megabits per second. The basemeans that the network uses baseband com-munication rather than broadband communi-cations, and the T stands for twisted (wire)cable.

2-D Attasi model a 2-D model describedby the equations

xi+1,j+1 = −A1A2xi,j + A1xi+1,j

+ A2xi,j+1 + Buij

yij = Cxij +Duij

i, j ∈ Z+ (the set of nonnegative integers).Here xij ∈ Rn is the local state vector,uij ∈ Rm is the input vector,yij ∈ Rp isthe output vector, andA1, A2, B,C,D arereal matrices. The model was introduced byAttasi in “Systemes lineaires homogenes a

deux indices,”IRIA Rapport Laboria,No.31, Sept. 1973.

2-D Fornasini–Marchesini model a 2-Dmodel described by the equations

xi+1,j+1 = A0xi,j + A1xi+1,j

+ A2xi,j+1 + Buij (1a)

yij = Cxij +Duij (1b)

i, j ∈ Z+ (the set of nonnegative integers)here xij ∈ Rn is the local state vector,uij ∈ Rm is the input vector,yij ∈ Rp isthe output vectorAk (k = 0,1,2), B,C,Dare real matrices. A 2-D model described bythe equations

xi+1,j+1 = A1xi+1,j + A2xi,j+1

+ B1ui+1,j + B2ui,j+1 (2)

i, j ∈ Z+ and (1b) is called the second 2-DFornasini–Marchesini model, wherexij , uij ,andyij are defined in the same way as for (1),Ak, Bk (k = 0,1,2) are real matrices. Themodel (1) is a particular case of (2).

2-D general model a 2-D model de-scribed by the equations

xi+1,j+1 = A0xi,j + A1xi+1,j

+ A2xi,j+1 + B0uij

+ B1ui+1,j + B2ui,j+1

yij = Cxij +Duij

i, j ∈ Z+ (the set of nonnegative integers)herexij ∈ Rn is the local state vector,uij ∈Rm is the input vector,yij ∈ Rp is the outputvector andAk,Bk (k = 0,1,2),C,D are realmatrices. In particular case forB1 = B2 = 0we obtain the first 2-D Fornasini–Marchesinimodel and forA0 = 0 andB0 = 0 we obtainthe second 2-D Fornasini–Marchesini model.

2-D polynomial matrix equation a 2-Dequation of the form

AX + BY = C (1)

whereA ∈ Rk×p [s], B ∈ Rk×q [s], C ∈Rk×m [s] are given, by a solution to (1) we


mean any pairX ∈ Rp×m [s], Y ∈ Rq×m [s]satisfying the equation. The equation (1)has a solution if and only if the matrices[A,B,C] and[A,B,0] are column equiva-lent or the greatest common left divisor ofAandB is a left divisor ofC. The 2-D equation

AX + YB = C (2)

A ∈ Rk×p [s], B ∈ Rq×m [s], C ∈ Rk×m [s]are given, is called the bilateral 2-D polyno-mial matrix equation. By a solution to (2) wemean any pairX ∈ Rp×m [s], Y ∈ Rk×q [s]satisfying the equation. The equation has asolution if and only if the matrices[

A 00 B

]and

[A C

0 B

]are equivalent.

2-D Roesser model a 2-D model de-scribed by the equations[xhi+1,jxvi,j+1

]=[A1 A2A3 A4

][xhijxvij

]+[B1B2

]uij

i, j ∈ Z+ (the set of nonnegative integers),

yij = C

[xhijxvij

]+Duij

Herexhij ∈ Rn1 andxvij ∈ Rn2 are the hori-zontal and vertical local state vectors, respec-tively, uij ∈ Rm is the input vector,yij ∈ Rpis the output vector andA1, A2, A3, A4, B1,B2, C, D are real matrices. The model wasintroduced by R.P. Roesser in “A discretestate-space model for linear image process-ing,” IEEE Trans. Autom. Contr.,AC-20,No. 1, 1975, pp. 1-10.

2-D shuffle algorithm an extension of theLuenberger shuffle algorithm for 1-D case.The 2-D shuffle algorithm can be used forchecking the regularity condition

det[Ez1z2 − A0 − A1z1 − A2z2] 6= 0

for some(z1, z2) ∈ C×C of the singular gen-eral model (Seesingular 2-D general model).

The algorithm is based on the row compres-sion of suitable matrices.

2-D Z-transform F(z1, z2) of a dis-crete 2-D functionfij satisfying the condi-tion fij = 0 for i < 0 or/andj < 0 isdefined by

F (z1, z2) =∞∑i=0

∞∑j=0

fij z−i1 z

−j2

An 2-D discretefij has the 2-D Z-transformif the sum

∞∑i=0

∞∑j=0

fij z−i1 z

−j2

exists.

2DEGFET Seehigh electron mobilitytransistor(HEMT).

2LG Seedouble phase ground fault.

3-dB bandwidth for a causal low-passor bandpass filter with a frequency functionH(jω) the frequency at which| H(jω) |dBis less than 3 dB down from the peak value| H(ωP ) |.3-level laser a laser in which the mostimportant transitions involve only three en-ergy states; usually refers to a laser in whichthe lower level of the laser transition is sepa-rated from the ground state by much less thanthe thermal energykT. Contrast with4-levellaser.

3-level system a quantum mechanicalsystem whose interaction with one or moreelectromagnetic fields can be described byconsidering primarily three energy levels.For example, the cascade, vee, and lambdasystems are 3-level systems.

4-level laser a laser in which the mostimportant transitions involve only four en-ergy states; usually refers to a laser in whichthe lower level of the laser transition is sep-arated from the ground state by much more


than the thermal energykT . Contrast with3-level laser.

45 Mbs DPCM for NTSC color videoa codec wherein a subjectively pleasing pic-ture is required at the receiver. This doesnot require transparent coding quality typicalof TV signals. The output bit-rate for videomatches the DS3 44.736 Megabits per secondrate. The coding is done by PCM coding theNTSC composite video signal at three timesthe color subcarrier frequency using 8 bit perpixel. Prediction of current pixel is obtainedby averaging the pixel three after current and681 pixels before next to maintain the sub-carrier phase. A leak factor is chosen beforecomputing prediction error to main the quali-

ty of the image. For example a leak factor of3132 the prediction decay is maintained at thecenter of the dynamic range.

X−L = 128+ 31

32

(X− − 128

).

Finally, a clipper at the coder and decoderis employed to prevent quantization errors.

90% withstand voltage a measure ofthe practical lightning or switching-surge im-pulse withstand capability of a piece of powerequipment. This voltage withstand level istwo standard deviations above the BIL of theequipment.


Aa posteriori probability Seeposteriorstatistics.

a priori probability Seeprior statistics.

A-mode display returned ultrasoundechoes displayed as amplitude versus depthinto the body.

A-site in a ferroelectric material with thechemical formula ABO3, the crystalline lo-cation of the A atom.

A/D Seeanalog-to-digital converter.

AAL SeeATM adaptation layer.

ABC Seeabsorbing boundary condition.

ABCD propagation of an optical raythrough a system can be described by a sim-ple 2×2 matrix. In ray optics, the character-istic of a system is given by the correspond-ing ray matrix relating the ray’s position fromthe axis and slope at the input to those at theoutput.

ABCD formalism analytic method usingtwo-by-two ABCD matrices for propagatingGaussian beams and light rays in a wide va-riety of optical systems.

ABCD law analytic formula for trans-forming a Gaussian beam parameter fromone reference plane to another in paraxial op-tics, sometimes called the Kogelnik transfor-mation. ABCD refers to the ABCD matrix.

ABCD matrix the matrix containingABCD parameters.SeeABCD parameters.

ABCD parameters a convenient mathe-matical form that can be used to characterize

two-port networks. Sometimes referred toas chain parameters. ABCD parameters arewidely used to model cascaded connectionsof two-port microwave networks, in whichcase the ABCD matrix is defined for eachtwo-port network. ABCD parameters canalso be used in analytic formalisms for prop-agating Gaussian beams and light rays. Raymatrices and beam matrices are similar butare often regarded as distinct.

ABC parameters have a particularly use-ful property in circuit analysis where thecomposite ABCD parameters of two cas-caded networks are the matrix products ofthe ABCD parameters of the two individualcircuits. ABCD parameters are defined as[

v1i1

]=[A B

C D

] [v2i2

]wherev1 andv2 are the voltages on ports oneand two, andi1 andi2 are the branch currentsinto ports one and two.

aberration an imperfection of an opticalsystem that leads to a blurred or a distortedimage.

abnormal event any external or program-generated event that makes further normalprogram execution impossible or undesir-able, resulting in a system interrupt. Exam-ples of abnormal events include system de-tection of power failure; attempt to divide by0; attempt to execute privileged instructionwithout privileged status; memory parity er-ror.

abort (1) in computer systems, to termi-nate the attempt to complete the transaction,usually because there is a deadlock or be-cause completing the transaction would re-sult in a system state that is not compati-ble with “correct” behavior, as defined by aconsistency model, such as sequential con-sistency.

(2) in an accelerator, terminating the ac-celeration process prematurely, either by in-hibiting the injection mechanism or by re-moving circulating beam to some sort of


dump. This is generally done to prevent in-jury to some personnel or damage to acceler-ator components.

ABR Seeavailable bit rate.

absolute address an address within aninstruction that directly indicates a location inthe program’s address space.Compare withrelative addressing.

absolute addressing an addressing modewhere the address of the instruction operandin memory is a part of the instruction so thatno calculation of an effective address by theCPU is necessary.

For example, in the Motorola M68000 ar-chitecture instruction ADD 5000,D1, a 16-bitword operand, stored in memory at the wordaddress 5000, is added to the lower word inregister D1. The address “5000” is an exam-ple of using the absolute addressing mode.See alsoaddressing mode.

absolute encoder an optical devicemounted to the shaft of a motor consistingof a disc with a pattern and light sources anddetectors. The combination of light detectorsreceiving light depends on the position of therotor and the pattern employed (typically theGray code). Thus, absolute position infor-mation is obtained. The higher the resolutionrequired, the larger the number of detectorsneeded.See alsoencoder.

absolute moment Thepth order absolutemomentµp of a random variableX is theexpectation of the absolute value ofX raisedto thepth power:

µp = E[|X|]p.

See alsocentral moment, central absolutemoment. See alsoexpectation.

absolute pressure units to measure gaspressure in a vacuum chamber with zero be-ing a perfect vacuum. Normally referred toas psia (pounds per square inch absolute).

absolute sensitivity denotedS(y, x), issimply the partial derivative ofy with respectto x, i.e., S(y, x) = ∂y/∂x, and is used toestablish the relationships between absolutechanges.Seesensitivity, sensitivity measure,relative sensitivity, semi-relative sensitivity.

absolute stability occurs when the net-work functionH(s) has only left half-planepoles.

absorber generic term used to describematerial used to absorb electromagnetic en-ergy. Generally made of polyurethanefoam and impregnated with carbon (and fire-retardant salts), it is most frequently used toline the walls, floors and ceilings of anechoicchambers to reduce or eliminate reflectionsfrom these surfaces.

absorbing boundary condition (ABC) afictitious boundary introduced in differentialequation methods to truncate the computa-tional space at a finite distance without, inprinciple, creating any reflections.

absorption (1) process that dissipates en-ergy and causes a decrease in the amplitudeand intensity of a propagating wave betweenan input and output reference plane.

(2) reduction in the number of photons of aspecific wavelength or energy incident upona material. Energy transferred to the materialmay result in a change in the electronic struc-ture, or in the relative movement of atoms inthe material (vibration or rotation).

(3) process by which atoms or moleculesstick to a surface. If a bond is formed, it istermed chemisorption, while the normal caseis physisorption. The absorption process pro-ceeds due to, and is supported by, the fact thatthis is a lower energy state.

absorption coefficient (1) in a passive de-vice, the negative ratio of the power absorbed(pabsorbed= pin−pout) ratioed to the power in(pin = pincident−preflected) per unit length (l),usually expressed in units of 1/wavelength or1/meter.


(2) factor describing the fractional atten-uation of light with distance traversed in amedium, generally expressed as an exponen-tial factor, such ask in the functione−kx ,with units of (length)-1. Also called attenu-ation coefficient.

absorption cross section energy ab-sorbed by the scattering medium, normal-ized to the wavenumber. It has dimensionsof area.

absorption edge the optical wavelengthor photon energy corresponding to the sep-aration of valence and conduction bands insolids; at shorter wavelengths, or higher pho-ton energies than the absorption edge, the ab-sorption increases strongly.

absorption grating (1) a diffractiongrating where alternate grating periods areopaque.

(2) an optical grating characterized byspatially periodic variation in the absorptionof light. Absorption gratings are generallyless efficient than phase gratings.

absorption optical fiber the amount ofoptical power in an optical fiber capturedby defect and impurity centers in the energybandgap of the fiber material and lost in theform of longwave infrared radiation.

AC Seealternating current.

AC bridge one of a wide group ofbridge circuits used for measurements of re-sistances, inductances, and capacitances, andto provide AC signal in the bridge transducersincluding resistors, inductors, and capacitors.

The Wheatstone bridge can be used witha sinusoidal power supply, and with an ACdetector (headphones, oscilloscope), one canuse essentially the same procedure for mea-surement of resistors as in DC applications.Only a small number of other AC bridges areused in modern electric and electronic equip-ment. A strong selection factor was the factthat in a standard capacitor the electrical pa-

rameter are closest to the parameters of anideal capacitor. Hence, not only a capaci-tance is measured in terms of capacitance (inresistive ratio arms bridges), but the induc-tance as well is measured in terms of capac-itance (Hay and Owen bridges).

The AC bridges with ratio arms that aretightly coupled inductances allow measure-ment of a very small difference between cur-rents in these inductances, and this fact isused in very sensitive capacitance transduc-ers.

AC circuit electrical network in which thevoltage polarity and directions of current flowchange continuously, and often periodically.Thus, such networks contain alternating cur-rents as opposed to direct currents, therebygiving rise to the term.

AC coupling a method of connecting twocircuits that allows displacement current toflow while preventing conductive currents.Reactive impedance devices (e.g., capacitorsand inductive transformers) are used to pro-vide continuity of alternating current flowbetween two circuits while simultaneouslyblocking the flow of direct current.

AC motor an electromechanical sys-tem that converts alternating current electri-cal power into mechanical power.

AC plasma display a display that em-ploys an internal capacitive dielectric layerto limit the gas discharge current.

AC steady-state power the averagepower delivered by a sinusoidal source to anetwork, expressed as

P =| V | · | I | cos(θ)

where√

2· | V | and√

2· | I | are the peakvalues, respectively, of the AC steady-statevoltage and current at the terminals.θ rep-resents the phase angle by which the voltageleads the current.


AC/AC converter a power electronicsdevice in which an AC input voltage of somemagnitude, frequency, and number of phasesis changed to an AC output with changes toany of the previously mentioned parameters.AC/AC converters usually rectify the inputsource to a DC voltage and then invert theDC voltage to the desired AC voltage.

AC/DC converter Seerectifier.

AC-DC integrated system a power sys-tem containing both AC and DC transmissionlines.

ACARS aircraft communications ad-dressing and reporting. A digital commu-nications link using the VHF spectrum fortwo-way transmission of data between an air-craft and ground. It is used primarily in civilaviation applications.

ACC Seeautomatic chroma control.

accelerated testing tests conducted athigher stress levels than normal operation butin a shorter period of time for the specificpurpose to induce failure faster.

accelerating power the excess electricpower at a synchronous machine unit whichcannot be transmitted to the load because ofa short circuit near its terminals. This energygives rise to increasing rotor angle.

acceleration error the final steady dif-ference between a parabolic setpoint and theprocess output in a unity feedback controlsystem. Thus it is the asymptotic error in po-sition that arises in a closed loop system thatis commanded to move with constant acceler-ation. See alsoposition error, velocity error.

acceleration error constant a gainKafrom which acceleration errorea is read-ily determined. The acceleration error con-stant is a concept that is useful in the designof unity feedback control systems, since ittransforms a constraint on the final acceler-

ation error to a constraint on the gain of theopen loop system. The relevant equationsareea = 1

KaandKa = lims→inf tys

2q(s),where q(s) is the transfer function modelof the open loop system, including the con-troller and the process in cascade, ands isthe Laplace variable.See alsoposition errorconstant, velocity error constant.

accelerator (1) a positive electrode in avacuum tube to accelerate emitted electronsfrom its cathode by coulomb force in a de-sired direction.

(2) a machine used to impart large kineticenergies to charged particles such as elec-trons, protons, and atomic nuclei. The ac-celerated particles are used to probe nuclearor subnuclear phenomena in industrial andmedical applications.

acceptable delay the voice signal de-lay that results in inconvenience in the voicecommunication. A typically quoted value is300 ms.

acceptance in an accelerator, it defineshow "large" a beam will fit without scrap-ing into the limiting aperture of a transportline. The acceptance is the phase-space vol-ume within which the beam must lie to betransmitted through an optical system with-out losses. From an experimenters pointof view acceptance is the phase-space vol-ume intercepted by an experimenter’s detec-tor system.

acceptor (1) an impurity in a semicon-ductor that donates a free hole to the valenceband.

(2) a dopant species that traps electrons,especially with regard to semiconductors.

access channel a channel in a communi-cations network that is typically allocated forthe purpose of setting up calls or communi-cation sessions. Typically the users share theaccess channel using some multiple accessalgorithm such as ALOHA or CSMA.


access control a means of allowing ac-cess to an object based on the type of ac-cess sought, the accessor’s privileges, and theowner’s policy.

access control list a list of items associ-ated with a file or other object; the list con-tains the identities of users that are permittedaccess to the associated file. There is infor-mation (usually in the form of a set of bits)about the types of access (such as read, write,or delete) permitted to the user.

access control matrix a tabular repre-sentation of the modes of access permittedfrom active entities (programs or processes)to passive entities (objects, files, or devices).A typical format associates a row with an ac-tive entity or subject and a column with anobject; the modes of access permitted fromthat active entity to the associated passive en-tity are listed in the table entry.

access line a communication line thatconnects a user’s terminal equipment to aswitching node.

access mechanism a circuit board or anintegrated chip that allows a given part of acomputer system to access another part. Thisis typically performed by using a specific ac-cess protocol.

access protocol a set of rules that estab-lishes communication among different parts.These can involve both hardware and soft-ware specifications.

access right permission to perform anoperation on an object, usually specified asthe type of operation that is permitted, suchas read, write, or delete. Access rights canbe included in access control lists, capabilitylists, or in an overall access control matrix.

access time the total time needed to re-trieve data from memory. For a disk drivethis is the sum of the time to position theread/write head over the desired track and the

time until the desired data rotates under thehead. (LW)

accidental rate the rate of false coinci-dences in the electronic counter experimentproduced by products of the reactions of morethan one beam particle within the time reso-lution of the apparatus.

accumulation (1) an increase in the ma-jority carrier concentration of a region ofsemiconductor due to an externally appliedelectric field.

accumulator (1) a register in the CPU(processor) that stores one of the operandsprior to the execution of an operation, andinto which the result of the operation isstored. An accumulator serves as an implicitsource and destination of many of the pro-cessor instructions. For example, register Aof the Intel 8085 is an accumulator.See alsoCPU .

(2) the storage ring in which successivepulses of particles are collected to create aparticle beam of reasonable intensity for col-liding beams.

achievable rate region for a multipleterminal communications system, a set ofrate-vectors for which there exist codes suchthat the probability of making a decoding er-ror can be made arbitrarily small.See alsocapacity region, multiple access channel.

achromatic the quality of a transport lineor optical system where particle momentumhas no effect on its trajectory through the sys-tem. In an achromatic device or system, theoutput beam displacement or divergence (orboth) is independent of the input beam’s mo-mentum. If a system of lenses is achromatic,all particles of the same momentum will haveequal path lengths through the system.

ACI Seeadjacent channel interference.


acknowledge (1) a signal which indicatesthat some operation, such as a data transfer,has successfully been completed.

(2) to detect the successful completion ofan operation and produce a signal indicatingthe success.

acoustic attenuation the degree of am-plitude suppression suffered by the acous-tic wave traveling along the acousto-opticmedium.

acoustic laser a laser (or maser) in whichthe amplified field consists of soundwaves orphonons rather than electromagnetic waves;phonon laser or phaser.

acoustic memory a form of circulatingmemory in which information is encoded inacoustic waves, typically propagated througha trough of mercury. Now obsolete.

acoustic velocity the velocity of theacoustic signal traveling along the acousto-optic medium.

acoustic wave a propagating periodicpressure wave with amplitude representingeither longitudinal or shear particle displace-ment within the wave medium; shear wavesare prohibited in gaseous and liquid media.

acousto-optic cell a device consisting ofa photo-elastic medium in which a propa-gating acoustic wave causes refractive-indexchanges, proportional to acoustic wave am-plitude, that act as a phase grating for diffrac-tion of light. See alsoBragg cell.

acousto-optic channelized radiometerSee acousto-optic instantaneous spectrumanalyzer in Bragg mode.

acousto-optic correlator an optical sys-tem that consists of at least one acousto-optic cell, imaging optics between cells andfixed masks, and photodetectors whose out-puts correspond to the correlation function ofthe acoustic wave signal within one cell with

another signal in a second cell, or with fixedsignals on a mask.

acousto-optic deflector device devicewhere acousto-optic interaction deflects theincident beam linearly as a function of theinput frequency of the RF signal driving thedevice.

acousto-optic device descriptor ofacousto-optic cells of any design; generallydescribes a cell plus its transducer struc-ture(s), and may encompass either bulk,guided-wave, or fiber-optic devices.

acousto-optic effect the interaction oflight with sound waves and in particular themodification of the properties of a light waveby its interactions with an electrically con-trollable sound wave. See alsoBrillouinscattering.

acousto-optic frequency excisor similarto an acousto-optic spectrum analyzer wherethe RF temporal spectrum is spatially and se-lectively blocked to filter the RF signal feed-ing the Bragg cell.

acousto-optic instantaneous spectrum an-alyzer in Bragg mode device in which thetemporal spectrum of a radio frequency sig-nal is instantaneously and spatially resolvedin the optical domain using a Fourier trans-form lens and a RF signal-fed Bragg cell.

acousto-optic modulator a device thatmodifies the amplitude or phase of a lightwave by means of the acousto-optic effect.

acousto-optic processor an optical sys-tem that incorporates acousto-optic cells con-figured to perform any of a number of math-ematical functions such as Fourier trans-form, ambiguity transforms, and other time-frequency transforms.

acousto-optic scanner a device that usesan acoustic wave in a photoelastic medium


to deflect light to different angular positionsbased on the frequency of the acoustic wave.

acousto-optic space integrating convolverdevice that is the same as an acousto-optic

space integrating convolver except that it im-plements the convolution operation.

acousto-optic space integrating correlatoran acousto-optic implementation of the cor-relation function where two RF signals arespatially impressed on two diffracted beamsfrom Bragg cells, and a Fourier transformlens spatially integrates these beams onto apoint sensor that generates a photo currentrepresenting the correlation function.

acousto-optic spectrum analyzer anacousto-optic processor that produces at aphotodetector output array the Fourier de-composition of the electrical drive signal ofan acousto-optic device.

acousto-optic time integrating convolversame as the acousto-optic time integrating

correlator, except implements the signal con-volution operation. Seeacousto-optic timeintegrating correlator.

acousto-optic time integrating correlatoran acousto-optic implementation of the cor-relation function where two RF signals arespatially impressed on two diffracted beamsfrom Bragg cells, and a time integrating sen-sor generates the spatially distributed corre-lation results.

acousto-optic triple product processorsignal processor that implements a triple inte-gration operation using generally both spaceand time dimensions.

acousto-optic tunable filter (AOTF) anacousto-optic device that selects specific op-tical frequencies from a broadband opticalbeam, depending on the number and frequen-cies of acoustic waves generated in the de-vice.

acousto-optics the area of study of in-teraction of light and sound in media, andits utilization in applications such as signalprocessing and filtering.

ACP Seeadjacent channel power.

acquisition (1) in digital communica-tions systems, the process of acquiring syn-chronism with the received signal. Thereare several levels of acquisitions, and for agiven communication system several of themhave to be performed in the process of settingup a communication link: frequency, phase,spreading code, symbol, frame, etc.

(2) in analog communications systems,the process of initially estimating signal pa-rameters (for example carrier frequency off-set, phase offset) required in order to begindemodulation of the received signal.

(3) in vision processing, the process bywhich a scene (physical phenomenon) isconverted into a suitable format that al-lows for its storage or retrieval.See alsosynchronization.

across the line starter a motor starter thatapplies full line voltage to the motor to start.This is also referred to as “hard starting” be-cause it causes high starting currents. Largermotors require reduced voltage or “soft start-ing.”

ACRR Seeadjacent channel reuse ratio.

ACSR aluminum cable, steel-reinforced.A kind of overhead electric power conduc-tor made up of a central stranded steel cableoverlaid with strands of aluminum.

ACT Seeanticomet tail.

action potential a propagating change inthe conductivity and potential across a nervecell’s membrane; a nerve impulse in commonparlance.

activation function in an artificial neuralnetwork, a function that maps the net output


of a neuron to a smaller set of values. Thisset is usually[0,1]. Typical functions are thesigmoid function or singularity functions likethe step or ramp.

active contour a deformable templatematching method that, by minimizing theenergy function associated with a specificmodel (i.e., a specific characterization of theshape of an object), deforms the model inconformation to salient image features.

active device a device that can convertenergy from a DC bias source to a signal atan RF frequency. Active devices are requiredin oscillators and amplifiers.

active filter (1) a filter that has an en-ergy gain greater than one, that is, a filter thatoutputs more energy than it absorbs.

(2) a form of power electronic converterdesigned to effectively cancel harmonic cur-rents by injecting currents that are equal andopposite to, or 180◦ out of phase with, the tar-get harmonics. Active filters allow the out-put current to be controlled and provide sta-ble operation against AC source impedancevariations without interfering with the systemimpedance.

The main type of active filter is the seriestype in which a voltage is added in series withan existing bus voltage. The other type is theparallel type in which a current is injectedinto the bus and cancels the line current har-monics.

active impedance the impedance at theinput of a single antenna element of an ar-ray with all the other elements of the arrayexcited.

active layer Seeactive region.

active learning a form of machine learn-ing where the learning system is able to in-teract with its environment so as to affect thegeneration of training data.

active load a transistor connected so as toreplace a function that would conventionallybe performed by a passive component suchas a resistor, capacitor, or inductor.

active load-pull measurement a mea-surement method where transfer characteris-tics of a device can be measured by electri-cally changing the load impedance seen fromthe device. In an active load-pull measure-ment, the load impedance is defined by usingan output signal from the device and an in-jected signal from the output of the device.

active logic a digital logic that operatesall of the time in the active, dissipative regionof the electronic amplifiers from which it isconstructed. The output of such a gate isdetermined primarily by the gate and not bythe load.

active magnetic bearing a magneticbearing that requires input energy for stablesupport during operation. Generally imple-mented with one or more electromagnets andcontrollers.

active mixer a mixer that uses three termi-nal devices such as FET rather than diodes asnonlinear element. One advantage of activemixers is that they can provide conversiongain.

active network an electrical networkthat contains some solid state devices such asbipolar junction transistors (BJTs) or metal-oxide-silicon field effect transistors (FETs)operating in their active region of the volt-age vs. current characteristic. To ensure thatthese devices are operating in the active re-gion, they must be supplied with proper DCbiasing.

active neuron a neuron with a non-zerooutput. Most neurons have an activationthreshold. The output of such a neuron haszero output until this threshold is reached.

active power Seereal power.


active power line conditioner a devicewhich senses disturbances on a power lineand injects compensating voltages or currentsto restore the line’s proper waveform.

active RC filter an electronic circuitmade up of resistors, capacitors, and opera-tional amplifiers that provide well-controlledlinear frequency-dependent functions, e.g.,low-, high-, and bandpass filters.

active redundancy a circuit redundancytechnique that assures fault-tolerance by de-tecting the existence of faults and performingsome action to remove the faulty hardware,e.g., by standby sparing.

active region semiconductor materialdoped such that electrons and/or holes arefree to move when the material is biased. Inthe final fabricated device, the active regionsare usually confined to very small portions ofthe wafer material.

active-high (1) a logic signal having itsasserted state as the logic ONE state.

(2) a logic signal having the logic ONEstate as the higher voltage of the two states.

active-low (1) a logic signal having itsasserted state as the logic ZERO state.

(2) a logic signal having its logic ONEstate as the lower voltage of the two states;inverted logic.

actuator (1) a transducer that convertselectrical, hydraulic, or pneumatic energy toeffective motion. For example in robots, ac-tuators set the manipulator in motion throughactuation of the joints. Industrial robotsare equipped with motors that are typicallyelectric, hydraulic, or pneumatic.See alsoindustrial robot.

(2) in computers, a device, usually me-chanical in nature, that is controlled by acomputer, e.g., a printer paper mechanism ora disk drive head positioning mechanism.

ACTV Seeadvanced compatible tele-vision.

acuity sharpness. The ability of the eyeto discern between two small objects closelyspaced, as on a display.

adaptability the capability of a system tochange to suit the prevailing conditions, espe-cially by automatic adjustment of parametersthrough some initialization procedure or bytraining.

adaptation layer control layer of a mul-tilayer controller, situated above the directcontrol layer and — usually — also above theoptimizing control layer, required to intro-duce changes into the decision mechanismsof the layer (or layers) below this adaptationlayer; for example adaptation layer of the in-dustrial controller may be responsible for ad-justing the model used by the optimizing con-trol and the decision rules used by the direct(regulation) control mechanisms.

adapter a typical term from personalcomputers. A circuit board containing theinterface toward an additional peripheral de-vice. For example, a graphic adapter (inter-face boards like EGA, VGA, CGA), a gamecontroller, a SCSI controller, a PCMCI inter-face, etc.

adaptive algorithm (1) a method for ad-justing the parameters of a filter to satisfy anobjective (e.g., minimize a cost function).

(2) an algorithm whose properties are ad-justed continuously during execution withthe objective of optimizing some criterion.

adaptive antenna antenna, or array ofantennas, whose performance characteristicscan be adapted by some means; e.g., thepattern of an array can be changed whenthe phasing of each of the array elements ischanged.

adaptive array an array that adapts itselfto maximize the reception of a desired sig-


nal and null all interfering or jamming sig-nals. This is achieved by finding the correctweights (input excitations) to the elementscomprising the array.

adaptive coding a coding scheme thatadapts itself in some fashion to its input oroutput.

adaptive coding of transform coefficientscoding technique that is carried out by

threshold sampling and exploiting maskingeffects by variable quantization for differ-ent blocks. High detail blocks are codedwith more quantization error than low de-tail blocks. This is done to take into ac-count masking and boundary distortion ef-fects. Transform coding becomes more at-tractive compared with DPCM when adap-tive coding is used. The main drawback ofadaptive transform coding is its sensitivityto transmission bit errors due to synchro-nization problems at the decoder.See alsoDPCM.

adaptive control a control methodologyin which control parameters are continuouslyand automatically adjusted in response tobe measured/estimated process variables toachieve near-optimum system performance.

adaptive critic learning technique wherethe system learns to evaluate the actions of asystem (usually a controller) so as to providea reinforcement signal that is an estimate ofthe future value of the system’s current ac-tion.

adaptive differential pulse code modula-tion (ADPCM) a modulation scheme inwhich only the difference between successivesignal samples is encoded for transmission,and the quantization of the coding is adaptedto the characteristics of the signal source.

adaptive filtering a filtering strategy inwhich filter coefficients or governing param-eters evolve over time according to some up-dating strategy to optimize some criterion.

adaptive FIR filter a finite impulse re-sponse structure filter with adjustable coef-ficients. The adjustment is controlled by anadaptation algorithm such as the least meansquare (LMS) algorithm. They are usedextensively in adaptive echo cancellers andequalizers in communication systems.

adaptive fuzzy system fuzzy inferencesystem that can be trained on a data setthrough the same learning techniques usedfor neural networks. Adaptive fuzzy systemsare able to incorporate domain knowledgeabout the target system given from humanexperts in the form of fuzzy rules and numer-ical data in the form of input–output data setsof the system to be modeled.See alsoneuralnetwork, fuzzy inference system.

adaptive intrafield predictors a tech-nique used for picture signal prediction basedon local properties of the signal or side infor-mation if portions of local properties havenot been transmitted. Intrafield methods re-quire correlation with local information forprediction purposes.

A common technique is to use a mea-sure of the directional correlation based onlocal pixels that have already been transmit-ted. A predictor is chosen from a set to giveminimum prediction error. For example, theprevious line or previous pixel can be usedfor prediction, and the switching can then bedone as follows:

∩X = predictor for elementX

={A if ‖B − C‖ < ‖A− B‖C otherwise

An extension of this concept is called con-tour prediction where the direction of pixelAis determined by searching amongE, B, C,orG.

adaptive logic network tree-structurednetwork whose leaves are the inputs andwhose root is the output. The first hiddenlayer consists of linear threshold units and the


remaining layers are elementary logic gates,usually AND and OR gates. Each linearthreshold unit is trained to fit input data inthose regions of the input space where it isactive (i.e., where it contributes to the overallnetwork function).

adaptive manipulator controller a con-troller that uses an adaptation process which,based on observation of the manipulator po-sition and velocity, readjusts the parametersin the nonlinear model until the errors dis-appear. An adaptive manipulator controlleris depicted in the figure below. Such a sys-tem would learn its own dynamic properties.The adaptive manipulator control scheme

Adaptive manipulator control scheme.

presented in the figure belongs to the jointspace control schemes.See alsojoint spacecontrol.

adaptive predictor a digital filter whosecoefficients can be varied, according to someerror minimization algorithm, such that it canpredict the value of a signal sayN samplingtime intervals into the future. The adaptivepredictor is useful in many interference can-cellation applications.

adaptive resonance theory (ART) networkA clustering network developed to allow thelearning of new information without destroy-ing what has already been learnt. Each clus-ter is represented by a prototype and learningis achieved by comparing a new input pat-tern with each prototype. If a prototype isfound that is acceptably close to that input,

the new pattern is added to that prototype’scluster and the prototype is adjusted so asto move closer to the new input. If no pro-totype is acceptable, the pattern becomes anew prototype around which a new clustermay develop.

adaptive vector quantization term thatrefers to methods for vector quantization thatare designed to adaptively track changes inthe input signal.

ADC Seeanalog-to-digital converter.

ADCPM Seeadaptive differential pulsecode modulation.

add instruction a machine instructionthat causes two numeric operands to be addedtogether. The operands may be from machineregisters, memory, or from the instruction it-self, and the result may be placed in a ma-chine register or in memory.

adder a logic circuit used for adding bi-nary numbers.

additive acousto-optic processingacousto-optic signal processing where thesummation of acousto-optic modulated lightwaves is used to implement the signal pro-cessing operation.

additive polarity polarity designation ofa transformer in which terminals of the samepolarity on the low- and high-voltage coilsare physically adjacent to each other on thetransformer casing. With additive polarity, ashort between two adjacent terminals resultsin the sum of the two coil voltages appearingbetween the remaining terminals. Additivepolarity is generally used for transformers upto 500kVA and 34.5kV. Larger units use sub-tractive polarity. See the diagram below.Seealsosubtractive polarity.

additive white Gaussian noise (AWGN)the simplest form of channel degradation ina communication system in which the source


Transformer with additive polarity.

of errors in the channel can be modeled asthe addition of random noise with a Gaus-sian distribution and a constant (white) powerspectrum.See alsothermal noise.

address a unique identifier for the placewhere information is stored (as opposed tothe contents actually stored there). Most stor-age devices may be regarded by the user as alinear array, such as bytes or words in RAMor sectors on a disk. The address is then justan ordinal number of the physical or logicalposition. In some disks, the address may becompound, consisting of the cylinder or trackand the sector within that cylinder.

In more complex systems, the addressmay be a “name” that is more relevant to theuser but must be translated by the underlyingsoftware or hardware.

address aliasing Seecache aliasing.

address bus the set of wires or trackson a backplane, printed circuit board, or in-tegrated circuit to carry binary address sig-nals between different parts of a computer.The number of bits of address bus (the widthof the bus) determines the maximum size ofmemory that can be addressed. Modern mi-crochips have 32 address lines, thus 4 giga-bytes of main memory can be accessed.

address decoder logic that decodes anaddress.

1. A partial decoder responds to a smallrange of addresses and is used when recog-nizing particular device addresses on an I/Oaddress bus, or when recognizing that ad-dresses belong to a particular memory mod-ule.

2. A full decoder takesN bits and assertsone of 2N outputs, and is used within mem-ories (often within RAM chips themselves).

address error an exception (error inter-rupt) caused by a program’s attempt to accessunaligned words or long words on a proces-sor that does not accommodate such requests.The address error is detected within the CPU.This contrasts with problems that arise in ac-cessing the memory itself, where a logic cir-cuit external to the CPU itself must detect andsignal the error to cause the CPU to processthe exception. Such external problems arecalled bus errors.

address field the portion of a programinstruction word that holds an address.

address generation interlock (AGI) amechanism to stall the pipeline for one cyclewhen an address used in one machine cycleis being calculated or loaded in the previouscycle. Address generation interlocks causethe CPU to be delayed for a cycle. (AGIson the Pentium are even more important toremove, since two execution time slots arelost).

address locking a mechanism to protecta specific memory address so that it can beaccessed exclusively by a single processor.

address map a table that associates a baseaddress in main memory with an object (orpage) number.

address mapping the translation of vir-tual address into real (i.e., physical) ad-dresses for memory access.See alsovirtualmemory.

address register a register used primarilyto hold the address of a location in memory.The location can contain an operand or anexecutable instruction.

address size prefix a part of a machineinstruction that provides information as to the


length or size of the address fields in the in-struction.

address space an area of memory seen orused by a program and generally managed asa continuous range of addresses. Many com-puters use separate address spaces for codeand data; some have other address spacesfor system. An address space is usually sub-ject to protection, with references to a spacechecked for valid addresses and access (suchas read only).

The physical address space of a computer(232 bytes, and up to 264 bytes) is often largerthan the installed memory. Some parts of theaddress range (often at extreme addresses)may be reserved for input–output device ad-dresses. See alsobyte, memory, memorymapped I/O.

address translation Seeaddress mapping.

addressing (1) in processors: a mecha-nism to refer to a device or storage location byan identifying number, character, or group ofcharacters. That may contain a piece of dataor a program step.

(2) in networks, the process of identify-ing a network component, for instance, theunique address of a node on a local area net-work.

addressing fault an error that halts themapper when it cannot locate a referencedobject in main memory.

addressing mode a form of specifyingthe address (location) of an operand in aninstruction. Some of the addressing modesfound in most processors are direct or registerdirect, where the operand is in a CPU register;register indirect (or simply indirect), wherea CPU register contains the address of theoperand in memory; immediate, where theoperand is a part of the instruction.See alsocentral processing unit.

addressing range numbers that definethe number of memory locations addressableby the CPU. For a processor with one addressspace, the range is determined by the numberof signal lines on the address bus of the CPU.

adequate service in terms of the block-ing probability, term associated with a fixedblocking. A typically quoted value may be2. See alsoblocking.

adiabatic a system that has no heat trans-fer with the environment.

adiabatic cooling a process where thetemperature of a system is reduced withoutany heat being exchanged between the sys-tem and its surroundings. In particle beamacceleration this term is used to describe theprocess in the particle source storage ringwhere beam emittances are reduced withoutaffecting beam energy.

adiabatic following an approximationmade when some states in a quantum me-chanical system respond to perturbationsmore quickly than the other states. In thisapproximation the rapidly responding statesare assumed to depend only on the instanta-neous values of the other states and are saidto “follow” those states.

adiabatic passage a technique for the cre-ation of a long-lived coherence in a quantummechanical system by manipulating electro-magnetic field intensities so that the systemalways remains in an eigenstate. In practice,this involves changing field strengths on atime scale slower than the inverse of the en-ergy spacing between relevant eigenstates ofthe system. For example, consider a lambdasystem in which only one field is present ini-tially and all population starts out in the un-coupled ground state. If a field is graduallyturned on to couple this initial state to the ex-cited state, the system can remain transparentby evolving in such a way that it is alwaysmathematically equivalent to the dark statethat would be produced by coherent popu-


lation trapping. Adiabatic passage is oftenused for selective transfer of population be-tween two long-lived states of a multistatesystem, especially in cases where the two-step process of absorption followed by spon-taneous decay (optical pumping) would tendto populate many other states.

adjacency graph a graph in which eachnode represents an object, component, or fea-ture in an image. An edge between two nodesindicates two components that are touchingor connected in the image.

adjacent channel interference (ACI) theinterference caused by an adjacent frequencyband, e.g., in a system with frequency divi-sion duplex (FDD). Classified as either in-band or out-of-band adjacent channel inter-ference (ACI). The in-band ACI occurs whenthe center frequency of interfering signal fallswithin the band of the desired signal. Theout-of-band ACI occurs when the center fre-quency of interfering signal falls outside thebandwidth of the desired signal.

adjacent channel leakage power Seeadjacent channel power.

adjacent channel power (ACP) a powerof distortion components generated in adja-cent channel, which is caused by a nonlinear-ity of high-power amplifier amplifying a dig-itally modulated signal such as QPSK, QAM,etc. Adjacent channel power is defined as aratio of signal power in channel and leakagepower in adjacent channel.

adjacent channel reuse ratio (ACRR)the reuse ratio between radio communicationcells using adjacent radio channels.See alsoreuse ratio.

adjacent channels radio channels occu-pying radio frequency allocationsnandn±1.

adjoint network a network with an iden-tical structure to the original one, but withpossibly different elements. As an exam-

ple, for a network described by the nodal ad-mittance matrix, its adjoint network is repre-sented by the transposed admittance matrixof the original network. The adjoint networkis a basic tool in the computer-aided sensi-tivity analysis of electronic and microwavecircuits.

adjustable-speed drive See variablespeed DC drive, variable speed AC drive.

admissible matrix a matrixM− that canbe obtained by fixing the free parameters ofthe matrixM at some particular values.M−is said to be admissible with respect toM.

admittance the reciprocal of theimpedance of an electric circuit.

admittance inverter an idealized de-vice or set of matrix parameters that func-tions electrically like a quarter-wave losslesstransmission line of characteristic impedanceJ at each frequency, thus transforming theload admittance (YLOAD) by +90 degrees andmodifying the magnitude, resulting in an in-put admittance (Yin).

Yin = J 2

Yload

admittance matrix the inverse of theimpedance matrix in the method of moments.

ADP Seeammonium dihydrogen phosphate.

ADPCM Seeadaptive differential pulsecode modulation.

ADSL Seeasymmetric digital subscriberline.

adsorbent the material of an adsorber,for example, silica gel, alumina, and char-coal. Adsorbent materials are characterizedby high surface to volume ratio.

adsorber (1) condensation of a gas on thesolid material.


(2) material that attracts and holds (by Vander Waal forces) molecular layers of densegases (i.e., very near condensation temper-atures) on porous high surface/volume ratiomaterials.

ADTV Seeadvanced digital television.

advanced compatible television (ACTV)an extended definition television system thatcan operate with existing bandwidths on ex-isting receivers and is compatible with theNTSC broadcasting system. The ACTV sys-tem was proposed by the Advanced Televi-sion Research Consortium and was the firsthigh definition television (HDTV) system.HDTV system was tested by the FCC July17, 1992. The additional picture informationneeded to increase the picture width and toincrease the resolution to the HDTV formatis transmitted in an augmented channel asan alternative to simulcast transmission.SeeAdvanced Television Research Consortium.

advanced digital television (ADTV)a high definition television (HDTV) digitaltransmission television system was proposedto the Federal Communications Commissionby the Advanced Television Research Con-sortium. The ADTV system introduced alayered system to separately describe the dig-ital transmission system, the video compres-sion system, and the data packet transportsystem. The video compression method usesa MPEG++ standard that provides for com-patibility with multimedia computing.SeeAdvanced Television Research Consortium.

advanced mobile phone system (AMPS)a standard for a cellular radio communi-

cations network originally developed in the1970s by AT&T and later adopted as an in-dustry standard by the U.S.-based Telecom-munications Industries Association (TIA). Itis the first cellular standard widely deployedin North America. It is also referred to as theanalog cellular system. Frequency modula-tion with 30 kHz channels is used.

Advanced Television Research Consor-tium an organization consisting of DavidSarnoff Research Center, Thompson Con-sumer Electronics, North American PhilipsCorporation, NBC, and Compression Labo-ratories.

aeolian vibration a high-frequency me-chanical vibration of electric power linescaused by wind.

aerial cable any fully-insulated electricpower cable which is carried overhead uponpoles, as opposed to the use of the more usualoverhead bare conductors.

aerodynamic head Seedisk head.

AFC Seeautomatic frequency control.

affine transform a geometric imagetransformation including one or more transla-tions, rotations, scales, and shears that is rep-resented by a 4× 4 matrix allowing multiplegeometric transformations in one transformstep. Affine transformations are purely lin-ear and do not include perspective or warpingtransformations.

AFM Seeatomic force microscope.

AFT Seeautomatic fine tuning.

AFV Seeaudio follow-video switcher.

AGC See automatic gain controlorautomatic generation control.

agent a computational entity that actson behalf of other entities in an autonomousfashion.

agent-based system an applicationwhose component are agents.See alsoagent.

aggregation an operation performed onsystem variables whose purpose is to collectthem in a way enabling order and/or uncer-tainty reduction. For linear systems both


continuous-time and discrete-time state ag-gregation is obtained by linear transforma-tion of the original state represented by anaggregation matrixG endowed with the fol-lowing properties:

GA = A∗G;GB = B∗;CG′ = C∗;

whereA,B,C are original system matrices(respectively state, input, and output ones)andA∗, B∗, C∗ are aggregated system ma-trices. The aggregation is an eigenvalues-preservation approach and it provides orderreduction by neglecting some of the systemmodes.

For uncertainties, the aggregation definessome deterministic measures for a set of un-certain variables. For stochastic model ofuncertainty the aggregation may be given bymean value, higher stochastic models or otherstatistical characteristics, while set member-ship uncertainties could be aggregated bytheir maximal or minimal values, mass centerof the set or higher inertial moments.

AGI Seeaddress generation interlock.

Aiken, Howard Hathaway (1900–1973)Born: Hoboken, New Jersey, U.S.A.

Aiken is best known as the inventor ofthe Mark I and Mark II computers. Whilenot commercially successful, these machineswere significant in the development of themodern computer. The Mark I was essen-tially a mechanical computer. The Mark IIwas an electronic computer. Unlike UNI-VAC ( SeeEckert, John Presper) these ma-chines had a stored memory. Aiken was aprofessor of mathematics at Harvard. He wasgiven the assignment to develop these com-puters by the Navy department. Among hiscolleagues in this project were three IBM sci-entists and Grace Hopper. It was while work-ing on the Mark I that Grace Hopper pulledthe first “bug” from a computer.

air bridge a bridge made of metal stripsuspended in air that can connect componentson an integrated circuit in such a way as to

cross over another strip. Air bridges are alsoused to suspend metalization in spiral induc-tors off of the semi-conducting substrate in away that can lead to improved performancein some cases.

air capacitor a fixed or variable capacitorin which air is the dielectric material betweenthe capacitor’s plates.

air circuit breaker a power circuitbreaker where the power contacts operate inair. Some versions employ an air blast toextend and clear the arc on contact opening,while others employ arc chutes with mag-netic or thermal assists.

air core transformer two or more coilsplaced so that they are linked by the same fluxwith an air core. With an air core the flux isnot confined.

air gap Seemagnetic recording air gap.

air ionization chamber a device used tomonitor neutron flux.

air line a coaxial transmission line inwhich the volume between the inner andouter conductors are air-filled.

air terminal a lightning rod; any devicewhich extends upward into the air from astructure for purposes of lightning protection.

air-blast circuit breaker a circuit breakerin which the arc which forms between thecontacts on opening is extinguished with ablast of high-pressure air.

air-gap line the line that is obtained bycontinuing the linear portion of the saturationcurve of a synchronous machine or a DC ma-chine. The figure shows a plot of generatedvoltage vs. field current at constant machinespeed. Initially, an increase in field currentyields a linear increase in the generated volt-age, but as the iron becomes saturated, thevoltage rolls off. The air-gap line gives the


Plot of generated voltage vs. field current at con-

stant machine speed.

voltage that would be obtained without satu-ration.

air-gap voltage the internal voltage of asynchronous machine that is generated by theair gap flux. Also referred to as the voltagebehind leakage reactance.

airline a precision coaxial transmissionline with air dielectric used in a variety ofcalibration techniques and measurements asan impedance standard and to establish a ref-erence plane.

airy disk the central portion of the far-field optical diffraction pattern.

AlAs aluminum arsenide.

albedo the ratio between the total scat-tered intensity and the whole extracted fromthe incident light by scattering and absorp-tion.

ALC Seeautomatic level control.

AlGaAs symbol for aluminum galliumarsenide.

algebraic reconstruction the process ofreconstructing an imagex from a noise-corrupted and blurred imagey. An arbitraryimage is selected as the initial condition of

an iterative algorithm for solving a set of lin-ear equations. A set of linear constraints isspecified. In each iteration one constraint isapplied to a linear equation. The constraintsare repeated in a cyclic fashion until conver-gence is reached. The linear constraints arevectors in a vector space with specified basisimages for the type of problem to be solved.

algorithm (1) a systematic and precise,step-by-step procedure (such as a recipe, aprogram, or set of programs) for solving acertain kind of problem or accomplishing atask, for instance converting a particular kindof input data to a particular kind of outputdata, or controlling a machine tool. An algo-rithm can be executed by a machine.

(2) in image processing, algorithms can beeither sequential, parallel, or ordered. In se-quential algorithms, pixels are scanned andprocessed in a particular raster-scan order.As a given pixel is processed, all previouslyscanned pixels have updated (processed) val-ues, while all pixels not yet scanned have old(unprocessed) values. The algorithm’s resultwill in general depend on the order of scan-ning.

In a parallel algorithm, each pixel is pro-cessed independently of any changes in theothers, and its new value is written in a newimage, such that the algorithm’s result doesnot depend on the order of pixel processing.

In an ordered algorithm, pixels are put inan ordered queue, where priority depends onsome value attached to each pixel. At eachtime step, the first pixel in the queue is takenout of it and processed, leading to a possi-ble modification of priority of pixels in thequeue. By default, an algorithm is usuallyconsidered as parallel, unless stated other-wise.

algorithmic state machine (ASM) a se-quential logic circuit whose design is directlyspecified by the algorithm for the task the ma-chine is to accomplish.


aliasing (1) in signal processing, distor-tion introduced in a digital signal when it isundersampled.

In all digital systems the signals should befiltered before they are sampled to eliminatesignal components with frequencies abovethe Nyquist frequency,

ωN = ωs/2 = π/T ,

whereT is a sampling time, are eliminated.If this filtering is not done, signal componentswith frequencies

ω > ωN

will appear as low-frequency componentswith the frequency

ωa = |((ω + ωN) mod ωs)− ωN |

The prefilters introduced before a sampler arecalled anti-aliasing filters (common choicesare second- or fourth-order Butterworth, in-tegral time absolute error (ITAE), or Besselfilters).

(2) in computer graphics, distortion dueto the discrete nature of digital images thatcauses straight lines to appear jagged.

(3) in computer software, a single objecthaving two different identities, such as namesin memory space. Aliasing can make it diffi-cult to determine whether two names (or ac-cess paths to reach an object) that appear to bedifferent really access the identical object; asystem designed to find parallelism when twoaccesses really reach different objects willhave trouble achieving correct (functional)operation if aliasing is present.

alignment (1) the requirement that a da-tum (or block of data) be mapped at an ad-dress with certain characteristics, usually thatthe address modulo the size of the datum orblock be zero. For example, the address of anaturally aligned long word is a multiple offour.

(2) the act of positioning the image of aspecific point on a photomask to a specificpoint on the wafer to be printed.

(3) the process of determining the time orphase shift of a certain signal so that part ofit may be matched with another signal.Seealso image registration.

all-digital synchronization synchroniza-tion algorithm, where the analog-to-digitalconversion takes place as early as possible toassist digital implementation of the synchro-nizer. In most cases, an all-digital synchro-nization approach leads to optimal maximumlikelihood algorithms.

all-optical network an optical communi-cations network where the role of electronicsis reduced to basic supervisory and controlfunctions. All-optical devices are used ex-clusively between the nodes to re-configurethe network which enables the greatest use offiber bandwidth.

all-optical switch an optically addresseddevice whose optical transmission can beswitched between two possible states bychanges in the incident optical power.

all-pass system a system with unit mag-nitude and poles and zeroes that are complexconjugate reciprocals of each other. An all-pass system with a pole atz = a and a zeroat z = 1

a∗ is

Hap(z) = z−1 − a∗

1 − az−1.

alley arm a crossarm meant for use inan alleyway or other confined area in whichpoles must be placed close to buildings.Seecrossarm.

allocate to create a block of storage of agiven size in some memory, which is not tobe used for any other purpose until expresslyfreed.

allocation the act of allocating.See alsoallocate.


allocation of authority process by whichthe authority (scope of competence) is allo-cated to various decision units; this allocationmay result form the natural reasons or be aproduct of system partitioning.

almost sure convergence for a stochas-tic process, the property of the sample valuesconverging to a random variable with proba-bility one (for almost all sample paths).

alnico a permanent magnet material con-sisting mainly of aluminum, nickel, cobalt,and iron, which has a relatively low-energyproduct and high residual flux density. Analnico is most suitable for high-temperatureapplications.

ALOHA a random access, multiple ac-cess protocol, originally developed by Nor-man Abramson at the University of Hawaii in1970. A given user transmits a message whenthe message is generated without regard forcoordination with the other users sharing thechannel. Messages involved in collisions areretransmitted according to some retransmis-sion algorithm. Literally, “aloha” is a greet-ing in the Hawaiian native language.

alpha channel a grayscale image associ-ated with the color channels of an image thatdictates the opacity/transparency of the cor-responding color channel pixels. If the colorchannels are multiplied by the alpha chan-nel when stored, the image is referred to aspremultiplied; otherwise, it is known as un-premultiplied.

alpha particle a subatomic particleemitted by ceramic packaging materials thatcauses soft errors in memory integrated cir-cuits.

alpha particle noise this type of noiseoccurs exclusively in small semiconductorcapacitors, when an energetic alpha particle,either from cosmic rays or from the packag-ing or substrate itself, traverses the capaci-tor, discharging it, thereby creating an error

in the stored charge. Such an accumulationof errors in a digital system has the effect ofcreating a noise signal.

alpha-cut the set of all crisp, or nonfuzzy,elements whose membership function inA isgreater than or equal to a given value,α.

alphanumeric mode relates to alpha-betic characters, digits, and other characterssuch as punctuation marks. Alphanumericis a mode of operation of a graphic terminalor other input/output device. The graphicsterminal should toggle between graphic andalphanumeric data.

alternate channel power a measure ofthe linearity of a digitally modulated system.The amount of energy from a digitally trans-mitted RF signal that is transferred from theintended channel to one which is two chan-nels away. It is the ratio (in decibels) of thepower measured in the alternate channel tothe total transmitted power.

alternating current (AC) a periodic cur-rent the average value of which over a periodis zero.

alternating current machine an electro-mechanical system that either converts alter-nating current electrical power into mechan-ical power (AC motor), or converts mechan-ical power into alternating current electricalpower (AC generator, or alternator). SomeAC machines are designed to perform eitherof these functions, depending on the energysource to the dynamo.

alternator-rectifier exciter a source offield current of a synchronous machine de-rived from the rectified output voltage of analternator. The components of the exciterconsist of the alternator and the power rec-tifier (including possible gate circuitry), ex-clusive of all input control elements. Therectifier circuits may be stationary, or rotatewith the alternator, which may be driven by


a motor, prime mover, or by the shaft of thesynchronous machine.

ALU Seearithmetic and logic unit.

AM Seeamplitude modulation.

AM to PM conversion phase variationsof an output signal, due to passing through anactive device, where the phase of the outputsignal varies in response with the amplitudeof the input signal.

AM video the amplitude modulated videocarrier wave is produced by an amplitudemodulated video transmitter where the am-plitude of the wave form varies in step withthe video signal similar to that shown in thefigure.

amateur radio The practice and study ofelectronic communications as an avocation;most often referring to those persons possess-ing a license earned by examination (in theU.S., the Federal Communications Commis-sion grants such licenses).

ambient field the background magneticfield level existing in the environment, with-out contribution from specific magnetic fieldsources.

ambient temperature the temperatureof the air or liquid surrounding any electricalpart or device. Usually refers to the effectof such temperature in aiding or retarding re-moval of heat by radiation and convectionfrom the part or device in question.

ambiguity in artificial intelligence, thepresence of more than one meaning or possi-bility.

Amdahl’s law states that the speedupfactor of a multiprocessor system is given by

S(n) = n

1 + (n− 1)f

where there arenprocessors andf is the frac-tion of computational that must be performedsequentially (by one processor alone). Theremaining part of the computation is assumedto be divided inton equal parts each executedby a separate processor but simultaneously.The speedup factor tends to 1/f asn → ∞,which demonstrates that under the assump-tions given, the maximum speedup is con-strained by the serial fraction.

American National Standards Institute(ANSI) The U.S. organization that rec-ommends standards for metrology, drawingsymbology and numerous other facets forproducts and industries.


American standard code for informationinterchange (ASCII) a binary code com-prised of seven digits, originally used totransmit telegraph signal information.

ammeter an instrument for measuringelectric current in amperes.

ammonia maser first maser, invented byCharles H. Townes. Such a maser operatesat microwave frequencies.

ammonium dihydrogen phosphate (ADP)a strong linear electro-optic material. Its

chemical formula isNH4H2PO4. See alsopotassium dihydrogen phosphate (KDP).

amorphous alloy a ferromagnetic mate-rial with very low coercive force (i.e., a nar-row hysteresis loop). The material is formedas a very thin ribbon, by freezing the moltingalloy before it can crystallize, thus providinga random molecular orientation.

amortisseur winding See damperwinding.

ampacity the maximum current whichcan be safely carried by a conductor underspecified conditions.

ampere interrupting rating the inter-rupting rating of a device expressed in amps(often rms symmetrical amps).See alsoMVA interrupting rating.

Ampere’s Law a fundamental rela-tionship in electromagnetic theory. In afairly general form it is expressed by one ofMaxwell’s equations,

∇ × H(r , t) = ∂D(r , t)∂t

+ J(r , t)

wheret is the time,r is the coordinate vector,and the other vectors are defined asD(r , t)electric displacement;H(r , t), magnetic fieldstrength;J(r , t), electric current density.

Ampere, Andre Marie (1775–1836)Born: Lyon, France

Ampere is best known for his pioneeringwork in the field of Electrodynamics. Duringhis emotionally troubled life, he held severalprofessorships: at Bourg, Lyon, and at theEcole Polytechnic in Paris. While Ampereworked in several sciences, the work of theDanish physicist Hans Christian Oerstad onthe electric deflection of a compass needle, asdemonstrated to him by Dominique Arago,caused Ampere’s great interest in electro-magnetism. His seminal work,Notes on theTheory of Electrodynamic Phenomena De-duced Solely from Experiment,establishedthe mathematical formulations for electro-magnetics including what is now known asAmpere’s Law. It can be said that Amperefounded the field of electromagnetics. He ishonored for this by the naming of the unit ofelectric current as the ampere.

amperometric sensor an electrochem-ical sensor that determines the amount ofa substance by means of an oxidation–reduction reaction involving that substance.Electrons are transferred as a part of the re-action, so that the electrical current throughthe sensor is related to the amount of the sub-stance seen by the sensor.

amplidyne a special generator that actslike a DC power amplifier by using com-pensation coils and a short circuit across itsbrushes to precisely and fastly control highpowers with low level control signals.

amplified spontaneous emission sponta-neous emission that has been enhanced in am-plitude and perhaps modified in spectrum bypropagation through an amplifying medium,usually the medium in which it was first gen-erated.

amplifier a circuit element that has alinear input-output signal relationship, withgain in voltage, current, and/or power.Seealsobalanced amplifier, feedback amplifier,feedforward amplifier, laser amplifier, maseramplifier, optical amplifier, single-endedamplifier.


ASCII Code ChartHex Char Hex Char Hex Char Hex Char

00 nul 20 sp 40 @ 60 ‘01 soh 21 ! 41 A 61 a02 stx 22 " 42 B 62 a03 etx 23 # 43 C 63 c04 eot 24 $ 44 D 64 d05 enq 25 % 45 E 65 e06 ack 26 & 46 F 66 f07 bel 27 ’ 47 G 67 g08 bs 28 ( 48 H 68 h09 ht 29 ) 49 I 69 i0A lf 2A * 4A J 6A j0B vt 2B 4B K 6B k0C ff 2C , 4C L 6C l0D cr 2D - 4D M 6D m0E so 2E . 4E N 6E n0F si 2F / 4F O 6F o10 dle 30 0 50 P 70 p11 dc1 31 1 51 Q 71 q12 dc2 32 2 52 R 72 r13 dc3 33 3 53 S 73 s14 dc4 34 4 54 T 74 t15 nak 35 5 55 U 75 u16 syn 36 6 56 V 76 v17 etb 37 7 57 W 77 w18 can 38 8 58 X 78 x19 em 39 9 59 Y 79 y1A sub 3A : 5A Z 7A z1B esc 3B ; 5B [ 7B {1C fs 3C < 5C \ 7C |1D gs 3D = 5D ] 7D }1E rs 3E > 5E ˆ 7E ˜1F us 3F ? 5F _ 7F


amplitron a classic crossed-field am-plifier in which output current is obtainedprimarily by secondary emission from thenegative electrode that serves as a cathodethroughout all or most of the interactionspace.

amplitude descriptor of the strength of awave disturbance such as an electromagneticor acoustic wave.

amplitude equations a form of theSchrodinger equation that describes the evo-lution of a quantum mechanical system interms of only the coefficients of the preferredbasis states. These coefficients are knownas quantum mechanical amplitudes and con-tain both magnitude and phase information.Amplitude equations are often used to gainphysical insight into interactions of quantumsystems with electromagnetic fields.See alsoSchrodinger wave equation (SWE).

amplitude linearity qualitative measureof the extent to which the output ampli-tude of a device is a faithful reproduction ofits input, with no new frequency harmonicsadded. A perfectly linear device would out-put a scaled version of its input, where theshape of the input waveform has been un-altered (i.e., there is no distortion of the in-put waveform). Viewed in the frequency do-main, the output signal would contain onlythose spectral components found in the in-put signal, and each frequency line would bescaled by the same amount (i.e., by the gainof the device).

amplitude modulation (AM) the processof modulating a signalx(t) by a carrier wavec(t) for transmission:

y(t) = c(t)x(t),

wherey(t) is the signal to be transmitted.c(t) is either a complex exponential of theform

c(t) = ej (ωct+θc)

or a sinusoidal signal of the form

c(t) = cos(ωct + θc).

ωc is referred to as thecarrier frequency. AMhas the effect of shifting the frequency spec-trum ofx(t)byωc. The signal is recovered byshifting the spectrum ofx(t) back to its orig-inal form. See alsofrequency modulation.

amplitude response the magnitude of thesteady-state response of a fixed, linear systemto a unit-amplitude input sinusoid.

amplitude spectrum the magnitude ofthe Fourier transform|F(ω)|,−∞ < ω <

∞ of a signalf (t). For example, the ampli-tude spectrum of a rectangular pulse of unitwidth is given in the following figure:SeealsoFourier transform.

Amplitude spectrum.

amplitude stabilization circuit a circuitused to obtain a precise oscillation amplitudeof oscillators. These circuits are used in in-strumentation when it is required to increasethe purity of output signal and reduce the fre-quency depression (especially in Meachem-bridge oscillator with crystal) of the main har-monic by higher harmonics (van der Pol ef-fect). Three types of circuits are used:

1. An element of large inertia (tungstenlamp, thermistor) is included in the circuit ata point where it can change the magnitude offeedback, but not affect the frequency.


2. A controlled resistor (usually an FEToperating in a triode regime) that is also partof the feedback circuit (the DC control signalis obtained with a rectifier and a filter of largetime constant).

3.An automatic gain control circuit wherethe DC control signal obtained from a recti-fier and filter is used to change the bias ofoscillator active element.

amplitude-modulated link a transmitter–receiver system that utilizes amplitude-modulation for the transmission of signal fre-quencies.

amplitude-shift keying (ASK) a mod-ulation technique in which each group ofsource bits determines the amplitude of themodulated carrier.

AMPS See advanced mobile phonesystem.

AMR Seeautomated meter reading.

analog Seeanalog signal, analog data.

analog data data represented in a contin-uous form with respect to continuous time,as contrasted with digital data represented ina discrete (discontinuous) form in a sequenceof time instant.

analog multiplier a device or a circuitthat generates an analog output signal that isproportional to the product or multiplicationof two analog input signals.

analog optical computing optical com-puting that involves two-dimensional analogoperations such as correlation and complexspatial frequency filtering primarily basedon the property of lens to perform two-dimensional Fourier transform. In analogoptical computing, operations to be per-formed are matched with and based on al-ready known optical phenomena.

analog signal a signal represented in acontinuous form with respect to continuoustime, as contrasted with digital signal repre-sented in a discrete (discontinuous) form ina sequence of time instant.See alsoanalogdata.

analog signal conditioning an interfacebetween the sensor or transducer output,which represents an analog or physical world,and the analog-to-digital converter.

analog-to-digital (A/D) conversion amethod by which a continuously varying sig-nal (voltage) is sampled at regularly occur-ring intervals. Each sample is quantized to adiscrete value by comparisons to preestab-lished reference levels. These quantizedsamples are then formatted to the requireddigital output (e.g., binary pulse code words).The A/D converter is “clocked” to provideupdated outputs at regular intervals. In ordernot to lose any baseband information, sam-pling must occur at a rate higher than twicethe highest incoming signal frequency com-ponent.See alsoNyquist rate.

analog-to-digital (A/D) converter a de-vice that changes an analog signal to a digitalsignal of corresponding magnitude. This de-vice is also called an encoder, ADC, or A/Cconverter.

analysis-by-synthesis coding refers tothe class of source coding algorithms wherethe coding is based on parametric synthetiza-tion of the source signal at the encoder. Thesynthesized signal is analyzed, and the pa-rameters that give the “best” result are cho-sen and then transmitted (in coded form).Based on the received parameters the signalis resynthesized at the receiver.

analysis filter a filter in the analysis sec-tion of a sub-band analysis and synthesis sys-tem.


analyte the substance being measured bya chemical or bioanalytical sensor and instru-mentation system.

analytic signal refers to a signal that has aFourier transform that is zero valued for neg-ative frequencies; i.e., the signal has a one-sided spectrum.

analytical Jacobian a mathematical rep-resentation computed via differentiation ofthe direct kinematic equation with respect tothe joint variablesq. Formally one can write

x =[φ

p

]=[Jφ(q)

Jp(q)

]q = JA(q)q where the

analytical Jacobian isJA(q) = ∂k(q)∂q

. Seeexternal spacefor notation used in these equa-tions. The analytical Jacobian is differentfrom the geometric Jacobian, since the end-effector angular velocity with respect to thebase frame is not given byφ. Both Jacobiansare related asJ = TA(φ)JA whereTA(φ) isa matrix that depends on the particular repre-sentation of the orientation representation. InparticularTA(φ) is an identity matrix whenequivalent axis of rotation in the task spaceis the same as the equivalent axis of rota-tion of the end-effector.See alsogeometricJacobian.

anamorphic lenses a lens system havinga difference in optical magnification alongthe two mutually perpendicular axes (ver-tical plane or tilt vs. horizontal plane orpanorama).

and SeeAND.

AND the Boolean operator that imple-ments the conjunction of two predicates. Thetruth table for∧ ≡ X andY is

X Y X ∧ YF F F

F T F

T F F

T T T

n-ary ands can be obtained as conjunction ofbinary ands.

AND gate a device that implements theBoolean AND operation.SeeAND.

angle diversity a diversity technique usedin radio communications based on receivinga signal over multiple arrival angles. Thesignal components are typically affected byuncorrelated fading processes and are com-bined in the receiver to improve performance.The main combining methods are selectiondiversity, equal gain combining, and maxi-mal ratio combining.

angle modulation a type of modulationwhere either the frequency (FM) or the phase(PM) of a carrier are varied.

angle of arrival (AOA) the direction toa source emitting a signal impinging on asensor array. Also called direction of arrival(DOA).

angstrom popular unit not officially rec-ognized as part of the SI unit system. Equalto 10−10 meters. Abbreviated_A. Named af-ter Anders_Angstrom (1814–1874).

angular alignment loss the optical powerloss in an optical connection between two op-tical fibers, between an optical source and afiber, or between an optical fiber and a detec-tor caused by the angular misalignment of theaxes of the source and fiber, the two fibers,or the fiber and detector.

angular frequency the rate of change ofthe phase of a wave in radians per second.

anisotropic direction-dependent.

anisotropic diffraction diffraction whenthe refractive indices for the incident anddiffracted optical waves are different.

anisotropic diffusion a process of pro-gressive image smoothing as a function of atime variablet , such that the degree and ori-entation of smoothing at a point varies ac-cording to certain parameters measured at


that point (e.g., gray-level gradient, curva-ture, etc.) in order to smooth image noisewhile preserving crisp edges. The progres-sively smoothed imageI (x, y, t) (wherex, yare spatial coordinates andt is time) satisfiesthe differential equation

∂I/∂t = div(c∇I ),where the diffusion factorc is a decreasingfunction of the spatial gradient∇I . Whencis constant, this reduces to the heat diffusionequation

∂I/∂t = c1I.

Other mathematical formulations have beengiven, where edge-preserving smoothing isrealized by a selective diffusion in the di-rection perpendicular to the gradient.Seealso multiresolution analysis, mathematicalmorphology.

anisotropic etch an etch with an etchrate that is direction-dependent. In wet etch-ing, the direction dependence has to do withcrystallographic axis – some planes etch atdifferent rates than others.

anisotropic medium (1) a medium inwhich the index of refraction varies withthe light propagation direction within themedium. In such a medium, the constitutiverelation involves a tensor.

(2) a medium that exhibits anisotropy. Ex-amples are anisotropic crystals, ferrites inthe presence of a static magnetic field, andplasma in the presence of a static magneticfield.

anisotropic scatterer inhomogeneousmedium, usually consisting of suspension ofanisotropic molecules, capable of producingeffects like birefringence or dichroism. Assuch, its dielectric permittivity is a tensor act-ing differently upon each component of theelectromagnetic field.

anisotropy (1) the degree of variation ina property such as index of refraction withlight propagation direction.

(2) dependence of the response of amedium on the direction of the fields, for ex-ample, thex component of the electric dis-placement might depend in part on they com-ponent of the fields.

annealing a process often used in semi-conductor processing to cause a change inmaterials or device properties to improve thecircuit performance and/or reliability.Seealsosimulated annealing.

annealing schedule specifies the se-quence of temperature values that are to beused in an application of simulated anneal-ing and also specifies the number of param-eter changes that are to be attempted at eachtemperature.

annihilation a process in which a particleand its anti-particle meet and convert sponta-neously into photons.

annul bit a bit that is used to reduce theeffect of pipeline breaks by executing the in-struction after a branch instruction. The an-nul bit in a branch allows one to ignore thedelay-slot instruction if the branch goes thewrong way. With the annul bit not set, thedelayed instruction is executed. If it is set,the delayed instruction is annulled.

annular cathode a cathode of a vacuumtube with the shape of the emitting surface ofthe cathode is annular. The annular cathodecan produce a hollow electron beam.

annular illumination a type of off-axisillumination where a doughnut-shaped (an-nular) ring of light is used as the source.

anode the positive electrode of a device.Contrast withcathode.

anomalous dispersion decrease of the in-dex of refraction with increasing frequency;tends to occur near the center of absorbingtransitions or in the wings of amplifying tran-sitions.


ANSI American National Standards In-stitute, a body which administers numerousindustrial standards in the USA includingseveral which pertain to electric utility con-struction practices.SeeAmerican NationalStandards Institute.

antenna a device used to couple energyfrom a guiding structure (transmission line,waveguide, etc.) into a propagation medium,such as free space, and vice versa. It providesdirectivity and gain for the transmission andreception of electromagnetic waves.

antenna beamwidth the effective an-gular extent of the antenna radiation patternusually specified between points of fixed am-plitude relative to the main lobe gain (e.g.,−3 dB points).

antenna diversity a diversity techniquebased on the use of multiple antennas eitherat the receiver (receiver antenna diversity) orat the transmitter (transmitter antenna diver-sity) in a radio communication link. If theseparation of antennas is sufficient, the signalcomponents are affected by different fadingprocesses and are combined in the receiverto improve performance.See alsoRAKEreceiver. Contrast withangle diversity.

antenna gain the maximum ratio of anantenna’s ability to focus or receive power ina given direction relative to a standard; thestandard is usually an isotropic radiator or adipole. The gain includes the efficiency ofthe antenna.

antenna noise temperature the effectivenoise temperature of the antenna radiation re-sistance appearing at the antenna terminals.At a given frequency, the antenna noise tem-perature,Ta ( K), can be calculated asPn

kBwherePn is the noise power available at theantenna terminals (W),k is Boltzmann’s con-stant (1.38×10−23 J/0K), andB is the band-width (Hz). The antenna noise is the resultof thermal noise generated in ohmic losses in

the antenna structure and noise received bythe antenna from external radiating sources.

antenna pattern graph or chart repre-senting the absolute or normalized antennagain as a function of angle (typically azimuthor elevation) and used to describe the direc-tional properties of an antenna. In the nearfield, the antenna pattern is a function of thedistance from the antenna whereas in the farfield, the pattern is independent of distancefrom the antenna.

antenna Q ratio of the energy stored to theenergy dissipated (ohmically or via radiation)per cycle.

antenna synthesis the process of de-termining or designing an antenna to yielda given radiation pattern. Several synthesismethods exist. Some are closed form solu-tions and some use numerical techniques.

anthropomorphic manipulator a ma-nipulator that consists of two shoulder joints,one for rotation about a vertical axis and onefor elevation out of the horizontal plane, anelbow joint with axis parallel to the shoulderelevation joint, and two or three wrist joints atthe end of the manipulator (see figure). Ananthropomorphic manipulator is sometimescalled a jointed, elbow, or articulated manip-ulator.

An anthropomorphic manipulator.


anti-plugging a feature to prevent a mo-tor from reversing direction directly acrossthe line. The purpose of the anti-pluggingcoil and contact is to prevent the motor fromstarting in the opposite direction until thespeed has slowed enough where the currentand torque surges are within acceptable lev-els when changing direction.

anti-Stokes scattering the scattering oflight accompanied by a shift to higher fre-quencies.Contrast withStokes Law of lightscattering.

antialiasing filter typically, a filter thatprovides a prefiltering operation to ensurethat the frequency components of a signalabove the Nyquist frequency are sufficientlyattenuated so that, when aliased, they willcause a negligible distortion to the sampledsignal.See alsoaliasing,Nyquist frequency.

anticollision radar a type of radar, gen-erally operating in the millimeter wave fre-quency range, used to prevent collision be-tween moving vehicles.

anticomet tail (ACT) a special type ofelectron gun designed to handle highlightsby increasing beam current with a defocusedbeam during line retrace.

antidependency a potential conflict be-tween two instructions when the second in-struction alters an operand which is read bythe first instruction. For correct results, thefirst instruction must read the operand beforethe second alters it. Also called a write-after-read hazard.

antidots regions of repulsive potential,but which are configured so that particles(usually electrons) can pass around the poten-tial and proceed past it. In the limiting case,a repulsive Coulomb potential is the simplestantidot structure.

antiferromagnetic materials in which theinternal magnetic moments line up antiparal-

lel, resulting in permeabilities slightly greaterthan unity; unlike paramagnetic substances,these materials exhibit hysteresis and have aCurie temperature. Examples include man-ganese oxide, nickel oxide, and ferrous sul-fide.

antifuse a fuse-like device that when ac-tivated becomes low-impedance.

antiparticle a particle having the samemass as a given fundamental particle, butwhose other properties, while having thesame magnitude, may be of opposite sign.Each particle has a partner called an antipar-ticle. For example, electrical charge in thecase of the electron and positron, magneticmoment in the case of the neutron and an-tineutron. On collision a particle and its an-tiparticle may mutually annihilate with theemission of radiation. Some properties ofthe antiparticle will be identical in magni-tude but opposite in sign to the particle it ispaired with.

antipodal symmetry created by simulta-neously mirroring an object in both the X andY axes.

antiproton antiparticle to the proton. Itis a strongly interacting baryon carrying unitnegative charge. It has mass of 938 MeV andcarries spin 1/2.

antireflection coating Seeantireflectivecoating.

antireflective coating (ARC) a coatingplaced on top or below the layer of pho-toresist to reduce the reflection of light, andhence reduce the detrimental effects of stand-ing waves or thin film interference.

AOA Seeangle of arrival.

AOTF Seeacousto-optic tunable filter.

APART/PADE a computer code for anal-ysis of stray light in optical systems devel-


oped by the University of Arizona and BRO,Inc.

APC-7 connector common term for am-phenol precision connector - 7mm. A “sex-less” coaxial connector with butt contact be-tween both the inner and outer conductorscapable of low standing wave ratios to fre-quencies up to 18 GHz.

APD Seeavalanche photodiode.

aperiodic convolution the convolution oftwo sequences.Seeconvolution.

aperiodic signal a signal that is not pe-riodic, i.e., one for whichx(t) 6= x(t + T ).This means that the signalx(t) has a propertythat is changed by a time shiftT . See alsoperiodic signal.

aperiodic waveform this phrase is usedto describe a waveform that does not repeatitself in a uniform, periodic manner.Com-pare withperiodic waveform.

aperture (1) an opening to a cavity, orwave-guide, from which radiation is eitherreceived or transmitted. Typically used asantenna or a coupling element.

(2) a physical space available for beam tooccupy in a device. Aperture limitations arethe physical size of the vacuum chamber; amagnetic field anomaly may deflect the beamso that the full available aperture cannot beused.

aperture antenna an antenna with aphysical opening, hole, or slit.

aperture correction signal compensa-tion used to correct the distortion caused bythe non-zero aperture of a scanning electronbeam. A standardized measure of the selec-tivity of a circuit or system. The−3 dB(or half-power) band width is taken to bethe difference between the upper (f2) andlower (f1) frequencies where the gain vs. fre-

quency response curve has decreased−3 dBfrom the passband reference gain. Note thatf1 andf2 define the response passband bymarking the points at which the output powerhas decreased to one-half the value of the in-put power. For band widths extending downto DC, the upper−3 dB frequency is cited asthe 3 dB bandwidth.

aperture coupling a method of couplinga transmission line to an antenna in whichfields leak through an aperture in a metallicground plane separating the line from the an-tenna.

aperture efficiency a figure of merit thatdetermines how much of the incident energyis captured by an aperture. It depends on thephysical dimensions of the aperture.

aperture problem given a sequence ofimages over time we would like to infer themotion (optical flow) field. Based on localimage information (i.e., based on the valuesof those pixels falling within some aperture)only the component of motion along the gray-level gradient can be inferred; that the com-ponent of motion perpendicular to the gray-level gradient can only be known by resortingto global methods is known as the apertureproblem.Seeoptical flow, optical flux.

APL Seeaverage picture level.

APLC Seeactive power line conditioner.

apodization (1) a deliberate variation inthe transmission of an optical aperture as afunction of distance from the center or edges,in order to control optical transfer functions.

(2) a deliberate variation in the strength ofa signal with time.

apparent concurrency within an intervalof time more than one process executes on acomputer, although at the instruction level,instructions from only one process run at anysingle point in time.See alsoconcurrency.


apparent mean thermal conductivitythe effective thermal conductivity of an as-semblage of material (Pearlite, super insula-tion) between specified temperatures.

apparent power (1) in an AC system,the product of voltage,E and current,I . Ap-parent power (or total power) is composed oftwo mutually independent components — anactive component (real power), and a reac-tive component (imaginary power). Appar-ent power is denoted byS, and has the unitof voltamperes.

(2) the scalar product of the voltage andcurrent delivered to the load. It can also beexpressed as the vectorS = P + jQ, whereP = real power andQ = reactive power.

application-specific integrated circuit(ASIC) an integrated circuit designed forone particular application.

appropriate technology the technologythat will accomplish a task adequately giventhe resources available. Adequacy can beverified by determining that increasing thetechnological content of the solution resultsin diminishing gains or increasing costs.

approximate coding a process, definedwith respect to exact coding, that deals withirreversible and information-lossy process-ing of two-level pictures to improve compres-sion ratio with significant degradation of pic-ture quality. Exact coding schemes dependon the ability to predict the color of a pixelor the progression of a contour from line toline. Irreversible processing techniques try toreduce prediction errors by maintaining thecontinuity of the contours from line to line.With predictive coding the number of pix-els can be changed to reduce those havingnonzero prediction error. With block codingthe compression efficiency can be improvedby increasing the probability of occurrenceof the all zero block. The third approximateblock coding scheme is pattern matching. Inthis scheme the identification codes of therepeated patterns are transmitted to the re-

ceiver. A library of patterns is maintained forcontinuous checking.See alsoexact coding.

approximate reasoning an inferenceprocedure used to derive conclusions froma set of fuzzy if-then rules and some con-ditions (facts). The most used approximatereasoning methods are based on the general-ized modus ponens.See alsofuzzy IF-THENrule, generalized modus ponens, linguisticvariable.

approximately controllable systeman infinite-dimensional stationary linear dy-namical system where the attainable setK∞is dense in the infinite-dimensional statespaceX. The set is said to be approximatelycontrollable in [0, T ] if the attainable setK(0, T ) is dense in the infinite-dimensionalstate spaceX.

Approximate controllability in[0, T ] al-ways implies approximate controllability.The converse statement is not always true.

Ar + laser laser in which the activemedium consists of singly ionized argonatoms. Ar+ lasers have several wavelengthsin the visible portion of the spectrum.

Arago, Dominique Francois (1786–1853) Born: Estagel, France

Arago is best known for the breadth andthe volume of his contribution to the studyof light and for his work with Ampere on thedevelopment of electrodynamics. Arago dis-covered that iron could be magnetized by thepassage of current through a wire and, thephenomenon of magnetic rotation. It wasleft to Michael Faraday to properly explainthis phenomenon. Arago spent a significantamount of time involved in politics and suc-ceeded Jean Fourier as the permanent secre-tary to The Academy of Sciences in 1830. Ithas been suggested that Arago’s enthusiasmand work ethic were an inspiration to manycontemporary scientists.


arbiter a unit that decides when multi-ple requestors may have access to a sharedresource.

arbitrary reference frame a two-dimensional space that rotates at an unspec-ified angular velocityω. In electric ma-chines/power system analysis, an orthog-onal coordinate axis is established in thisspace upon which fictitious windings areplaced. A linear transformation is estab-lished in which the physical variables of thesystem (voltage, current, flux linkage) arereferred to variables of the fictitious wind-ings. See alsorotor reference frame, stationary referenceframe, synchronous reference frame.

arbitration Seebus arbitration.

ARC Seeantireflective coating.

arc detector a device placed within amicrowave power tube or within one or moreof the external cavities of a microwave powertube whose purpose is to sense the presenceof an overvoltage arc.

arc fault interrupter the mechanism thatbreaks the fault current arc in a power circuitbreaker.

arc lamp lamp made by driving a highcurrent across a gap between two electrodes.Some types operate in air consuming the elec-trode, for example, a carbon arc in whichthe electrode material is made as a rod andfed into the discharge to replace what is con-sumed. Others operate in a vacuum envelopthat reduces the electrode consumption.

arc resistance period of time that the sur-face of an insulating material can be submit-ted to the action of an electrical arc withoutbecoming conductive.

architecture Seecomputer architecture.

arcing fault Seearcing ground.

arcing ground a ground fault on a powerline which alternately clears and restrikes,causing high, repetitive voltage surges.

areal density a measure for the improve-ment in the capacity of a disk. It is the prod-uct of the number of tracks per inch and thenumber of bits per inch, i.e., it is the numberof bits per square inch.

argon ion laser SeeAr+ laser.

argument (1) an address or value that ispassed to a procedure or function call, as away of communicating cleanly across proce-dure/function boundaries.

(2) a piece of data given to a hardwareoperator block.

arithmetic and logic unit (ALU) a com-binational logic circuit that can perform basicarithmetic and logical operations on n-bit bi-nary operands.

arithmetic coding a method (due to Elias,Pasco, Rissanen and others) for lossless datacompression. This incremental coding algo-rithm works efficiently for long block lengthsand achieves an average length within onebit of the entropy for the block. The namecomes from the fact that the method utilizesthe structures of binary expansions of the realnumbers in the unit interval.

arithmetic instruction a machine in-struction that performs computation, such asaddition or multiplication.

arithmetic operation any of the follow-ing operations and combination thereof: ad-dition, subtraction, multiplication, division.

arithmetic radian center frequency thelinear radian center frequency, it is the mid-


point between the higher (ωH ) and lower(ωL) band edges, expressed in units of ra-dians/second. The band edges are usuallydefined as the highest and lowest frequencieswithin a contiguous band of interest at whichthe loss equalsLAmax, the maximum attenu-ation loss across the band.

ωoa = ωH + ωL

2

arithmetic shift a shift in which it is as-sumed that the data being shifted is integerarithmetic in nature; as a result, the sign bitis not shifted, thereby maintaining the arith-metic sign of the shifted result.See alsological shift.

arithmetic–logic unit Seearithmetic andlogic unit.

arm a part of a robot. A robot is composedof an arm (or mainframe) and a wrist plusa tool. For many industrial robots the armsubassembly can move with three degrees offreedom. Hence, the arm subassembly is thepositioning mechanism.See alsoindustrialrobot.

arm pin a pin insulator .

ARMA Seeauto-regressive moving-aver-age model.

armature the magnetic circuit of a ro-tating electrical machine, including the maincurrent carrying winding, in which an alter-nating voltage is induced by the magneticfield.

armature circuit components of the ma-chine that carry armature current. For ex-ample, in a DC machine the armature cir-cuit could consist of the armature windings,brushes, series field winding, compensat-ing windings, interpoles, starting resistor(s),main-line contacts, and overload sensor.

armature current limiting a conditionwherein the stator currents are clamped at the

maximum allowable limit due to excessiveheating of the stator.

armature reaction (1) in DC machines, adistortion of the field flux caused by the fluxcreated by the armature current. Armaturereaction in a DC machine causes lower fluxat one pole-tip and higher flux at the other,which may lead to magnetic saturation. Italso shifts the neutral axis, causing sparkingon the commutator.

(2) in AC synchronous machines, a volt-age “drop” caused by the armature current.In the steady state model of the synchronousmachine, the armature reaction is accountedfor by a component of the synchronous reac-tance.

armature voltage control a method ofcontrolling the speed of a DC motor by vary-ing the voltage applied to the armature whilekeeping the voltage applied to the field circuitconstant.

armature winding an arrangement ofcoils carrying the main current, typicallywound on the stator of a synchronous ma-chine or the rotor of a DC machine, in whichan alternating voltage is induced by the mag-netic field.

armless construction a method of distri-bution line construction, often used for aes-thetic purposes, in which pin insulators aremounted on steel brackets bolted directly toa utility pole without the use of a crossarm.

Armstrong oscillator Hartley oscillatorsare usually not used at VHF of higher fre-quencies. Similarly, the circuit is avoided atvery low audio frequencies. It is importantto distinguish the Hartley oscillator from theArmstrong topology. In the Armstrong oscil-lator, no ohmic connection exists between thetwo inductors. Instead, coupling is entirelymagnetic.

Armstrong, Edwin Howard (1890–1954) Born: New York, New York, U.S.A.


Armstrong is best known as the developerof frequency modulation (FM) radio and in-ventor of the superheterodyne receiver. Arm-strong spent most of his career at ColumbiaUniversity. During his life, his inventionsmade him quite wealthy. The superhetero-dyne receiver was purchased as a way for themilitary to detect the spark plug ignitions ofapproaching aircraft. Patent fights with LeeDeForest and the difficulty in promoting FMradio led to bitterness and frustration whichmany felt led to his suicide.

ARQ Seeautomatic repeat request.

array several antennas arranged togetherin space and interconnected to produce a de-sired radiation pattern.

array factor in antenna theory, the re-sulting radiation pattern of an array wheneach antenna in the array is replaced by anisotropic radiator.

array processor an array of processorelements operating in lockstep in response toa single instruction and performing compu-tations on data that are distributed across theprocessor elements.

array signal processing signal process-ing techniques used for extracting informa-tion based on signals from several (identical)sensors, for example an antenna array con-sisting of several antenna elements.

arrester discharge current the currentin an arrester during a surge.

arrester discharge voltage the voltagein an arrester during a surge.

ART network Seeadaptive resonancetheory network.

artifact an error or aberration in a signalthat is the result of aliasing, a quantizationerror, some form of noise, or the distorting

effects of some type of processing.See alsooutlier.

artificial constraint an additional con-straint in accordance with the natural con-straints to specify desired motion or forceapplication. An artificial constraint occursalong the tangents and normals of the con-straint surface. An artificial force con-straint is specified along surface normals,and an artificial position constraint along tan-gents and hence consistency with the natu-ral constraints is preserved.See alsonaturalconstraint.

artificial dielectric a dielectric materialthat has been modified to alter its proper-ties. Common modifications include micro-machining to remove material from the sub-strate under planar patch antenna to improveradiation properties and the fabrication of pe-riodic arrays of holes to realize guiding orphotonic bandgap structures.

artificial intelligence the study of com-puter techniques that emulate aspects of hu-man intelligence, such as speech recognition,logical inference, and ability to reason frompartial information.

artificial neural network a set of nodescalled neurons and a set of connections be-tween the neurons that is intended to performintellectual operations in a manner not un-like that of the neurons in the human brain.In particular, artificial neural networks havebeen designed and used for performing pat-tern recognition operations.See alsopatternrecognition, perceptron.

artificial neuron an elementary analog ofa biological neuron with weighted inputs, aninternal threshold, and a single output. Whenthe activation of the neuron equals or exceedsthe threshold, the output takes the value+1,which is an analog of the firing of a biologicalneuron. When the activation is less than thethreshold, the output takes on the value 0 (inthe binary case) or−1 (in the bipolar case)


representing the quiescent state of a biologi-cal neuron.

artificial skin artificial skin is a de-vice which, when pressed against the surfaceby an object, causes local deformations thatare measured as continuous resistance varia-tions. The latter are transformed into electri-cal signals whose amplitude is proportionalto the force being applied to any given pointon the surface of the material of the device.

ASAP/RABET acronym for a computercode for optical systems by BRO, Inc., forstandard optical analysis and stray-light anal-ysis such as light-scattering.

ASCII SeeAmerican standard code forinformation interchange.

ASCR Seeasymmetrical silicon controlledrectifier.

ASIC Seeapplication-specific integratedcircuit.

ASK Seeamplitude-shift keying.

askarel a trade name for an insulating oil.

ASM Seealgorithmic state machine.

aspect ratio (1) the size invariant ratio oflength to width for a rectangular box enclos-ing a shape, the orientation of the box beingchosen to maximize the ratio. This measureis used to characterize object shapes as a pre-liminary to, or as a quick procedure for, ob-ject recognition.

(2) the ratio of width to height for animage or display.

(3) in television or motion pictures, thealgebraic ratio of picture width to height. Atpresent, the television format in the UnitedStates consists of a width to height ratio of 4to 3.

aspheric description of optical elementswhose curved surfaces are not spherical, of-

ten used to reduce aberrations in optical sys-tems.

assembler (1) a computer program thattranslates an assembly-code text file to an ob-ject file suitable for linking.

(2) a program for converting assemblylanguage into machine code.

assembly language a programming lan-guage that represents machine code in asymbolic, easier-to-read form. See alsoassembler.

assert (1) raising the voltage on a wire tothe “high” state, usually as a signal to someother unit.

(2) to make an assertion.

assertion (1) a Boolean expression forstating the right behavior of the program or,if hardware implemented, of a circuit.

(2) a logical expression specifying a pro-gram state that must exist or a set of condi-tions that program variables must satisfy at aparticular point during program execution.

associate mode an operating mode ofcontent addressable memories, in which astored data item is retrieved that contains afield that matches a given key.

associated reference directions amethod assigning the current and voltage di-rections to an electrical element so that a pos-itive current-voltage product always meansthat the element is absorbing power from thenetwork and a negative product always meansthat the element is delivering power to thenetwork. This method of assigning direc-tions is used in most circuit simulation pro-grams.

associative memory a memory in whicheach storage location is selected by its con-tents and then an associated data location canbe accessed. Requires a comparative witheach storage location and hence is more com-plex than random access memory. Used in


fully associative cache memory and in sometranslation look-aside buffers or page transla-tion tables of the hardware to support virtualmemory. Given the user-space address of apage it returns the physical address of thatpage in main memory. Also called contentaddressable memory (CAM).

associative processor a parallel proces-sor consisting of a number of processing el-ements, memory modules, and input–outputdevices under a single control unit. The ca-pability of the processing elements is usuallylimited to the bit-serial operations.

associativity In a cache, the number oflines in a set. Ann-way set associative cachehas n lines in each set. (Note: the term“block” is also used for “line.”)

astable multivibrator the circuit that isobtained from a closed-loop regenerative sys-tem that includes two similar amplifiers ofhigh gain connected with each other via cou-pling circuits with reactance elements. Morefrequently are used RC-coupling circuits(free-running RC-multivibrators, emitter-coupled multivibrators), yet RL-circuits,usually as transformer coils, may be used aswell (magnetic multivibrators).

astigmatism adefect associated with op-tical and electrostatic lenses where the mag-nification is not the same in two orthogonalplanes; common where beam propagation isnot along the axis of rotation of the system.

asymmetric digital subscriber line (ADSL)a digital subscriber line (DSL) in which therate from central switching office (CO) tocustomer premise is much faster than the ratefrom customer premise to CO.

asymmetric multip rocessor (1) a ma-chine with multiple processors, in which thetime to access a specific memory address isdifferent depending on which processor per-forms the request.

(2) in contrast with a symmetric multipro-cessor, asymmetric multiprocessor is a mul-tiprocessor in which the processors are notassigned equal tasks. The controller (mas-ter) processor(s) are assigning tasks to (slave)processors and controlling I/O for them.

asymmetric multivibrator a multivibra-tor where the output voltage represents a trainof narrow pulses. Most asymmetric multivi-brators use a slow charge of a large timing ca-pacitor by a small current (or via a large resis-tor) and a fast discharge of this capacitor viaa switch. The charge process determines theduration of space; the mark duration, whichcoincides with the time allowed for dischargeof the timing capacitor, is usually determinedby a small time constant of the circuit con-trolling the switch. Asymmetric multivibra-tors find applications in voltage-to-frequencyconverters. Also called multivibrators with asmall mark/space ratio.

asymmetric resonator standing-waveresonator in which either the reflectivities orthe curvatures of the primary mirrors are un-equal.

asymmetrical silicon controlled rectifier(ASCR) (1) an inverter grade SCR fab-

ricated to have limited reverse voltage capa-bility. Fabrication with asymmetrical volt-age blocking capability in the forward andreverse direction permits reduction of turn-on time, turn-off time, and conduction drop.

(2) a thyristor that has limited conduc-tion in the reverse direction to gain increasedswitching speed and low forward voltagedrop. See alsosilicon controlled rectifier(SCR).

asymptotic 2-D observer a system de-scribed by the equations

zi+1,j+1 = F1zi+1,j + F2zi,j+1

+G1ui+1,j +G2ui,j+1

+H1yi+1,j +H2yi,j+1

xi,j = Lzi,j +Kyi,j


i, j ∈ Z+ (the set of nonnegative integers) iscalled a full-order asymptotic observer of thesecond generalized Fornasini–Marchesini 2-D model

Exi+1,j+1 = A1xi+1,j + A2xi,j+1

+ B1ui+1,j + B2ui,j+1

yi,j = Cxi,j +Dui,j

i, j ∈ Z+ if

limi,j→∞

[xi,j − xi,j

] = 0

for any ui,j , yi,j and boundary conditionsxi0 for i ∈ Z+ andx0j for j ∈ Z+ wherezi,j ∈ Rn is the local state vector of the ob-server at the point(i, j), uij ∈ Rm is theinput,yi,j ∈ Rp is the output, andxi,j ∈ Rnis the local semistate vector of the model,F1,F2,G1,G2,H1,H2,L,K,E,A1,A2,B1,B2,C,D are real matrices of appropriate dimen-sions withE possibly singular or rectangular.In a similar way a full-order asymptotic ob-server can be defined for other types of the2-D generalized models.

asymptotic stability (1) an equilibriumstate of a system of ordinary differentialequations or of a system of difference equa-tions is asymptotically stable (in the sense ofLyapunov) if it is stable and the system tra-jectories converge to the equilibrium state astime goes to infinity, that is, the equilibriumxeq is asymptotically stable if it is stable and

x(t) → xeq ast → ∞ .

(2) a measure of system damping with re-gard to a power system’s ability to reach itsoriginal steady state after a disturbance.

asymptotic tracking refers to the abil-ity of a unity feedback control to follow itssetpoint exactly with zero error once all tran-sients have decayed away. Clearly this is onlyachieved by stable systems.

asymptotically stable equilibrium a sta-ble equilibrium point such that all solutions

that start “sufficiently close,” approach thispoint in time.See alsostable equilibrium.

asymptotically stable in the large theequilibrium state of a stable dynamic systemdescribed by a first-order vector differentialequation is said to be asymptotically stable inthe large if its region of attraction is the entirespace<n. See alsoregion of attraction.

asymptotically stable state the equilib-rium state of a dynamic system described by afirst-order vector differential equation is saidto be asymptotically stable if it is both con-vergent and stable.See alsostable stateandconvergent state.

asynchronous not synchronous.

asynchronous AC systems AC systemseither with different operating frequencies orthat are not in synchronism.

asynchronous bus a bus in which thetiming of bus transactions is achieved withtwo basic “handshaking” signals, a requestsignal from the source to the destination andan acknowledge signal from the destinationto the source. The transaction begins withthe request to the destination. The acknowl-edge signal is generated when the destinationis ready to accept the transaction. Avoidsthe necessity to know system delays in ad-vance and allows different timing for differ-ent transactions.See alsosynchronous bus.

asynchronous circuit (1) a sequentiallogic circuit without a system clock.

(2) a circuit implementing an asyn-chronous system.

asynchronous demodulation a tech-nique for extracting the information-carryingwaveform from a modulated signal with-out requiring a phase-synchronized carrierfor demodulation. See alsosynchronousdemodulation.


asynchronous machine Seeinductionmachine.

asynchronous operation a term to indi-cate that a circuit can operate or a communi-cation system can transmit information whenready without having to wait for a synchro-nizing clock pulse.

asynchronous system a (computer, cir-cuit, device) system in which events are notexecuted in a regular time relationship, thatis, they are timing-independent. Each eventor operation is performed upon receipt of asignal generated by the completion of a pre-vious event or operation, or upon availabilityof the system resources required by the eventor operation.

asynchronous transfer mode (ATM)method of multiplexing messages onto achannel in which channel time is divided intosmall, fixed-length slots or cells. In ATMsystems the binding of messages to slots isdone dynamically, allowing dynamic band-width allocation. ATM is asynchronous inthe sense that the recurrence of cells contain-ing information from an individual user is notnecessarily periodic.

asynchronous updating one unit at atime is selected from within a neural networkto have its output updated. Updating an out-put at any time is achieved by determiningthe value of the unit’s activation function atthat time.

AT bus bus typically used in personalcomputer IBM AT for connecting adaptersand additional memory boards. It is calledalso 16 bit ISA bus since it presents a data busat 16 bit. It presents an additional connectorwith respect to the classical ISA bus (at 8 bit)of IBM PCs based on Intel 8088.See alsoEISA.

Atanasoff, John Vincent (1903– ) Born:Hamilton, New York, U.S.A.

Atanasoff is best known for his invention,along with Clifford Berry, of the first digi-tal computer, known as the ABC (Atanasoff-Berry Computer). Unlike the many WorldWar II computer pioneers, Atanasoff’s in-terest in the topic dated to his Ph.D. the-sis research at the University of Wisconsin.After graduation Atanasoff taught physicsand mathematics at Iowa State College andcontinued to work on the problem of solv-ing lengthy calculation by electronic means.Legend has it that Atanasoff worked out thebasic structure for his new machine whilehaving a drink at an Illinois road house.Clifford Berry, an electrical engineer joinedAtanasoff to help with the construction ofthe device based on Atanasoff’s ideas. JohnMauchly, another computer pioneer oftenvisited and consulted with Atanasoff. Thesediscussions resulted in a later lawsuit that es-tablished Atanasoff as the first person to buildan electronic digital computer.

ATM Seeasynchronous transfer mode.

ATM adaptation layer (AAL) a layer inthe ATM protocol hierarchy that adapts the(small) cell-sized payloads to a form moresuitable for use by higher layer protocols. Forexample, AAL5 performs segmentation andreassembly to map between 48-byte payloadsand variable length data segments.

atmosphere a convenient measure ofpressure. 1 std atm = 14.696 psia (poundsper square inch absolute).

atmospheric attenuation decrease in theamplitude of a signal propagating through theatmosphere, due primarily to absorption andscatter.

atmospheric duct a thin layer of atmo-sphere near the earth that acts as a waveguide,the electromagnetic field, trapped within theduct, can travel over long distances with verylittle attenuation.


atom a particle of matter indivisible bychemical means, which is chemically neutral.It is the fundamental building block of thechemical elements.

atomic beam a source of atoms trav-eling primarily in one direction. In prac-tice, atomic beams are usually realized bythe expansion of an atomic vapor into a vac-uum through a small aperture. The resultingexpanding cloud of atoms is usually madenearly unidirectional by a collimator thatblocks or otherwise removes all atoms notpropagating within a narrow range of angles.

atomic force microscope (AFM) amicroscope in which a sharp probe tip isscanned across a surface, with piezoelectricceramics being used to control position inthree dimensions. The lateral (in-plane, or x-y) positions are raster scanned, while the ver-tical dimension is controlled by a feedbackcircuit that maintains constant force. The im-age produced is a topograph showing surfaceheight as a function of position in the plane.

atomic instruction an instruction thatconsists of discrete operations that are all exe-cuted as a single and indivisible unit, withoutinterruption by other system events.See alsotest-and-set instruction, atomic transaction.

atomic transaction the same as an atomicinstruction, except that the notion of beingatomic applies to a transaction, which maybe a sequence of operations, no intermedi-ate states of which may be seen or operatedupon by another transaction.See alsoatomicinstruction.

atomic transition coupling of energylevels in an atom by means of absorption oremission processes.

atomic vapor a material composed ofatoms that preferentially exist as monomersin the vapor phase.

ATRC SeeAdvanced Television ResearchConsortium.

attachment one of the events which pre-cede a lightning stroke to the earth. Attach-ment occurs when the stepped leader fromthe thundercloud makes contact with one ofseveral streamers which emanate from theground or structures on the earth. The returnstroke follows immediately.Seestreamer,stepped leader, return stroke.

attachment process a process that occursin lightning when one or more stepped leaderbranches approach within a hundred metersor so of the ground and the electric field atthe ground increases above the critical break-down field of the surrounding air. At that timeone or more upward-going discharges is ini-tiated. After traveling a few tens of meters,one of the upward discharges, which is essen-tially at ground potential, contracts the tip ofone branch of the stepped leader, which is ata high potential, completing the leader pathto ground.

attainable set for discrete system the setof all the possible ends of system trajectoriesat timet1 starting from zero initial conditionsat timet0. DenteodK(t0, t1).K(t0, t1) is defined for zero initial state as

follows

K (k0, k1) = (x ∈ Rn :

x =j=k1−1∑j=k0

F (k1, j + 1) B(j)u(j) :

u(j) ∈ Rm)Therefore, controllability in[k0, k1] for

discrete dynamical system is equivalent to thecondition

K (t0, t1) = Rn

Using the concept of the attainable set itis possible to express the remaining types ofcontrollability for discrete system.


attenuated total reflection the phe-nomenon associated with the appearance of areflection minimum identified with the gen-eration of surface waves at the metal — airinterface in a prism, air, metal arrangement.

attenuation the exponential decreasewith distance, in the amplitude of an elec-tric signal traveling along a very long trans-mission line due to losses in the support-ing medium. In electromagnetic systems at-tenuation is due to conductor and dielectriclosses. In fiber optic systems attenuationarises from intrinsic material properties (ab-sorption and Rayleigh scattering) and fromwaveguide properties such as bending, mi-crobending, splices, and connectors.

attenuation coefficient Seeabsorptioncoefficient.

attenuation constant the real part of thecomplex propagation constant for an electro-magnetic wave.

attenuator a device or network that ab-sorbs part of a signal while passing the re-mainder with minimal distortion.

attractor an asymptotic state of a dy-namical system of which there are three basictypes. Either (i) the system comes to rest andthe attractor is a fixed point in state space,(ii) the system settles into a periodic motionknown as a limit cycle, or (iii) the systementers a chaotic motion, in which case theattractor is called strange.

attribute a special function in Pawlak’sinformation system. Pawlak’s informationsystemS is a pair(U,A) where the setU iscalled the universe and hasn members de-notedxi , while the setA consists onm func-tion on the universeU. These functions arecalled the attributes and denotedaj . The at-tributes are vector-valued functions that maybe interpreted, for example, as issues undernegotiation by the members of the universeU. An example of an attribute is a function of

the form

aj : U → {−1,0,1}n .

attribute set a set of vectors (signals)lying in metric space that possess prescribedproperties.

audio science of processing signals thatare within the frequency range of hearing,that is, roughly between 20 hertz and 20 kilo-hertz. Also, name for this kind of signal.

audio channels the portion of the cir-cuit containing frequencies that correspondto the audible sound waves. Audio frequen-cies range from approximately 15 hertz to20,000 hertz.

audio coding the process of compress-ing an audio signal for storage on a digitalcomputer or transmission over a digital com-munication channel.

audio follow-video switcher (AFV) aswitcher that simultaneously switches thevideo and audio information. The term isassociated with the action of the audio signaland corresponding video signal switching to-gether.

augmented code a code constructed fromanother code by adding one or more code-words to the original code.

aural subcarrier in a composite tele-vision signal, the frequency division multi-plexed carrier placed outside the visual pass-band that carries the audio modulation. Inthe NTSC (United States) system, it is placed4.5 Mhz higher than the visual carrier.

auto-regressive moving-average model(ARMA) the discrete-time input–outputmodel in which the current output dependsboth on its past values (auto-regressive part)and the present and/or past values of the input(moving-average part).


autoassociative backpropagation networka multilayer perceptron network that is

trained by presenting the same data at boththe input and output to effect a self-mapping.Such networks may be used for dimensionalreduction by constraining a middle, hiddenlayer to have fewer neurons than the inputand output layers.

autobank an array of autotransformers .

autoconfiguration a process that deter-mines what hardware actually exists duringthe current instance of the running kernelat static configuration time. It is done bythe autoconfiguration software that asks thedevices to identify themselves and accom-plishes other tasks associated with events oc-curring during the autoconfiguration of de-vices. For instance, PCI devices have auto-configuration capabilities and do not have tobe configured by users.

autocorrelation a measure of the statis-tical dependence between two samples of thesame random process. For a random processX(t), the auto-correlation is the expectation

Rxx (t1, t2) = E [X (t1)X (t2)] .

See alsocross-correlation.

autocorrelation function the expectedvalue of the product of two random vari-ables generated from a random process fortwo time instants; it represents their inter-dependence. The Fourier transform of theautocorreclation function is the power spec-trum (power spectral density) for the randomprocess.

autocorrelator a circuit that computesthe autocorrelation function.

autocovariance (1) for a random processf (t), a measure of the variability of the mean-removed process:

Cf (t1, t2) = E[f (t1)f (t2)

T]

−E[f (t1)]E[f (t2)T ].

(2) for a random vectorx, a measure of themean-square variability of a random vectorx

about its mean:

3x = E[(x − E[x])(x − E[x])T

].

See alsoautocorrelation, covariance.

autodecrementing (1) an addressingmode in which the value in a register is decre-mented by one word when used as an address.

(2) in high-level languages, operation

i −− ⇒ i = i − 1

wherei is arbitrary variable, register or mem-ory location.

(3) in machine code, more generally, theprocessor decrements the contents of the reg-ister by the size of the operand data type;then the register contains the address of theoperand. The register may be decrementedby 1, 2, 4, 8, or 16 for byte, word, long-word, quadword, or octaword operands, re-spectively.

autoincrementing (1) an addressingmode in which the value in a register is incre-mented by one word when used as an address.

(2) in high-level languages: operation

i ++ ⇒ i = i + 1

wherei is arbitrary variable, register or mem-ory location.

(3) in machine code, after evaluating theoperand address contained in the register, theprocessor increments the contents of the reg-ister by 1, 2, 4, 8, or 16 for a byte, word, long-word, quadword, or octaword, respectively.

automated meter reading (AMR) theuse of meters which have the capability oftransmitting at the least consumption infor-mation to the utility through some means ofelectronic communication.

automatic (1) property pertaining to aprocess or a device that functions without in-tervention by a human operator under speci-fied conditions.


(2) a spring-loaded tension sleeve intowhich a conductor or other wire is insertedfor tensioning and attachment to a pole orother fixture.

automatic allocation allocation of mem-ory space to hold one or more objects whoselifetimes match the lifetime of the activationof a module, such as a subroutine. Automaticallocations are usually made upon entry to asubroutine.

automatic black-level control electroniccircuitry used to maintain the black levels ofthe video signal at a predetermined level. Theblack level reference is either derived fromthe image or from the back porch of the hor-izontal blanking interval.

automatic chroma control (ACC) ACCis used to correct the level of the input chromasignal. Typically, the ACC circuitry makescorrections to the chroma, based on the rela-tive degeneration of the color burst referencesignal, since this signal will have been sub-jected to the same degradation.

automatic circuit recloser Seerecloser.

automatic fine tuning (AFT) one ofthe input circuits of a color television re-ceiver specifically designed to maintain thecorrect oscillator frequency of the tunerfor best color reproduction of the picture.The circuit is sometimes called the auto-matic frequency control.See alsoautomaticfrequency control (AFC).

automatic focusing on an optical disk,the process in which the distance from the ob-jective focal plane of the disk is continuouslymonitored and fed back to the disk controlsystem in order to keep the disk constantly infocus.

automatic frequency control (AFC) (1)an automatic feedback control system thatis used to maintain active power balance bymeans of the speed governor system. In an

interconnected system, scheduled power in-terchanges are maintained by means of con-trolling area generations.

(2) electronic circuitry used to keep the re-ceived signal properly placed within the de-sired IF frequency range. In televisions, theAFC circuitry is also called the AFT or “auto-matic fine tuning” section. The AFC circuitwill generate an error signal if the input fre-quency to the IF drifts above or below theIF frequency. The error signal is fed backto vary the local oscillator frequency in thetuner section.See alsoautomatic fine tuning(AFT).

automatic gain control (AGC) a methodto control the power of the received signalin order to be able to use the full dynamicrange of the receiver and to prevent receiversaturation.

automatic generation control (AGC)phrase describing the computer-based pro-cess by which electric utilities control indi-vidual generating stations to maintain systemfrequency and net interchange of power on ahighly interconnected transmission grid. Au-tomatic generation control (AGC) systemsmonitor grid frequency, actual and scheduledpower flows, and individual plant output tomaintain balance between actual and sched-uled power production, both within transmis-sion control areas and at individual gener-ating stations. Control is generally accom-plished by adjusting the speed control (ordroop) characteristics of individual generat-ing units. Control actions are determined byplanned production schedules and power ex-change agreements among participating util-ities.

automatic level control (ALC ) a feed-back system where an RF signal from asource is sampled, detected, and sent to avoltage controlled attenuator to maintain aconstant amplitude output over a specifiedband of frequencies.


automatic repeat request (ARQ) an er-ror control scheme for channels with feed-back. The transmitted data is encoded forerror detection and a detected error results ina retransmission request.

automatic tracking on an optical disk,the process in which the position of the diskhead relative to the disk surface is constantlymonitored and fed back to the disk controlsystem in order to keep the read/write beamconstantly on track.

automatic transfer switch a self-actingswitch which transfers one or more load con-ductor connections from one power source toanother.

automatic voltage regulator (AVR) anautomatic feedback control system that is re-sponsible for maintaining a scheduled volt-age either at the terminals of a synchronousgenerator or at the high-side bus of the gen-erator step-up transformer. The control isbrought about by changing the level of ex-citation.

automation refers to the bringing to-gether of machine tools, materials handlingprocess, and controls with little worker inter-vention, including

1. a continuous flow production processthat integrates various mechanisms to pro-duce an item with relatively few or no workeroperations, usually through electronic con-trol;

2. self-regulating machines (feedback)that can perform highly precise operations insequence; and

3. electronic computing machines.In common use, however, the term is of-

ten used in reference to any type of advancedmechanization or as a synonym for techno-logical progress; more specifically, it is usu-ally associated with cybernetics.

automaton (1) a fundamental conceptin mathematics, computer engineering, androbotics.

(2) a machine that follows sequence of in-structions.

(3) any automated device (robots, me-chanical and electromechanical chess au-tomata). Automata (plural of automa-ton) theory studies various types of au-tomata, their properties and limitations.Seealsocellular automaton, finite state machine(FSM).

autonomic that part of the nervous systemwhich controls the internal organs.

autonomous operation operation of a se-quential circuit in which no external signals,other than clock signals, are applied. Thenecessary logic inputs are derived internallyusing feedback circuits.

autonomous system a dynamic systemdescribed by a first-order vector differentialequation that is unforced and stationary. Inother words, such a system is governed by anequation of the form

x(t) = f (x(t))

Seealso unforced system and stationary .

autoregressive (AR) apth order autore-gressive process is a discrete random pro-cess that is generated by passing white noisethrough an all-pole digital filter havingppoles. Alternatively,x[n] is apth order ARprocess if

x[n] =n−1∑i=n−p

α[i]x[i] + q[n].

Autoregressive processes are often used tomodel signals since they exhibit several use-ful properties.See alsomoving average.

autotransformer a power transformerthat has a single continuous winding perphase, part of this winding being commonto both the primary and the secondary sides.As a result, these voltages are not isolated


but the transformer is reduced in weight andsize. Autotransformers are most suited forrelatively small changes in voltage. Threephase autotransformers are by necessity con-nected in a wye configuration.

autotransformer starter a single three-phase autotransformer or three single phasetransformer used to start induction motors ata reduced voltage.

auxiliary memory See secondarymemory.

auxiliary relay a relay employed inpower system protection schemes that doesnot directly sense fault presence and loca-tion. Typical auxiliary relays include lockoutrelays, reclosing relays, and circuit breakeranti-pump relays.

auxiliary winding a winding designed tobe energized occasionally for a specific pur-pose, such as starting a single-phase motor.The power to the winding may be controlledby various means including a timer, centrifu-gal switch, current sensing relay, or voltage(counter EMF) sensing relay.

availability the probability that a systemis operating correctly and is available to per-form list functions at the instant of timet .Also defined as the value

1 − outage

See alsooutage.

available bit rate (ABR) ATM conges-tion control algorithm that enables a source todiscover the bit rate available between it anda destination in a network. The source trans-mits a resource manager cell containing thedesired bit rate; each switch this cell passesthrough adjusts the bit rate down to what itcan support. Upon reaching the destination,the cell contains the available bit rate and isreturned to the source.

available power gain ratio of poweravailable from a network to the power avail-able from the source.

avalanche breakdown process that oc-curs in a semiconductor space charge regionunder a sufficiently high voltage such that thenet electron/hole generation rate due to im-pact ionization exceeds certain critical value,causing the current to rise indefinitely dueto a positive feedback mechanism. The I-Rheating caused during this process can per-manently degrade or destroy the material.

avalanche injection the physics wherebyelectrons highly energized in avalanche cur-rent at a semiconductor junction can pene-trate into a dielectric.

avalanche photodiode (APD) a photodi-ode (detector) that provides internal currentgain. Used in optical communication sys-tems when there is limited optical power atthe receiver.

average optical power time average ofthe optical power carried by a non-CW opti-cal beam.

average picture level (APL) describesthe average (mean) changes in a video signaldue to a changing brightness of the visual im-age. The APL is typically expressed in termsof a percentage (10-15% for dark pictures and75-90% for bright pictures). Changes in theAPL can effect linearity unless DC restora-tion or clamping circuits are included in thevideo circuitry.

average power the average value, takenover an interval in time, of the instantaneouspower. The time interval is usually one pe-riod of the signal.

average-value model a mathematicalrepresentation in which the average value ofvariables are used to model a system. In elec-tric machines and drives, system variables aretypically averaged over various switching in-


tervals. This eliminates the high-frequencydynamics, but preserves the slower dynam-ics of the system.

averaging the sum ofN samples, im-ages or functions, followed by division ofthe result byN . Has the effect of reduc-ing noise levels. See alsoblurring, imagesmoothing, mean filter, noise smoothing,noise suppression, smoothing.

AVR Seeautomatic voltage regulator.

AWG American Wire Gauge, a systemof wire sizing used in the USA especiallyin smaller conductors used in residential andcommercial wiring.

AWGN Seeadditive white Gaussiannoise.

axon the conducting portion of a nervefiber — a roughly tubular structure whosewall is composed of the cellular membraneand is filled with an ionic medium.

Ayrton, William Edward (1847–1908)Born: London, England

Ayrton is best known as the inventor ofa number of electrical measurement devicesand as an engineering educator. Ayrton’searly work was with the Indian TelegraphService, after which he studied with WilliamThomson (Lord Kelvin) in Glasgow. Af-ter several more telegraph assignments Ayr-ton traveled to Tokyo, where he establishedthe first electrical engineering teaching lab-oratory at the Imperial Engineering College.Among his many inventions he is creditedwith the ammeter and an improved voltmeter.His wife Bertha was also an active researcherand became the first woman to be admittedto the Institute of Electrical Engineers.

azimuth recording a recording schemewhereby the data is recorded at an acute anglefrom the direction of movement of the record-ing medium. Used in the recording schemeof video information, FM radio, and audio inVCRs.


BB coefficient Seeloss coefficient.

B-ISDN See broadband integratedservices digital network.

B-mode display returned ultrasoundechoes displayed as brightness or gray-scalelevels corresponding to the amplitude versusdepth into the body.

B-site in a ferroelectric material with thechemical formula ABO3, the crystalline lo-cation of the B atom.

B-spline the shortest cubic spline consist-ing of different three-degree polynomial onfour intervals; it can be obtained by convolv-ing four box functions.

Babbage, Charles (1792–1871) Born:Totnes, England

Babbage is best known for his ideas onmechanical computation. Babbage is said tohave been disgusted with the very inaccuratelogarithm tables of his day, as well as ap-palled by the amount of time and people ittook to compute them. Babbage attemptedto solve the problem by building mechan-ical computing engines. The government-funded Difference Engine was beyond thetechnology of the craftsman who attemptedto build it. Undeterred, Babbage followedthis failure with the larger and more complexAnalytical Engine (also unfinished). Theideas behind the Analytical Engine formedthe basis for Howard Aiken’s 1944 MarkI computer. Babbage’s assistant, Ada Au-gusta, the Countess of Lovelace and the poetLord Byron’s daughter, is honored as the firstprogrammer for her work and because hermeticulous notes preserved the descriptionsof Babbage’s machines.

Babinet principle principle in optics thatstates that the diffraction patterns producedby complementary screens are the same ex-cept for the central spot. It can be rigorouslyproved both for acoustic and electromagneticwaves. The Babinet’s principle for scalarfields is the following: letp be the resul-tant field inz > 0 due to the incident fieldpifrom z < 0 and letpt be the total field whenthe same incident wave falls on the comple-mentary screen. Then, inz > 0,

p + pt = pi

back in a motor, the end that supports themajor coupling or driving pulley.

back EMF Seecounter-EMF.

back end that portion of the nuclear fuelcycle which commences with the removal ofspent fuel from the reactor.

back porch a 4.7 microsecond region inthe horizontal blanking interval of the NTSCcomposite video signal that contains a burstof eight to ten cycles of the 3.579545 MHz(3.58 MHz) color subcarrier. The back porchoccupies 7% of the total horizontal line time;starting at the end of the horizontal line syncsignal and ending with the start of the video.

backbone wiring that runs within and be-tween floors of a building and connects local-area network segments together.

backfeed in power distribution work,power which flows from the secondary linesinto the primary lines through the distributiontransformer,e.g.,from an emergency genera-tor connected to customer load.

backflash an arc which forms along atower during a lightning strike due to hightower or footing impedance.

background (1) refers to the receivedvector power level of an electromagneticmeasurement (usually radar cross section)


with no target present. The backgroundincludes the collective unwanted power re-ceived from sources other than the desiredtarget under test such as positioners, foamcolumns, fixtures necessary to support a tar-get, and the room or ground environment.The background level is vectorially sub-tracted from the received level with the targetpresent to obtain the raw data set for a partic-ular target.

(2) any unwanted signal. The backgroundis a lower limit on the detection of small sig-nals when devices are used to make a mea-surement in an experimental set up. The mea-surement is a superposition of events from theexperiment itself and events from all othersources including the background.

background noise the noise that typicallyaffects a system but is produced independentof the system. This noise is typically due tothermal effects in materials, interpreted as therandom motion of electrons, and the intensitydepends on the temperature of the material.In radio channels, background noise is typi-cally due to radiation that is inherent to theuniverse and due mainly to radiation fromastronomical bodies. There is a fundamen-tal lower bound on the intensity of such noisewhich is solely dependent on the universe andindependent of antenna and receiver design.See alsothermal noise, noise temperature,noise figure.

background subtraction for images, theremoval of stationary parts of a scene by sub-tracting two images taken at different times.For 1-D functions, the subtraction of a con-stant or slowly varying component of thefunction to better reveal rapid changes.

backing memory the largest and slowestlevel of a hierarchical or virtual memory, usu-ally a disk. It is used to store bulky programsor data (or parts thereof) not needed imme-diately, and need not be placed in the fasterbut more expensive main memory or RAM.Migration of data between RAM and back-ing memory is under combined hardware and

software control, loading data to RAM whenit is needed and returning it to the backingstore when it has been unused for a while.

backing storage Seebacking memory.

backoff a technique used in amplifierswhen operated near saturation that reducesintermodulation products for multiple carri-ers. In its implementation, the drive signalis reduced or backed off. Input backoff isthe difference in decibels between the inputpower required for saturation and that em-ployed. Output backoff refers to the reduc-tion in output power relative to saturation.

backplane Seebackplane bus.

backplane bus a special data bus espe-cially designed for easy access by users andallowing the connection of user devices tothe computer. It is usually a row of sockets,each presenting all the signals of the bus, andeach with appropriate guides so that printedcircuit cards can be inserted. A backplanediffers from a motherboard in that a back-plane normally contains no significant logiccircuitry and a motherboard contains a sig-nificant amount of circuitry, for example, theprocessor and the main memory.

backplane optical interconnect Seeboard-to-board optical interconnect.

backprojection an operator associatedwith the Radon transform

g(s, θ) =∫ +inf ty

−inf ty

∫ +inf ty

−inf tyf (x, y)

δ(x cosθ + y sinθ − s) dx dy.

The backprojection operator is defined as

b(x, y) =∫ π

0g(x cosθ + y sinθ, θ) dθ.

b(x, y) is called the backprojection ofg(s, θ). b(x, y) is the sum of all rays thatpass through the point(x, y).


backpropagation the way in which errorterms are propagated in a multilayer neuralnetwork. In a single layer feedforward net-work, the weights are changed if there are dif-ferences between the computed outputs andthe training patterns. For multiple layer net-works, there are no training patterns for theoutputs of intermediate (‘hidden’) layer neu-rons. Hence the errors between the outputsand the training patterns are propagated tothe nodes of the intermediate neurons. Theamount of error that is propagated is propor-tional to the strength of the connection.

backpropagation algorithm a super-vised learning algorithm that uses a formof steepest descent to assign changes to theweights in a feedforward network so as to re-duce the network error for a particular inputor set of inputs. Calculation of the modifica-tions to be made to the weights in the outputlayer allows calculation of the required mod-ifications in the preceding layer, and modi-fications to any further preceding layers aremade a layer at a time proceeding backwardstoward the input layer; hence the name of thealgorithm.

backscatter energy from a reflected elec-tromagnetic wave. In optics, the optical en-ergy that is scattered in the reverse directionfrom the transmitted optical energy in an op-tical fiber transmission link or network. Thebackscattered energy comes from impuritiesin the fiber; mechanical or environmental ef-fects that cause changes in the attenuationin the fiber; connectors, splices, couplers,and other components inserted into the op-tical fiber network; and faults or breaks inthe optical fiber.

backscattering the reflection of a por-tion of an electromagnetic wave back inthe direction of the wave source.See alsobackscatter.

backside bus a term for a separate busfrom the processor to the second level cache

(as opposed to the frontside bus connectingto the main memory).

backward error recovery a technique oferror recovery (also called rollback) in whichthe system operation is resumed from a point,prior to error occurrence, for which the pro-cessing was backed up.

backward wave interaction interactionbetween backward propagating microwaveelectric fields against an electron stream andthe electron in the electron beam. The di-rection of propagating microwaves and thedirection of motion of electrons in the beamare opposite each other.

backward wave oscillator (BWO) a mi-crowave oscillator tube that is based on abackward wave interaction.

balanced Seebalanced line.

balanced amplifier an amplifier in whichtwo single-ended amplifiers are operated inparallel with 90 degree hybrid. Balanced am-plifiers feature a low voltage standing waveratio because of an absorption of reflectedpower at the terminating resistor of the hy-brids.

balanced code a binary line code thatensures an equal number of logic ones andlogic zeros in the encoded bit sequence. Alsocalled a DC-free code because the continuouscomponent of the power spectral density ofa balanced encoded sequence falls to zero atzero frequency.

balanced line symmetric multiconduc-tor transmission line in which the voltage oneach conductor along the transmission linehas the same magnitude, but the phases aresuch that the voltage would sum to zero. Ina two conductor transmission line, the volt-ages would be equal and 180 degrees out ofphase. This is the equivalent of a virtualground plane or zero E-field plane at the ge-ometric center plane of the transmission line


cross section, or balanced with respect to vir-tual ground. Balanced wiring configurationsare often used to prevent noise problems suchas ground loops.Contrast withunbalancedline.

balanced load a load on a multi-phasepower line in which each line conductor seesthe same impedance.

balanced mixer a nonlinear 3-port device(two inputs, one output) used to translate aninput signal’s frequency component either upor down the frequency spectrum by generat-ing the sum and difference of two or morefrequencies present at its inputs. The threeports are termed RF (radio frequency), LO(local oscillator), and IF (intermediate fre-quency). A balanced mixer translates thefrequency components found in the RF in-put signal to the IF output in such a manneras to minimize the amount of LO noise arriv-ing at the IF. This reduces the mixer’s over-all noise figure and increases its sensitivity.Other advantages of these mixers include im-proved local oscillator isolation and linearityand higher power handling ability.

balanced modulator a modulator inwhich the carrier and modulating signal areintroduced so that the output contains the twosidebands without the carrier.

balanced operation in n-phase circuits(n > 1), an operating condition in which thevoltages (currents) of the phases are equal-amplitude sinusoids with phase-angles dis-placed by a specific angleφ. The angle(φ)is a function of the number of phases (n). Forn = 2, φ = 90 degrees, forn = 3, φ = 120degrees. In machine analysis the term “bal-anced” is also used to describe a machine thathas symmetrical phase windings.

balanced slope detector an arrangementof two detectors designed to convert an FMsignal to AM for detection. This is accom-plished by setting the IF center frequencyso that it falls on the most linear portion

of the response curve. Frequency changes(FM) will result in corresponding amplitudechanges that are then sent to an AM detector.The balanced version is two slope detectorsconnected in parallel and 180 degrees out ofphase.

ball grid array (BGA) a modern highI/O count packaging method. It reduces thepackage size and its pin-to-pin trace gap inorder to integrate more functions and relia-bility in a single space. It can have as many as324 pins. BGA sockets are high speed, highreliability, surface-mountable, and can be in-stalled without soldering. The related termsare PBGA—plastic ball grid array, CBGA—ceramic ball grid array, TBGA—tape auto-mated bonded ball grid array. The disadvan-tage of BGA packaging is that new tools andskills are required to mount or replace thechipset manually for repair purposes.

ballast a starting and control mecha-nism for fluorescent and other types of gas-discharge lamps. Initially a ballast suppliesthe necessary starting (or striking) voltage inorder to ionize the gas to establish an arc be-tween the two filaments in the lamp. Once thegas is ionized, the ballast controls the inputpower and thus the light output to maximizethe efficiency and life of the lamp.

balun a network for the transformationfrom an unbalanced transmission line, sys-tem or device to a balanced line, system ordevice. Baluns are also used for impedancetransformation. Derived from “balanced tounbalanced.”

In antenna systems, baluns are used toconnect dipole-type antennas to coaxial ca-ble, to balance the current on dipole arma-tures, and to prevent currents from excitingthe external surface of the coaxial shield.

Seealso balanced, unbalanced line.

band reference name for a range of fre-quencies. Current defined bands include thefollowing.


Band Name Reference RangeL-band 1.12 - 1.7 GHzX 8.2 - 12.4 GHzKu 12.4 - 18 GHzKa 26.5 - 40 GHzV 50 - 75 GHzW 75 - 110 GHz

band gap the energetic gap between theconduction and valence band edges of a ma-terial (usually referred to semiconductors).

band stop filter filter that exhibits fre-quency selective characteristic such that fre-quency components of an input signals passthrough unattenuated from input to output ex-cept for those frequency components coinci-dent with the filter stop-band region, whichare attenuated. The stop-band region of thefilter is defined as a frequency interval overwhich frequency components of the inputsignal are attenuated.

band structure the energy versus mo-mentum relationship for an electron in a pe-riodic crystal.

band-limited signal a signalx(t) is said tobe band limited if its Fourier transformX(ω)is zero for all frequenciesω > ωc, whereωcis called the cutoff frequency.

band-pass filter (1) a circuit whose trans-fer function, or frequency response,H(ω) iszero or is very small for frequencies not in aspecified frequency band. In a strict senseH(ω) = 0 for |ω| 6∈ [ω1, ω2] for some0 < ω1 < ω2. Compare withlow-passfilter, high-passfilter, notch,band-reject filter.

(2) an electronic or electrical circuit whichhas the response shown in the figure below.There are two cut-off frequencies,ωL andωH. In the passbandωL < ω < ωH , | N(jω) |is constant. In the stopbands,ω < ωL andω > ωH , | N(jω) | is very small and thereis practically no transmission of the signal.

Band-pass filter response.

band-pass network a configuration ofsolely passive components or combination ofactive and passive components that will at-tenuate all signals outside of the desired rangeof frequency.

band-pass signal a signal whose Fouriertransform or spectrum approaches zero out-side a given frequency band. Ideally, thespectrum should equal zero outside the band,but this is difficult to achieve in practice. Thismay be described mathematically as follows:letX(ω) be the Fourier transform of the sig-nal. Then, for a band-pass signal we haveX(ω) = 0 for |ω| 6∈ [ω1, ω2], for some0< ω1 < ω2.

band-reject filter Seeband-stop filter.

band-stop filter a filter which attenu-ates only within a finite frequency band andleaves the signal unaffected outside this band.See alsoband-pass filter, high-pass filter,low-pass filter.

bandgap energy in materials with bandenergy levels, the minimum energy neededto excite a charge carrier from a lower to anupper band.See alsoabsorption edge.

bandgap engineering in materials suchas compound semiconductors and superlat-tice structures, the fabrication of materialswith specific bandgap energies by varying thefractional proportions of the constituents andby varying superlattice layer thicknesses.

bandgap narrowing reduction of the for-bidden energy gap of a semiconducting ma-terial due to the narrowing influence of im-purities.


bandgap reference a voltage referencebased on the 1.205 V bandgap voltage of sil-icon.

bandgap wavelength the optical wave-length corresponding to a photon energyequal to the bandgap energy.

bandlimited a waveform is described asbandlimited if the frequency content of thesignal is constrained to lie within a finite bandof frequencies. This band is often describedby an upper limit, the Nyquist frequency, as-suming frequencies from DC up to his upperlimit may be present. This concept can be ex-tended to frequency bands that do not includeDC.

bandwidth (1) the frequency range ofa message or information processing systemmeasured in hertz.

(2) width of the spectral region over whichan amplifier (or absorber) has substantialgain (or loss); sometimes represented morespecifically as, for example, full width at halfmaximum.

(3) the property of a control system orcomponent describing the limits of sinu-soidal input frequencies to which the sys-tem/component will respond. It is usuallymeasured at the half-power points, which arethe upper and lower frequencies at which theoutput power is reduced by one half. Band-width is one measure of the frequency re-sponse of a system, i.e., the manner in whichit performs when sine waves are applied tothe input.

(4) the lowest frequency at which the ra-tio of the output power to the input powerof an optical fiber transmission system de-creases by one half (3 dB) compared to theratio measured at approximately zero mod-ulation frequency of the input optical powersource. Since signal distortion in an opti-cal fiber increases with distance in an opti-cal fiber, the bandwidth is also a function oflength and is usually given as the bandwidth-distance product for the optical fiber in mega-

hertzper kilometer. Seealsobandwidth-distanceproduct.

bandwidth efficiency the ratio of theinformation rate in bits per second to therequired bandwidth in hertz for any digitalmodulation technique.

bandwidth improvement (BI) a dB read-ing that is a comparison of the RF bandwidthof a receiver to the IF bandwidth. Designatedas BI, it is 10 logBrf /Bif .

bandwidth-distance product a measureof the information carrying capacity of an op-tical fiber which emphasizes that the band-width is a function of distance. For example,an optical fiber with a specification of 500MHz-km bandwidth-distance product wouldhave a 500 MHz bandwidth over 1 km, a50 MHz bandwidth over 10 km or a 1 GHzbandwidth over 0.5 km.See alsobandwidth,optical fiber.

bang-bang control control actionachieved by a command to the actuator thattells it to operate in either one direction or theother at any time with maximum energy.

Bang-bang control is an optimal or subop-timal piecewise constant control whose val-ues are defined by bounds imposed on the am-plitude of control components. The controlchanges its values according to the switchingfunction which may be found using Pontrya-gin maximum principle. The discontinuityof the bang-bang control leads to disconti-nuity of a value function for the consideredoptimal control problem. Typical problemswith bang-bang optimal control include time-optimal control for linear and bilinear controlsystems.

Bardeen, John (1908–1991) Born:Madison, Wisconsin, U.S.A.

Bardeen is best known for his two NobelPrizes. The first prize he received was in1956 for his development at Bell Labs, alongwith Walter Brattain and William Schockley,of the first transistor. When the three applied


for a patent for the device in 1948 they calledit a germanium transfer resistance unit; hencethe name transistor. This device was a sig-nificant step in the development of integratedcircuits. Bardeen’s second Nobel, which heshared with Leon Cooper and John R. Schri-effer, was for his work at the University ofIllinois in describing the theory of supercon-ductivity.

bare-hand refers to a method of servic-ing energized overhead conductors in whichthe line worker’s body is maintained at thesame potential as the conductor on which heis working, thus enabling the conductor to becontacted without danger of shock.

BARITT barrier injection transit time, amicrowave transit time device that uses injec-tion over a forward biased barrier and transittime delay through a reverse biased junctionto produce negative resistance at microwavefrequencies, useful in low power and self-oscillating mixer applications.

Barkhausen criterion two conditionsplaced on a feedback oscillator necessary forsustained oscillation. The Barkhausen crite-rion states

1. The total loop amplitude transmissionfactor must be at least unity.

2.The frequency of oscillation will be thatfrequency characterized by a total loop phasetransmission factor ofN2π radians.N is ei-ther zero or an integer. Simply, for sustainedoscillation, a disturbance that makes a com-plete trip around the feedback loop of the os-cillator must be returned at least as strongas the original disturbance and in phase withthat disturbance.

Barkhausen effect the series of irregularchanges in magnetization that occur when amagnetic material is subjected to a change inmagnetizing force.

Barkhausen noise noise arising in mag-netic read heads because the interlockingmagnetic domains cannot rotate freely in re-

sponse to an applied field. The response to anexternal magnetic field is randomly discon-tinuous as domains “stick,” and then release.Barkhausen noise is particularly important invery small heads and thin-film heads wherevery few domains are involved; in largerheads the effects of many domains tend to av-erage out and Barkhausen noise is relativelyless important.

barrel distortion a geometric distortionof a raster display in which vertical lines ap-pear to bow outward away from the displaycenter line. The bowing of the vertical linesincreases as the distance from the verticalcenter increases. The appearance of thesevertical lines is similar to the staffs of a bar-rel. Barrel distortion is a result of the over-correction for pincushion distortion.

barrel shifter an implementation of ashifter, which contains log2(max number ofbits shifted) stages, where each stage shiftsthe input by a different power of two numberof positions. It can be implemented as a com-binational array with compact layout that canshift the data by more than one bit using onlyone gate. For instance, for a 4-bit word, itcan execute instructions shl, shl2, shl3, andshl4. This shifter lends itself well to beingpipelined.

barrier layer layer of deposited glassadjacent to the inner tube surface to create abarrier against OH diffusion.

barrier voltage a voltage that developsacross the junction due to uncovered immo-bile ions on both sides of the junction. Ionsare uncovered due to the diffusion of carriersacross the junction.

Bartlett window a triangular windoww[n] of width 2M defined as follows:

w[n] ={1/2[1 + cos(πn/M)], −M ≤ n ≤ M

0, otherwise


Multiplying a signalx[n] by the finite dura-tion window signalw[n] leads to the trian-gularly scaled, finite duration signalz[n] =x[n]w[n], which is then processed. Window-ing is used in the spectral analysis of mea-sured signals and the design of finite impulseresponse, linear time invariant systems.

baryon a collective term for all stronglyinteracting particles with masses greater thanor equal to the mass of the proton. Theseinclude the proton, neutron, and hyperons.

base (1) the number of digits in a numbersystem (10 for decimal, 2 for binary).

(2) one of the three terminals of a bipolartransistor.

(3) a register’s value that is added to animmediate value or to the value in an indexregister in order to form the effective addressfor an instruction such as LOAD or STORE.

base address (1) an address to whichan index or displacement is added to locatethe desired information. The base addressmay be the start of an array or data structure,the start of a data buffer, the start of page inmemory, etc.

(2) as a simpler alternative to a full vir-tual memory, the code space or data space ofa program can be assumed to start at a con-venient starting address (usually 0) and relo-cated in its entirety into a continuous rangeof physical memory addresses. Translationof the addresses is performed by adding thecontents of an appropriate base address reg-ister to the user address.

base dynamic parameters a set of dy-namic parameters that appear in the canon-ical equations of motion. Canonical equa-tions of motion of robot dynamics do not in-clude linearly dependent equations. Theseare eliminated by making use of various pro-cedures. As a result, dynamic equations ofmotion contain only independent equationswhich are used for the purpose of control.Each base dynamic parameter is a linear com-bination of the inertial parameters of the in-

dividual links. Base dynamic parameters aresubject to the identification in adaptive con-trol schemes applied to robot control.

base frame a frame attached to the non-moving base of the manipulator. Sometimesthe base frame is called the reference frame.

base quantity Seeper-unit system.

base register the register that con-tains the component of a calculated addressthat exists in a register before the calcu-lation is performed (the register value in“register+immediate” addressing mode, forexample).

base register addressing addressing us-ing the base register. Base register is thesame as base address register, i.e., a general-purpose register that the programmer choosesto contain a base address.

base speed corresponds to speed at ratedtorque, rated current, and rated voltage condi-tions at the temperature rise specified in therating. It is the maximum speed at whicha motor can operate under constant torquecharacteristics or the minimum speed to op-erate at rated power.

base station the fixed transceiver in a mo-bile communication system.See alsofixedstation (FS).

base vector a unit vector in a coordinatedirection.

baseband in communication systems, theinformation-carrying signal that is modu-lated onto a carrier for transmission.

baseband signal in digital communica-tions, a signal that appears in the transmitterprior to passband modulation. For example,in the case of pulse amplitude modulation,s(t) = ∑

i bip(t − iT ) is a baseband signal,wherebi is the transmitted symbol at timei,


andp(t) is the baseband pulse shape (e.g.,raised-cosine).See alsolow-pass signal.

basic impulse insulation level (BIL) ameasurement of the impulse withstand capa-bility of a piece of electric power equipmentbased on its ability to withstand 50% of im-pulses applied at the BIL voltage.

basic input–output system (BIOS) partof a low-level operating system that directlycontrols input and output devices.

basic lightning impulse level (BIL) thestrength of insulation in terms of the with-stand voltage crest value using a standardvoltage level impulse.

basin of attraction the region in statespace from which a dynamical system movesasymptotically toward a particular attractor.

basis function one of a set of functionsused in the transformation or representationof some function of interest. A linear trans-formationT of continuous functions is of theform

y(s) = T {x(t)} =∫ +inf ty

−inf tyx(t)b(s, t) dt.

whereb(s, t) is a basis function. For discretesequencesT would be of the form

y[k] = T {x[n]} =+inf ty∑n=−inf ty

x[n]b[k, n].

The function to be transformed is projectedonto the basis function corresponding to thespecified value of the index variables or k.y(s) is the inner product ofx(t) and the basisfunction b(s, t). For the Laplace transformb(s, t) = e−st , and for the Fourier trans-form b(ω, t) = e−jωt . For the discrete-time

Fourier transformb[k, n] = ej2πknN , and for

the Z-transformb[z, n] = z−n.

BaTiO3 (barium titanate) a ferroelectriccrystalline material that is particularly use-

ful for photorefractive and optical multibeamcoupling.

battery one or more cells connected soas to produce energy.

baud the signaling rate, or rate of statetransitions, on a communications medium.One baud corresponds to one transition persecond. It is often confused with the datatransmission rate, measured in bits per sec-ond.

Numerically, it is the reciprocal of thelength (in seconds) of the shortest elementin a signaling code. For very low-speedmodems (up to 1200 bit/s) the baud rate andbit rate are usually identical. For example, at9600 baud, each bit has a duration of 1/9600seconds, or about 0.104 milliseconds.

Modems operating over analog telephonecircuits are bandwidth limited to about 2500baud; for higher user data speeds each tran-sition must establish one or more decod-able states according to amplitude or phasechanges. Thus, if there are 16 possible states,each can encode 4 bits of user data and thebit rate is 4 times the baud rate.

At high speeds, the reverse is true, withrun-length controlled codes needed to en-sure reliable reception and clock recovery.For example FDDI uses a 4B/5B coding inwhich a “nibble” of 4 data bits is encodedinto 5 bits for transmission. A user data rateof 100 Mbit/s corresponds to transmission at125 Mbaud.

baud rate Seebaud.

Baum–Welch algorithm the algorithmused to learn from examples the parametersof hidden Markov models. It is a special formof the EM algorithm.

Bayes envelope function given a the priordistribution of a parameter2 and a deci-sion functionφ, the Bayes envelope functionρ(F2) is defined as

ρ(F2) = minφr(F2, φ),


where r(F2, φ) is the Bayes risk functionevaluated with the prior distribution of theparameter2 and decision ruleφ.

Bayes risk function with respect to a priordistribution of a parameter2 and a decisionruleφ, the expected value of the loss functionwith respect to the prior distribution of theparameter and the observationX.

r(F2, φ) =∫2

∫X

L[θ, φ(x)]fX|2(x|θ)f|2(θ) dx dθ.

the loss function is the penalty incurred forestimating the parameter2 incorrectly. Thedecision ruleφ(x) is the estimated value ofthe parameter based on the measured obser-vationx.

Bayes’ rule Bayes’ rule relates the con-ditional probability of an eventA given Band the conditional probability of the eventB givenA:

P(A|B) = P(B|A)P (A)P (B)

.

Bayesian classifier a Bayesian classifieris a function of a realization of an observedrandom vectorX and returns a classificationw. The set of possible classes is finite. ABayesian classifier requires the conditionaldistribution function ofX givenw and theprior probabilities of each class. A Bayesianclassifier returns thewi such thatP(wi |X) ismaximized. By Bayes’ rule

P(wi |X) = f racP (X|wi)P (wi)P (X).SinceP(X) is the same for all classes, itcan be ignored and thewi that maximizesP(X|wi)P (wi) is returned as the classifica-tion.

Bayesian detector a detector that min-imizes the average of the false-alarm andmiss probabilities, weighted with respectto prior probabilities of signal-absent andsignal-present conditions.

Bayesian estimation an estimationscheme in which the parameter to be esti-mated is modeled as a random variable withknown probability density function. SeeBayesian estimator.

Bayesian estimator an estimator of agiven parameter2, where it is assumed that2 has a known distribution function and arelated random variableX that is called theobservation.X and2 are related by a con-ditional distribution function ofX given2.With P(X|2) andP(2) known, an estimateof 2 is made based on an observation ofX.P(2) is known as thea priori distribution of2.

Bayesian mean square estimator for arandom variableX and an observationY , therandom variable

X = E[X|Y ],where the joint density functionfXY (x, y)is known. See alsomean-square estimation,linear least squares estimator.

Bayesian reconstruction an algorithm inwhich an imageu is to be reconstructed froma noise-corrupted and blurred versionv.

v = f (Hu)+ η.

A prior distributionp(u|v)of the original im-age is assumed to be known. The equation

u = µu + RuHTDR−1

η [v − f (Hu)],whereRu is the covariance of the imageu,Rη is the covariance of the noiseη, andD isthe diagonal matrix of partial derivatives off evaluated atu. An initial point is chosenand a gradient descent algorithm is used tofind the closestu that minimizes the error.Simulated annealing is often used to avoidlocal minima.

Bayesian theory theory based on Bayes’rule, which allows one to relate thea priorianda posterioriprobabilities. IfP(ci) is thea priori probability that a pattern belongs to


classci , P(xk) is the probability of patternxk , P(xk |ci) is the class conditional proba-bility that the pattern isxk provided that itbelongs to classci , P(ci |xk) is the a pos-teriori conditional probability that the givenpattern class membership isci , given patternxk , then

P (ci |xk) = P (xk |ci) P (ci)P (xk)

.

The membership of the given pattern is de-termined by

maxciP (ci |xk) = max

ciP (xk |ci) P (ci) .

Hence, thea posterioriprobability can bedetermined as a function of thea priori prob-ability.

BCD Seebinary-coded decimal.

BCH code cyclic block forward errorcontrol codes developed by Bose and Chaud-huri, and independently by Hocquenghem.These codes are a superset of the Hammingcodes, and allow for correction of multipleerrors.

BCLA Seeblock carry lookahead adder.

beam (1) transverse spatial localizationof the power in a wave field.

(2) a slender unidirectional stream of par-ticles or radiation.

beam cooling the process by which a par-ticle beam’s phase space volume is reduced,while conserving Liouville’s theorem (emptyspaces between particles exist). Beam cool-ing is manifest by a reduction in the trans-verse beam size (betatron cooling) or by asmaller momentum spread (momentum cool-ing).

beam divergence the geometric spread-ing of a radiated electromagnetic beam as ittravels through space.

beam hardening a phenomenon thatoccurs when a polychromatic X-ray beampasses through a material. Lower energyphotons are absorbed more readily thanhigher energy photons, increasing the ef-fective energy of the beam as it propagatesthrough the material.

beam intensity the average number ofparticles in a beam passing a given point dur-ing a certain time interval. For example thenumber of protons (electrons) per pulse orprotons per second.

beam loading the beam being acceleratedby an RF cavity and it changes the gradientand phase of the RF in the cavity.

beam mode confined electromagneticfield distributions of a propagating wave thatmatch the boundary conditions imposed bya laser or aperture. For example, Hermite–Gaussian or Laguerre–Gaussian.

beam parameter one of several complexnumbers employed to characterize the prop-agation of a beam; most common parametercombines in its real and imaginary parts thephase front curvature and spot size of a Gaus-sian beam.

beam pulsing a method used to controlthe power output of a klystron in order to im-prove the operating efficiency of the device.

beam roll a periodic change in horizontaland/or vertical positions during spill. Thisdoes not include changes caused by humans.

beam solid angle a parameter that quali-tatively describes the angular distribution ofradiated power from an antenna. The val-ues range from very small numbers for veryfocused antennas to 4π steradians for anisotropic radiator.

beam stop a thick metal shield that movesinto the beam line to prevent beam from en-tering a specific area.


beam toroid a device used for measur-ing beam intensities by measuring the mag-netic field fluctuations produced by the pass-ing beam. The magnetic field fluctuationsproduce a current in a coil, that is woundaround a closed circular ring (torus) throughwhich the beam passes.

beam waist position at which a beam ismost highly confined; for Gaussian beams inreal media the position at which the phasefronts are flat.

beamformers system commonly used fordetecting and isolating signals that are prop-agating in a particular direction.

beamforming a form of filtering in spatialrather than time domain to obtain a desiredspatial impulse response in order to suppressor to reject signal components coming fromcertain directions. The technique involvesdirecting one or more beams in certain di-rections by adjusting, for example, the ele-ment excitation of an array antenna. Usedin communications applications to suppressother signals than the desired source signal.Also termed spatial filtering.

beamline a series of magnets placedaround a vacuum pipe that carry the protonbeam from one portion of the accelerator toanother. Also known as transport line.

beamsplitter any of a number of passiveoptical devices that divide an optical wave-front into two parts. Wavefront division maybe according to intensity, polarization, wave-length, spatial position, or other optical prop-erties.

beamwidth the angular width of the ma-jor lobe of a radiation pattern. It is usuallyat the half-power level, i.e., 3 dB below thepeak of the major lobe. It can also be speci-fied as the width between the nulls on eitherside of the major lobe (BWFN).

bearing currents current flow in thebearings of electrical machines, because ofelectromagnetic unbalance in the machineor from using highdv

dtinverters. The lat-

ter is able to charge up the stray capacitancepresent between the stator and rotor and be-tween the rotor and shaft and thus allows mo-tor bearing currents to flow, with resultingbearing damage.

beat frequencies the two frequencies,sum and difference frequencies, generatedduring the heterodyning process or during theamplitude-modulating process. For exam-ple, if a 500 kHz carrier signal is amplitude-modulated with a 1 kHz frequency, the beatfrequencies are 499 kHz and 501 kHz.

beat frequency oscillator an adjustableoscillator used in superheterodyne receiversgenerating a frequency when combined withthe final IF produces a difference or beat fre-quency in audio range.

becky a knot used to secure a handline.

bed of nails a test fixture for automatedcircuit qualification in which a printed wiringboard is placed in contact with a fixture thatcontacts the board at certain nodes requiredfor exercising the assembly.

bel Seedecibel.

Bell, Alexander Graham (1847–1922)Born: Edinburgh, Scotland

Bell is best know as the first patent holderfor a device to electronically transmit humanspeech. Bell’s early interest in the mecha-nisms of speech come from living with hisgrandfather, a London speech tutor. Workwith the deaf was to be a lifelong vocationfor Bell. Bell’s inventions were not limitedto the telephone. He was the first person totransmit speech without wires, he inventedthe gramophone, an early tape recorder, anair-cooling system, an iron lung, and he hadseveral patents in telegraphy.


bell insulator a type of strain insulator,shaped like saucer with ribs on its lower sideand frequently used in insulator strings.

Bello functions a group of alternativemethods of characterizing a wideband com-munication channel, named after their pro-poser, P. Bello. The four functions charac-terizing deterministic channels are the InputDelay-spread Function, the Output Doppler-spread function, the Time-variant TransferFunction and the Delay Doppler-spread func-tion.

BEM Seeboundary-element method.

benchmark standard tests that are usedto compare the performance of computers,processors, circuits, or algorithms.

bending loss in a fiber depends exponen-tially on the bend radiusR. It is proportionalto exp(−R/Rc) where the critical radius

Rc = a

2n (nco − ncl),

a is the fiber radius,nco is the refractive indexof the core, andncl is the refractive index ofthe cladding.

BER Seebit error rate.

Bernoulli distribution a random variableX with alphabet{0,1} and parameterα suchthat its probability mass function is

p(x) = (1 − α)xα1−x.

Bernoulli process a binary valued,discrete-time random process defined on anindex set corresponding to fixed incrementsin time. A typical example is a sequence ofcoin tosses where the values of the processare denoted as “Heads” or “Tails” dependingon the outcome of the tosses. The output val-ues of the process is a sequence of statisticallyindependent random variables with the sameprobability distribution. The two outcomesmay or may not have equal probabilities.

Berry, Clifford Edward Berry is bestknown as the co-developer, along with JohnVincent Atanasoff, of the first functioningelectronic digital computer. Berry was rec-ommended to Atanasoff by the Dean of En-gineering at Iowa State College as a mostpromising student who understood the elec-tronics well enough to help Atanasoff imple-ment his ideas for a computing machine. Un-fortunately, Berry’s contributions as a com-puting pioneer were not honored until afterhis death.

beryllium oxide a compound commonlyused in the production of ceramics for elec-trical applications and whose dust or fumesare toxic.

Bessel beam transverse wave ampli-tude distribution in which the radial varia-tion is approximately describable in terms oftruncated Bessel functions; collimation forBessel beams is sometimes considered bet-ter than for more usual polynomial-Gaussianbeams.

Bessel functions a collection of func-tions, denoted asJν(x)andYν(x), that satisfyBessel’s equation

x2d2f

dx2+ x

df

dx+(x2 − ν2

)f = 0 ,

wheref is equal to eitherJν or Yν ; ν is theorder of the function andx is its argument.Typically, Bessel functions arise in boundaryvalue problems that are based upon a cylin-drical coordinate system.

best-fit memory allocation a mem-ory allocator for variable-size segments mustsearch a table of available free spaces to findmemory space for a segment. In “best-fit” al-location, the free spaces are linked in increas-ing size and the search stops at the smallestspace of sufficient size.Compare withbuddymemory allocation.

beta function a measure of beam width.The beta function details how the beam


changes around the accelerator. There areseparate beta functions for thex andy planes.The square root ofbx is proportional to thebeam’sx-axis extent in phase space.

beta particle an electron or positron emit-ted from a radioactive source.

betatron oscillation stable oscillationsabout the equilibrium orbit in the horizontaland vertical planes. First studied in betatronoscillators, betatron oscillation is the trans-verse oscillation of particles in a circular ac-celerator about the equilibrium orbit. Therestoring force for the oscillation is providedby focusing components in the magnetic fieldthat act to bend a particle that is off the equi-librium orbit back toward it.

Beverage antenna simple traveling waveantenna consisting of an electrically long hor-izontal wire above ground with a terminationresistance between the end of the wire andground equal to the characteristic impedanceof the wire/ground transmission line.

Bezout identity of 2-D polynomial matri-ces a systems identity defined as follows:letNR(z1, z2),DR(z1, z2) (NL(z1, z2),DL(z1, z2)) be two right (left) coprime poly-nomial matrices, then there exists a polyno-mial matrix inz2, sayER(z2), (EL(z2)) andtwo polynomial matricesXR(z1, z2),YR(z1, z2) (XL(z1, z2),YL(z1, z2)) such that

XR (z1, z2)DR (z1, z2)

+ YR (z1, z2)NR (z1, z2) = ER (z2)

(NL (z1, z2) YL (z1, z2)

+ DL (z1, z2)XL (z1, z2) = EL (z2))

BGA Seeball grid array.

BI Seebandwidth improvement.

bi-anisotropic media (1) a class of ma-terial in which the electric and magnetic fluxdensities,D andB, are each linearly relatedto both the electric and magnetic field inten-

sities,E andH, via dyadic constitutive pa-rameters. The permittivity, permeability, andmagnetoelectric coupling parameters are ten-sor quantities.

(2) media for which the electric and mag-netic fields displacements,D andB respec-tively, are related to the electric and magneticfield strengthE andH by general dyadics.

bi-isotropic media media for which theelectric and magnetic fields displacements,DandB, respectively, are scalarly dependent byboth the electric and magnetic field strengthE and H. For these media the constitutiverelations are

D = εE + (χ − jκ)√µ0ε0 H

B = εH + (χ + jκ)√µ0ε0 E

whereε is the permittivity,µ the permeabil-ity, and the subscript 0 refers to free-space.Bi-isotropic media can be reciprocal (χ = 0)or nonreciprocal (χ 6= 0); nonchiral (κ = 0)or chiral (κ 6= 0).

bi-stable pertaining to a device with twostable states, e.g., bi-stable multivibrator; cir-cuit that has two possible output states andthat will remain in its current state withoutrequiring external inputs; a flip-flop.

bi-stable device Seeflip-flop.

bias the systematic (as opposed to ran-dom) error of an estimator.

bias current the arithmetic average ofthe currents that flow in the input leads of anop-amp.

bias lighting technique used in videotubes to correct for undesirable artifacts suchas lag. Applying a uniform light source tothe surface of the tube (the photoconductivelayer) will create a bias current in the tube,thereby minimizing the undesirable charac-teristics.

bias network a key aspect of microwavecircuit design is to apply the proper DC bias


to the appropriate terminals of transistors(e.g., FETs) without disturbing the AC mi-crowave operation of the circuit. In somecases, “on-chip” DC circuitry needs to bedesigned so as to provide stable bias volt-age/current conditions for the device evenwhen the chip DC supply voltages vary (dueto weakening batteries, etc.). The other as-pect of bias network design is to isolate theDC network from interfering with the AC orRF/microwave operation of the circuit, andvice-versa. In a lumped element design, thisis generally accomplished by a combinationof spiral inductors and MIM capacitors.

bias voltage or current the DC powerapplied to a transistor allowing it to operateas an active amplifying or signal generatingdevice. Typical voltage levels in GaAs FETsused in receivers are 1 to 7 volts between thedrain and source terminals, and 0 to−5 voltson, or between, the gate and source termi-nals. For microwave systems, DC voltagesand currents, provided by batteries or AC/DCconverters required to “bias” transistors toa region of operation where they will eitheramplify, mix or frequency translate, or gen-erate (oscillators) microwave energy. Sinceenergy can be neither created nor destroyed,microwave energy amplification or creationis accomplished at the expense of DC energy.

biasing the technique of applying a direct-current voltage to a transistor or an active net-work to establish the desired operating point.

bible nickname for the National ElectricalCode.

BIBO stability See bounded-inputbounded-output stability.

BIBO stability of 2-D linear system asystem described by the equation

yi,j =i∑

k=0

j∑l=0

gi−k,j−luk,l

i, j ∈ Z+ (the set of nonnegative integers)is said to be bounded-input bounded-output(BIBO) stable if for every constantM > 0there exist a constantN > 0 such that if‖uk,l‖ ≤ M for all k, l ∈ Z+, then‖yi,j‖ ≤N for all i, j ∈ Z+ whereuk,l ∈ Rm is theinput, yi,j ∈ Rp is the output,gi,j ∈ Rp×mis the matrix impulse response of the systemand‖v‖ denotes a norm of the vectorv. Thesystem is BIBO stable if and only if

∞∑i=0

∞∑j=0

∥∥gi,j∥∥ < ∞

BIBS Seebounded-input bounded-statestability.

BiCMOS integrated circuit technol-ogy/process that incorporates bipolar andcomplementary metal oxide semiconductordevices on the same die.

bicycle a chain-driven drill for boringholes.

bidirectional bus a bus that may carryinformation in either direction but not in bothsimultaneously.

bi-directional laser a ring laser with bothclockwise and counter-clockwise circulatingwaves. Useful as a rotation rate sensor.

bidirectional pattern a microphonepickup pattern resembling a figure eight, inwhich the device is most sensitive to soundson either side of the pickup element.

bi-directional resonator a standing-wave resonator or a ring-resonator in whichthe electromagnetic waves circulate in boththe clockwise and counter-clockwise direc-tions.

bidirectional transducer a surfaceacoustic wave (SAW) transducer whichlaunches energy from both acoustic portswhich are located at either end of the trans-ducer structure.


bidirectional transmission distributionfunction (BTDF) the optical scattering func-tion for transmissive optics. The scatteringfunction vs. angle is normalized to signal atzero degrees and with respect to solid angleof detector, including obliquity factor.

bifilar winding a two-wire winding. Itis often utilized in stepper motors to permit aunipolar power supply to produce alternatingmagnetic poles by energizing only half of thebifilar winding at any one time.

bifurcation a term from Chaos Theoryreferring to a sudden change in the qualitativebehavior of the solutions.

bifurcation diagram a diagram wherethe sampled variable is plotted versus a pa-rameter. The sampling period is equal to thesource period. Similar to a Poincare map.

big endian a storage scheme in which themost significant unit of data or an address isstored at the lowest memory address. Forexample, in a 32-bit, or four-byte word inmemory, the most significant byte would beassigned addressi, and the subsequent byteswould be assigned the addresses:i+1, i+2,and i + 3. Thus, the least significant bytewould have the highest address ofi + 3 ina computer implementing the big endian ad-dress assignment. “Big endian” computersinclude IBM 360, MIPS R2000, MotorolaM68000, SPARC, and their successors.

The little endian approach stores the leastsignificant unit at the lowest address. (Theterms big endian and little endian are takenfrom Jonathan Swift’s satirical story,Gul-liver’s Travels.

See alsolittle endian.

BIL Seebasic lightning impulse levelandbasic impulse insulation level.

bilateral Laplace transform a Laplacetransform of the form

L{f } =∫

−∞+∞f (t)e−st dt,

wheres is a complex number.SeeLaplacetransform.

bilateral Z-transform a Z-transform ofthe form

Zx =+∞∑n=−∞

x[n]z−n.

bilinear control systems a class of non-linear control system models that are linearin state, output and control variables treatedseparately but they contain the products ofthose variables. Such models arose naturallyin modeling the number of chemical pro-cesses where the controls are flow rates thatappear in the system equations as productswith state variables. The bilinear control sys-tems may also be used to model populationdynamics perturbed by control actions whichenter growth equations as multipliers of statevariables. Bilinear control systems can arisealso in connection with adaptive control nom-inally linear systems where uncertain pa-rameters regarded as additional state vari-ables leads to bilinear terms in model equa-tions. Bilinear time-continuous control sys-tems may be represented by the state equa-tions having the form

x = Ax + Bu+m∑i=1

Diuix

wherex is the state vector,u the control vec-tor with componentsui, i = 1,2, . . . , m,A,B,Di are matrices of the appropriate di-mensions.See alsopopulation dynamics.

bilinear interpolation interpolation of avalue in 2-D space from four surrounding val-ues by fitting a hyperbolic paraboloid. Thevalue at(x, y), denotedf (x, y) is interpo-lated usingf (x, y) = ax + by + cxy + d,wherea, b, c andd are obtained by substitut-ing the four surrounding locations and valuesinto the same formula and solving the systemof four simultaneous equations so formed.


bilinear transformation (1) conformalmapping of the complex plane of the formf (z) = az+b

cz+d , where the real valuesa, b,c, d satisfyad − bc 6= 0. Also called linearfractional transformation or Mobius transfor-mation.

(2) a special case of (1) is a mapping fromthe jω axis in thes-plane to the unit circle

|z| = 1 in thez- plane, given byx = 2T

1−z−1

1+z−1 ,whereT is the time interval between samples.

Such bilinear transformations are used inthe design of recursive digital filters fromequivalent analogue filters in the followingprocedure:1. define characteristic digital frequencies�i .2.prewarp these to analog frequenciesωi us-ingωi = 2

Ttan�iT2 1 ≤ i ≤ k.

3.design a suitable analog filter with frequen-ciesωi .4. use the bilinear transformation to replace

s in the analog filter withs = 2T

1−z−1

1+z−1 .

bimetal overload device an overload de-vice that employs a bimetal strip as the actu-ating element. The bimetal strip consists oftwo metals bonded together. When heated,the bimetal strip will bend due to the differ-ent coefficients of linear expansion of the twometals. The bending operates a set of con-tacts that automatically removes the affectedload from the source of electrical power.Seealsooverload heater, overload relay.

bimodal histogram a histogram withtwo main groupings of values, such as thesum of two displaced Gaussians.See alsohistogram.

binary (1) a signal or other informationitem that has two possible states.

(2) representation of quantities in base 2.

binary code a code, usually for error con-trol, in which the fundamental informationsymbols which the codewords consist of aretwo-valued or binary and these symbols areusually denoted by either “1” or “0,” Mathe-

matical operations for such codes are definedover the finite or Galois field consisting oftwo elements denoted by GF(2). The math-ematical operations for such a Galois fieldare addition and multiplication. For additionover GF(2) one finds that

1 + 1 = 0, 1 + 0 = 1 and 0+ 0 = 0.

For multiplication over GF(2) one finds that

1 · 1 = 1, 1 · 0 = 0 and 0· 0 = 0.

See alsoblock coding, convolutional coding,error control coding.

binary-coded decimal (BCD) (1) aweighted code using patterns of four bits torepresent each decimal position of a number.

(2) decimal digits 0 to 9, encoded by theirfour-bit binary representation. Thus: 0 =0000, 1 = 0001, 2 = 0010, 3 = 0011, 4 = 0100,5 = 0101, 6 = 0110, 7 = 0111, 8 = 1000, 9 =1001.

binary erase channel a channel wherean error detecting circuit is used and the er-roneous data is rejected as erasure asking forretransmission. The inputs are binary andthe outputs are ternary, i.e., 0,1 and erasure.Used for ARQ (automatic request for retrans-mission) type data communication.

binary hypothesis testing a special two-hypothesis case of the M-ary hypothesistesting problem. The problem is to as-sess the relative likelihoods of two hypothe-sesH1, H2, normally given prior statisticsP(H1), P (H2), and given observationsywhose dependencep(y|H1), p(y|H2) on thehypotheses is known. The receiver operat-ing characteristic is an effective means tovisualize the possible decision rules.Seealso m-ary hypothesis testing. See alsoconditional statistic, prior statistics, pos-terior statistics.

binary image an image whose pixels canhave only two values, 0 or 1 (i.e., “off” or


“on”). The set of pixels having value 1 (“on”)is called thefigure or foreground, while theset of pixels having value 0 (“off”) is calledthe background.

binary image coding compression oftwo-level (black/white) images, typicallydocuments. Bilevel coding is usually loss-less and exploits spatial homogeneity by run-length, relative address, quadtree, or chaincoding. Also called bilevel image coding.

binary notation Seebinary.

binary operator any mathematical op-erator that requires two data elements withwhich to perform the operation. Additionand Logical-AND are examples of binaryoperators; in contrast, negative signs andLogical-NOT are examples of unary opera-tors.

binary optics optical filters constructedwith only two amplitude or two phase val-ues to perform the functions of bulk opticalcomponents such as lenses.

binary phase frequency modulationconverting signals from a binary-digit pat-tern [pulse form] to a continuous wave form.FM is superseded by MFM (modified fre-quency modulation) is an encoding methodused in floppy disk drives and older harddrives. A competing scheme, known as RLL(run length limited), produces faster data ac-cess speeds and can increase a disk’s storagecapacity by up to 50 percent. MFM is super-seded by RLL, which is used on most newerhard drives.

binary phase grating a diffraction grat-ing where alternating grating lines that alterthe optical phase by 180◦ more than neigh-boring lines.

binary signal a signal that can only havetwo values: off and on, low and high, or zeroand one.

binary symmetric channel the binary-input, binary-output symmetric channel,where the channel noise and other distur-bances cause statistically independent errorsin the transmitted binary sequence with aver-age probability. The channel is memoryless.

binary tree recursively defined as a set ofnodes(n1, . . . nk) one of which is designatedthe root and the remainingk − 1 nodes format most two sub-trees.

binary tree predictive coding predictiveimage coding scheme in which pixels areordered in a pyramid of increasingly densemeshes. The sparsest mesh consists of sub-samples of the original image on a widelyspaced square lattice; succeeding meshesconsist of the pixels at the centers of thesquares (or diamonds) formed by all preced-ing meshes. Each mesh has twice the num-ber of pixels as its predecessor. Pixel val-ues are predicted by non-linear adaptive in-terpolation from surrounding points in pre-ceding meshes. The prediction errors, or dif-ferences, are quantized, ordered into a binarytree to provide efficient coding of zeros, andare then entropy coded.

binaural attribute psychoacoustic ef-fects (e.g., cocktail-party effect) that dependon the fact that we have two ears.

binocular imaging the formation of twoimages of a scene from two different po-sitions so that binocular vision can be per-formed, in a similar manner to the way hu-mans deploy two eyes.

binocular vision the use of two imagesof a scene, taken (often simultaneously) fromtwo different positions, to estimate depth ofvariouspoint features, once correspondencesbetween pairs of image features have beenestablished.

binomial coefficients the coefficients ofthe polynomial resulting from the expansion


of (a + b)n. These coefficients are equal to(n

k

)= n!k!(n− k)! ,

wheren is the order of the polynomial andkis the index of the coefficient. Thekth coef-ficient is multiplied by the termakbn−k.

binomial distribution the binomial dis-tribution is the distribution of a random vari-ableY that is the sum ofn random variablesthat areBernoulli distributed.

Y = X1 +X2 + · · · +Xn.

The probability mass function of such aY is

pY (k) =(n

k

)pk(1 − p)n−k,

wherep is the parameter of the Bernoulli dis-tribution of anyXi .

bioanalytical sensor a special caseof a chemical sensor for determining theamount of a biochemical substance. Thistype of sensor usually makes use of oneof the following types of biochemical reac-tions: enzyme-substrate, antigen-antibody,or ligand-receptor.

bioluminescence Seeluminescence.

biomass General term used for wood,wood wastes, sewage, cultivated herbaceousand other energy crops, and animal wastes.

biomedical sensor a device for interfac-ing an instrumentation system with a biolog-ical system such as biological specimen or anentire organism. The device serves the func-tion of detecting and measuring in a quanti-tative fashion a physiological property of thebiological system.

biometric verifier device that helps au-thenticate by measuring human characteris-tics.

biorthogonal filter bank a filter bank thatsatisfies the perfect reconstruction condition,

i.e., the product of the polyphase transferfunction of the analysis and synthesis filtersis a pure delay. In general, the analysis andsynthesis filters are different, as opposed tothe situation for an orthogonal filter bank.

biorthogonal wavelet a generalization oforthogonal wavelet bases, where two dual ba-sis functions span two sets of scaling spaces,Vj and Vj , and two sets of wavelet spaces,Wj andWj , with each scaling space orthog-onal to the dual wavelet space,i.e.,Vj ⊥ Wj

and Vj ⊥ Wj . See alsobiorthogonal filterbank.

BIOS Seebasic input–output system.

bipolar (1) a type of transistor thatuses both polarities of carriers (electrons andholes) in its operation as a junction transistor.

(2) a type of data encoding that uses bothpositive and negative voltage excursions.

bipolar device Seebipolar.

bipolar junction transistor (BJT) athree-terminal nonlinear device composed oftwo bipolar junctions (collector-base, base-emitter) in close proximity. In normal oper-ation, the voltage between base and emitterterminals is used to control the emitter cur-rent. The collector current either equals this(with BC junction in reverse bias), or goesinto saturation (the BC junction goes into for-ward bias). Used for medium power (700 A)and medium speed (10 kHz) applications.

In power electronics applications, BJTsare typically operated as switches, in eithertheir fully on or off states, to minimize losses.The base current flowing into the middle ofthe device controls the on–off state, wherecontinuous base current is required to be inthe on state. A disadvantage is the low cur-rent gain.

The base current is generally muchsmaller than collector and emitter currents,but not negligible as in MOSFETs.


bipolar memory memory in which a stor-age cell is constructed from bipolar junctiontransistors. See alsostatic random accessmemory (SRAM).

bipolar neuron a neuron with a signalbetween−1 and+1.

bipolar transistor Seebipolar junctiontransistor.

bipole DC system with two conductors,one positive and the other negative polarity.The rated voltage of a bipole is expressed as±100 kV, for example.

biquad an active filter whose transferfunction comprises a ratio of second-ordernumerator and denominator polynomials inthe frequency variable.

biquadratic transfer function a ratio-nal function that comprises a ratio of second-order numerator and denominator polynomi-als in the frequency variable.

bird’s beak feature seen in cross-sectional photomicrographs of silicon gatetransistors caused by encroachment of oxideunder the gate.

birefringence The property of certainmaterials to display different values of therefractive index for different polarizations ofa light beam.

birefringent fiber optical fiber thathas different speeds of propagation for lightlaunched along its (two distinct) polarizationaxes.

birefringent material material that canbe described by two or more refractive in-dices along the directions for principle axes.

birthmark a stamp on a wooden utilitypole which denotes its manufacturer, date ofmanufacture, size, and method of preserva-tion.

bispectra computation of the frequencydistribution of the EEG exhibiting nonlinearbehavior.

bispectrum the Fourier transform of thetriple correlation function. It preserves phaseinformation and uniquely represents a givenprocess in the frequency domain. It can beused to identify different types of nonlinearsystem response.

BIST Seebuilt-in self-test.

bistable pertaining to a device with twostable states. Examples: bistable multivibra-tor, flip-flop. See alsobistable system.

bistable device Seebistable.

bistable optical device a device whoseoptical transmission can take on two possiblevalues.

bistable system an optical system wherethe transmission can take on two possible val-ues.See alsobistable.

bistatic scattering a measure of the rera-diated power (back-scattered) from an illu-minated target in the direction other than thatof the illuminating source.

bit (1) the fundamental unit of informa-tion representation in a computer, short for“binary digit” and with two values usuallyrepresented by “0” and “1.” Bits are usu-ally aggregated into “bytes” (7 or 8 bits) or“words” (12–60 bits).

A single bit within a word may representthe coefficient of a power of 2 (in numbers),a logical TRUE/FALSE quantity (masks andBoolean quantities), or part of a character orother compound quantity. In practice, theseuses are often confused and interchanged.

(2) in Information Theory, the unit of in-formation. If an eventE occurs with aprobabilityP(E), it conveys information oflog2(1/P (E)) binary units or bits. Whena bit (binary digit) has equiprobable 0 and


1 values, it conveys exactly 1.0 bit (binaryunit) of information; the average informationis usually less than this.

bit allocation the allocation of bits to sym-bols with the aim of achieving some compres-sion of the data. Not all symbols occur withthe same frequency. Bit allocation attemptsto represent frequently occurring symbolswith fewer bits and assign more bits to sym-bols that rarely appear, subject to a constrainton the total number of bits available. Inthis way, the average string requires fewerbits. The chosen assignment of bits is usu-ally the one that minimizes the correspond-ing average coding distortion of the sourceover all possible bit assignments that satisfythe given constraint. Typically sub-sourceswith larger variances or energy are allocatedmore bits, corresponding to their greater im-portance.See alsotransform coding.

bit energy the energy contained inan information-bearing signal received at acommunications receiver per information bit.The power of an information bearing signalat a communications receiver divided by theinformation bit rate of the signal. Usuallydenoted byEb as in the signal to noise ratioEb/N0.

bit error rate (BER) the probability ofa single transmitted bit being incorrectly de-termined upon reception.

bit line used in, for example, RAM mem-ory devices (dynamic and static) to connectall memory cell outputs of one column to-gether using a shared signal line. In staticRAM, the “bit” line together with its com-plemented signal “-bit” feeds a “sense ampli-fier” (differential in this case) at the bottomof the column serving as a driver to the outputstage. The actual cell driving the bit line (and-bit) is controlled via an access transistor ineach cell. This transistor is turned on/off bya “word” line, a signal run across the cells ineach row.

bit parallel a method to transmit or pro-cess information in which several bits aretransmitted in parallel. Examples: a bit par-allel adder with 4-bit data has 8 input ports forthem (plus an initial carry bit); an 8-bit paral-lel port includes true 8-bit bi-directional datalines.

bit per second (bps) measure of trans-fer rate of a modem or a bus or any digitalcommunication support. (See alsobaudandbaud rate. bps and baud are not equivalentsince bps is a low-level measure and media;thus, it includes the number of bits sent for thelow-level protocol, while baud is typically re-ferred to a higher level of transmission).

bit period the time between successivebits in data transmission or data recording.At the transmitter (or recorder) the timing isestablished by a clock. At the receiver (orreader) an equivalent clock must be recoveredfrom the bit stream.

bit plane the binaryN×N image formedby selecting the same bit position of the pix-els when the pixels of anN × N image arerepresented usingk bits.

bit plane encoding lossless binary en-coding of the bit planes is termed bit planeencoding. The image is decomposed into aset ofk,N ×N bit planes from the least sig-nificant bit tok − 1 most significant bits andthen encoded for image compression.

bit rate (1) a measure of signaling speed;the number of bits transmitted per second.Bit rate and baud are related but not identi-cal. Bit rate is equal to baud times the numberof bits used to represent a line state. For ex-ample, if there are sixteen line states, eachline state encodes four bits, and the bit rate isthus four times the baud.See alsobaud.

(2) the number of bits that can be trans-mitted per unit time.

bit serial processing of one bit per clockcycle. If word length isW , then one sample


or word is processed inW clock cycles. Incontrast, allW bits of a word are processed inthe same clock cycle in a bit-parallel system.

For example: a bit serial adder with 4-bit data has one input signal for each of datathem, one bit for carry-in, and two 4-bit shiftregisters for data.

bit-line capacitance the equivalent ca-pacitance experienced in each “bit line” in aRAM or ROM device.See alsobit line.

bit-oriented block transfer (bitBLT) atype of processing used mainly for video in-formation characterized by minimal opera-tions performed on large data blocks; a pro-cessor designed for such operations. bit-BLT operations include transfers, masking,exclusive-OR, and similar logical functions.

bit-serial system a system that uses bitserial data transfer.

bit-slice processor a processor organiza-tion that performs separate computations (viamultiple processing units) separately uponsubsections of an incoming channel.

bitBLT Seebit-oriented block transfer.

bitmapped image a digital image com-posed of pixels. Bitmapped images areresolution-dependent, i.e., if the image isstretched, the resolution changes. Alsocalled a raster image.See alsoimage, pixel,vector image.

bits per pixel the number of bits used torepresent each pixel in a digital image. Typ-ical grayscale images have 8 bits per pixel,giving 256 different gray levels. True colorimages have 24 bits per pixel, or 8 bits foreach of the red, green, and blue pixels. Com-pressed image sizes are often represented inbits per pixel, i.e., the total number of bitsused to represent the compressed image di-vided by the total number of pixels.

BJT Seebipolar junction transistor.

black burst a TV black video signalcontaining horizontal and vertical sync, colorburst, and setup (i.e., a composite video blacksignal). Black burst is also called “colorblack.” A black burst signal is often usedin the video studio to provide synchronizingpulses.

black level the portion of the video sig-nal pertaining to the lower luminance (bright-ness) levels.

black start the task of re-starting an iso-lated power system which is completely de-energized. Most generating plants requiresubstantial external electric power to start.Thus a black start may be initiated by hand-starting gas turbine generators or by openingthe gates of a hydroelectric generator some-where in the system.

blackbody theoretically contrived objectthat gives rise to the so-called “black body ra-diation.” One might imagine a closed surfaceobject (say of metal) possessing one open-ing that connects the interior surface with theoutside world. When the object is heated, theopening becomes a perfect “black” radiator.Such radiation depends on temperature only.

blackout total loss of power to the entirepower system.

blanket an insulating rubber mat which isfitted temporarily over energized conductorsto protect nearby workers.

blanking the electronic control circuitrythat blanks the television raster during hori-zontal and vertical retrace.

blanking time the short time intervalwhen both switches in a leg of an inverterbridge must be off in order to prevent shortcircuiting the DC input. This is necessary be-cause non-ideal switches cannot turn on andoff instantaneously. Thus, after one switch isturned off in an inverter leg, the complimen-


tary switch is not turned on until the desig-nated blanking time has elapsed.

blind deconvolution the recovery of a sig-nalx[n] from y[n] — the convolution of thesignal with an unknown systemh[n]:

y[n] = h[n] ∗ x[n].

Occasionally, some knowledge ofh[n] isavailable (e.g., that it is a high-pass or low-pass filter). Frequently, detailed knowledgeis available about the structure ofx. See alsoconvolution.

blind via a via connected to either thepreliminary side or secondary side and oneor more internal layers of a multilayer pack-aging and interconnecting structure.

blink in computer display systems, a tech-nique in which a pixel is alternatively turnedon and off.

Bloch vector a set of linear combinationsof density matrix elements, written in vectorform, that can often be related to specific ob-servables in a quantum mechanical system.For example, in two-level systems the Blochvector components are 2Re(ρ12), 2Im(ρ12),andρ11 − ρ22, which are related to nonlin-ear refractive index, absorption, and popula-tion differences, respectively. The time evo-lution of two-level systems can be describedin terms of rotations of the Bloch vector.

block a group of sequential locations heldas one unit in a cache and selected as whole.Also called a line.See alsomemory block.

block cipher an encryption system inwhich a successive number of fundamen-tal plaintext information symbols, usuallytermed a block of plaintext information, areencrypted according to the encryption key.All information blocks are encrypted in thesame manner according to the transforma-tion determined by the encryption key. Thisimplies that two identical blocks of plaintext

information will always result in the same ci-phertext when a particular block cipher is em-ployed for encryption.See alsoencryption,stream cipher.

block code a mapping ofk input binarysymbols inton output symbols.

block coding (1) an error control codingtechnique in which a number of informationsymbols, and blocks, are protected againsttransmission errors by adding additional re-dundant symbols. The additional symbolsare usually calculated according to a mathe-matical transformation based on the so-calledgenerator polynomial of the code. A blockcode is typically characterized by the param-eters(n, k), wherek is the number of infor-mation symbols per data word, andn the fi-nal number of symbols in the code word afterthe addition of parity symbols or redundantsymbols. The rate of a block code is givenby k/n.

Typically, the lower the rate of a codethe greater the number of errors detectableand correctable by the code. Block codesin which the block of information symbolsand parity symbols are readily discernable,are known as systematic block codes. Thereceiver uses the parity symbols to deter-mine whether any of the symbols were re-ceived in error and either attempts to cor-rect errors or requests a retransmission ofthe information. See alsoautomatic repeatrequest, binary code, convolutional coding,error control coding.

(2) refers to (channel) coding schemesin which the input stream of informationsymbols is split into nonoverlapping blockswhich then are mapped into blocks of en-coded symbols (codewords). The mappingonly depends on the current message block.Compare withtrellis coding.

block diagram a diagrammatic represen-tation of system components and their inter-connections. In elementary linear systems,the blocks are often defined by transfer func-tions or state space equations while the inter-


connecting signals are given as Laplace trans-formations. Although the system blocks andsignals have the same mathematical form, theblocks represent operators that act on the in-coming signals while the signals representfunctions of time.

block matching the process of finding theclosest match between a block of samples ina signal and a block of equal size in anothersignal (or a different part of the same sig-nal) over a certain search range. Closenessis measured by correlation or an error metricsuch as mean square error. Used in data com-pression, motion estimation, vector quantiza-tion, and template matching schemes.

block multiplexer channel an I/O chan-nel can be assigned to more than one datatransfer at a time. It always transfers infor-mation in blocks, with the channel releasedfor competing transfers at the end of a block.See alsobyte multiplexer channel, selectorchannel.

block transfer the transmission of a sig-nificantly larger quantity of data than the min-imum size an interconnect is capable of trans-mitting, without sending the data as a numberof small independent transmissions (the goalbeing to reduce arbitration and address over-head).

block transform a transform that dividesthe image into several blocks and treats eachblock as an independent image. The trans-form is then applied to each block indepen-dently. This occurs in the JPEG standard im-age compression algorithm, where an imageis divided into 8× 8 blocks and the DCTis applied independently to each block. Usu-ally the blocks do not overlap each other, thatis, they have no signal samples in common.Seealso transform coding,lapped orthogonal transform.

block truncation coding (BTC) tech-nique whereby an image is segmented inton × n nonoverlaping blocks of pixels, and

a two-level quantizer is designed for eachblock. Encoding is essentially a local bina-rization process consisting of an×n bit mapindicating the reconstruction level associatedwith each pixel. Decoding is a simple pro-cess of associating the reconstructed value ateach pixel as per the bit map.

block carry lookahead adder (BCLA)an adder that uses two levels of carry looka-head logic.

block-diagram simulator a simulatorthat allows the user to simulate systems asa combination of block diagrams, each ofwhich performs a specific function. Eachfunction is described using a mathematicalequation or a transfer function.

blocked state Seeblocking.

blocked-rotor current Seelocked-rotorcurrent.

blocked-rotor test an induction motortest conducted with the shaft held so it cannotrotate. Typically about 25% of rated voltageis applied, often at reduced frequency and thecurrent is measured. The results are used todetermine the winding impedances referredto the stator.

blocking state entered if a new user findsall channels or access mechanisms busy andhence is denied service. Generally accom-panied by a busy signal. The call blockingprobability may be given by the Erlang B orErlang C formula.See alsoadequate service,multiple access interference (MAI).

blocking artifact the visibility in an im-age of rectangular subimages or blocks af-ter certain types of image processing. Alsocalled blocking effect distortion.

blocks world a visual domain, typicalof early studies on machine vision, in whichobjects are light, plane-faced solids over adark background.


blooming an area of the target that is un-stable due to insufficient beam current. Thearea normally appears as a white puddle with-out definition. Insufficient beam currentlymay be the result of low beam control set-ting.

blow up a relatively sudden and usuallycatastrophic increase in beam size generallycaused by some magnetic field error drivingthe beam to resonance.

Blumlein a water-filled transmission linethat serves as a pulse generator using a wavepropagation principle. The line is folded overon itself and is capable of voltage doublingacross its load due to having initially bothsides of the load on high potential.

Blumlein bridge an AC bridge, two armsof which are two serially connected tightlycoupled inductive coils. The point of connec-tion of these coils is usually grounded, andthe coupling is arranged in such a way thatfor the currents simultaneously entering orleaving the other ends of the coils the voltagedrop between the ends is close to zero. If oneof the currents is entering and another is leav-ing, then the voltage drop is essential. Thiscreates a sensitive current-comparing bridgehaving application in capacitance transduc-ers.

blurring (1) the defocusing effect pro-duced by the attenuation of high-frequencycomponents, e.g., obtained by local averag-ing operators, possibly applied directionally(motion blurring).

(2) the broadening of image features, rela-tive to those which would be seen in an idealimage, so that features partly merge into oneanother, thereby reducing resolution. The ef-fect also applies to 1-D and other types ofsignal.

BNC connector “Baby” N connector.Commonly used coaxial connector with bothmale and female versions used below mi-crowave frequencies.

board the physical structure that housesmultiple chips, and connects them with traces(busses).

board-to-board optical interconnectoptical interconnection in which the sourceand the detector are connected to electronicelements in two separate boards.

BOB Seebreak-out box.

Bode diagram SeeBode plot.

Bode plot a graphical characteriza-tion of the system frequency response:the magnitude of the frequency response|H(jω)|,−∞ < ω < ∞ in decibels, andthe phase angle6 H(jω),−∞ < ω < ∞,are plotted. For example, a system describedby the transfer function

H(s) = Y (s)

F (s)= s + 1

(s + 2)(s + 3)

has the Bode plot shown in the following fig-ure. See alsofrequency response.

Bode plot.

Bode–Fano criteria a set of rules fordetermining an upper limit on the bandwidthof an arbitrary matching network.

boiler a steam generator that converts thechemical energy stored in the fuel (coal, gas,etc.) to thermal energy by burning. The heat


evaporates the feedwater and generates high-pressure steam.

boiling water reactor a nuclear reactorfrom which heat is transferred in the form ofhigh-pressure steam.

bolted fault a bolted fault is a short cir-cuit fault with no fault resistance. Boltedfaults deliver the highest possible fault cur-rent for a given location and system config-uration, and are used in selecting equipmentwithstand and interrupting ratings and in thesetting of protective relays.

Boltzmann machine in its simplest form,a discrete time Hopfield network that em-ploys stochastic neurons and simulated an-nealing in its procedure for updating outputvalues. More generally it can have hiddenunits and be subjected to supervised trainingso as to learn probabilities of different out-puts for each class of inputs.

Boltzmann relation relates the densityof particles in one region to that in an ad-jacent region, with the potential energy be-tween both regions.

bond that which binds two atoms to-gether.

bond pad areas of metallization on the ICdie that permit the connection of fine wiresor circuit elements to the die.

bonded magnet a type of magnetconsisting of powdered permanent magnetmaterial, usually isotropic ceramic ferriteor neodymium-iron-boron, and a polymerbinder, typically rubber or epoxy, this magnetmaterial can be molded into complex shapes.

bonding the practice of ensuring a low-resistance path between metallic structuressuch as water lines, building frames, and ca-ble armor for the purpose of preventing light-ning arcs between them.

Boolean an operator or an expression ofGeorge Boole’s algebra (1847). A Booleanvariable or signal can assume only two val-ues: TRUE or FALSE. This concept has beenported in the field of electronic circuits byClaude Shannon (1938). He had the idea touse the Boole’s algebra for coding the statusof circuit: TRUE/FALSE as HIGH/LOW asCLOSE/OPEN, etc.

Boolean algebra the fundamental algebraat the basis of all computer operations. Seealso the other definitions with Boolean as thefirst word.

Boolean expression an expression ofthe Boole’s algebra, in which can appearBoolean variables/signals and Boolean oper-ators. Boolean expressions are used for de-scribing the behavior of digital equipmentsor stating properties/conditions in programs.

Boolean function common designationfor a binary function of binary variables.

Boolean logic the set of rules for logicaloperations on binary numbers.

Boolean operator the classical Booleanoperators are AND, OR, NOT. Other oper-ators such as XOR, NAND, NOR, etc., canbe easily obtained based on the fundamen-tal ones. In hardware these are implementedwith gates, see for example AND gate.

boost converter a circuit configuration inwhich a transistor is switched by PWM trig-ger pulses and a diode provides an inductor-current continuation path when the transistoris off. During the transistor on-time, the cur-rent builds up in the inductor. During thetransistor off-time, the voltage across the in-ductor reverses and adds to the input voltage,as a result, the output voltage is greater thanthe input voltage.

A boost converter can be viewed as areversed buck converter. The output volt-agevo is related to the input voltagevi byvo = vi/(1 − d) and it can be controlled by


Boost converter.

varying the duty ratiod. Its main applica-tion is in regulated DC power supplies andthe regenerative braking of DC motors. Alsocalled a step-up converter.

boot Seebootstrap.

boot record structure at the beginning ofa hard disk that specifies information neededfor the start up and initialization of a com-puter and its operating system. This recordis kept and displayed by the booting program.

bootstrap (1) a technique using positivefeedback to change the effective impedanceat a node, for example, to reduce capacitance.

(2) to initialize a computer system into aknown beginning state by loading the oper-ating system from a disc or other storage tocomputer’s working memory. This is done bya firmware boot program. Also called bootfor short.

boson an integral spin particle to whichBose-Einstien statistics apply. Such particlesdo not follow the Pauli exclusion principle.Photons, pions, alpha particles, and nuclei ofeven mass numbers are examples of bosons.

bottle slang for glass insulator.

bottom antireflective coating an antire-flective coating placed just below the pho-toresist to reduce reflections from the sub-strate.

bottom-up development an applicationdevelopment methodology that begins creat-ing basic building blocks and uses them tobuild more complex blocks for higher levelsof the system.

bound mode a type of mode of limitedspatial extension. Open waveguides can sup-port, apart for a continuous spectrum, alsoa few mode, which do not extend up to in-finity since they decay exponentially outsideof a certain region. In an optical waveguidethis is a mode whose field decays monoton-ically in the direction transverse to propaga-tion and which does not lose power to radi-ation. Bound modes can also be interpretedin terms of guided rays and total internal re-flection. Note: Except in a monomode fiber,the power in bound modes is predominantlycontained in the core of the fiber.See alsocontinuous spectrum.

boundary a curve that separates two setsof points.

boundary bus one of a set of buses whichdefine the boundary between the portion of apower system to be analyzed and the rest ofthe system. Boundary buses are connectedto both the internal and external systems.

boundary condition (1) the conditionssatisfied by a function at the boundary of itsinterval of definition. They are generally dis-tinguished in hard or soft also called Neu-mann (the normal derivative of the functionis equal to zero) or Dirichlet (the functionitself is equal to zero).

(2) the conditions satisfied from the elec-tromagnetic field at the boundary betweentwo different media.

(3) rules that govern the behavior of elec-tromagnetic fields as they move from onemedium into another medium.

boundary layer a method of smooth-ing out a discontinuous controller or a sliding


mode state estimator. For example, a bound-ary layer version of the discontinuous con-troller

u = −Us(e)/|s(e)| = −Usign(s(e)),

wheree is the control error ands is a functionof e, may have the form

u ={−Usign(s(e)) if |s(e)| > ν

−Us(e)/ν if |s(e)| ≤ ν,

whereν > 0 is called the boundary layerwidth.

boundary layer controller Seeboundarylayer.

boundary layer observer Seeboundarylayer state estimator.

boundary layer state estimator a con-tinuous version of a sliding mode type stateestimator, that is, a sliding mode type stateestimator in which the right-hand side of thedifferential equation describing the estima-tor is continuous due to the introduction of aboundary layer to smooth out the discontin-uous part of the estimator’s dynamics.

boundary scan a technique for applyingscan design concepts to control/observe val-ues of the signal pins of IC components byproviding a dedicated boundary-scan registercell for each signal I/O pin.

boundary scan interface a serial clockedinterface used to shift in test pattern or testinstruction and to shift out test responses inthe test mode. Boundary scan interface com-prises shift-in, shift-out, clock, reset, and testselect mode signals.

boundary scan path a technique that usesa standard serial test interface to assure easyaccess to chip or board test facilities such astest registers (in an external or internal scanpaths) or local BIST. In particular it assurescomplete controllability and observability ofall chip pins via shift in and shift out opera-tions.

boundary scan test a technique forapplying scan design concepts to con-trol/observe values of signal pins of IC com-ponents by providing a dedicated boundary-scan register cell for each signal I/O pin.

boundary value problem a mathemat-ical problem in which the unknown is a so-lution to a partial differential equation and issubject to a set of boundary conditions on theproblem domain.

boundary values of 2-D general modellet xi,j be a solution (semistate vector) to the2-D generalized model

xi+1,j+1 = A0xi,j + A1xi+1,j

+ A2xi,j+1 + B0ui,j

+ B1ui+1,j + B2ui,j+1

i, j ∈ Z+ (the set of nonnegative integers)whereui,j ∈ Rm is the input andAk,Bk (k =0,1,2) are real matrices of the model. Thevectorsxi,j ∈ Rn whose indices lie on theboundary of the rectangle[0, N1] × [0, N2],i.e., xi,0, xi,N2 for 1 ≤ i ≤ N1 andx0,j , xN1,j for 0 ≤ j ≤ N2, are calledboundary values of the solutionxi,j to the 2-D general model. The boundary values maybe also given in other ways.

boundary-element method (BEM) a nu-merical method (integral equation technique)well suited to problems involving structuresin which the dielectric constant does not varywith space.

bounded control Seesaturating control.

bounded distance decoding decoding ofan imperfectt-error correcting forward errorcorrection block code in which the correctederror patterns are limited to those witht orfewer errors, even though it would be possi-ble to correct some patterns with more thanterrors.

bounded function a functionx ∈ Xe issaid to be bounded if it belongs also to the


original (unextended) spaceX , whereX isa space of functions with its correspondingextensionXe. See alsoextended spaceandtruncation.

bounded state an equilibrium statexeof a dynamic system is said to be boundedif there exists a real numberB = B(x0, t0),wherex0 andt0 represent the initial values ofthe state and time, respectively, such that

‖ x(t) ‖< B ∀t ≥ t0

See alsostable state.

bounded-input bounded-output (BIBO)a signal that has a certain value at a cer-

tain instant in time, and this value does notequal infinity at any given instant of time. Abounded output is the signal resulting fromapplying the bounded-input signal to a stablesystem. See diagram below.

Bounded-input bounded-output system.

bounded-input bounded-output stabilitya linear dynamic system where a bounded

input yields a bounded zero-state response.More precisely, let be a bounded-input withas the least upper bound (i.e., there is a fixedfinite constant such that for everyt or k), ifthere exists a scalar such that for everyt (ork), the output satisfies, then the system is saidto be bounded-input bounded-output stable.

bounded-input bounded-state (BIBS) sta-bility if for every bounded input (SeeBIBO stability), and for arbitrary initial con-ditions, there exists a scalar such that the re-sultant state satisfies, then the system is saidto be bounded-input bounded-state stable.

bounds fault an error that holds the map-per whet it detects the offset requested intoan object exceeds the object’s size.

Boyle macromodel A SPICE computermodel for an op amp. Developed by G. R.Boyle in 1974.

Boys camera a rotating camera used tophotograph lightning and establish the mul-tiplicity of individual flashes in a lightningstroke.

BPI bits per inch.

bps Seebit per second.

Bragg angle the required angle of inci-dence for light into a Bragg cell to producea single diffraction order of maximum inten-sity. The sine of the Bragg angle is approx-imately the light wavelength divided by thegrating.

Bragg cell an acousto-optic cell designedwhere only a single diffraction order is pro-duced, generally by making the acoustic col-umn thick along the light propagation direc-tion.

Bragg cell radiometer similar to anacousto-optic spectrum analyzer in the Braggmode, but with generally much longer photo-integration times such as via a long integra-tion time photo detector array.

Bragg diffraction the interaction of lightwith a thick grating or acoustic wave, produc-ing a single diffraction order with maximumintensity.

Bragg diffraction regime regime wherethe acoustic beam width is sufficiently wideto produce only two diffracted beams, i.e.,the undiffracted main beam (also called thezero order or DC beam), and the principaldiffracted beam.

Bragg scattering the scattering of lightfrom a periodically varying refractive indexvariation in a thick medium, so-called byanalogy to the Bragg scattering of X-raysfrom the atomic arrays in a crystal. For in-


stance, an acousto-optic modulator can besaid to operate in the Bragg regime or al-ternatively in the Raman–Nath regime.SeealsoRaman–Nath diffraction regime.

braking operating condition in an electricmotor in which the torque developed betweenthe stator and rotor coils opposes the direc-tion of rotation of the rotor. Typical brakingmethods in DC machines include “plugging”in which the polarity of either the field orthe armature coil, but not both, is reversedwhile the rotor is turning, “dynamic brak-ing” in which generator action in the arma-ture is used to dissipate rotor energy througha braking resistor, and “regenerative brak-ing” in which generator action in the rotoris used to dissipate rotor energy by return-ing electric power to the power source as therotor slows. Typical braking methods in ACmachines include switching of the phase se-quence of the supply voltage, dynamic brak-ing through the armature coils, and varyingthe frequency of the AC supply voltage.Seealsophase sequence.

braking resistor resistive elements whichcan be switched into the electrical system tocreate additional load in the event of a tran-sient disturbance, thus limiting the generatorrotor acceleration such that the system canmore readily return to synchronism.

branch address the address of the in-struction to be executed after a branch in-struction if the conditions of the branch aresatisfied. Also called a branch target address.

branch circuit the three components ofan electrical circuit are source, load, and in-terconnecting circuit conductors. A branchcircuit is an electrical circuit designed to de-liver power to the lowest-order load(s) servedon a facility. It includes the overcurrent de-vice, circuit conductors, and the load itself.

branch current the current in a branchof a circuit.

branch history table a hardware compo-nent that holds the branch addresses of pre-viously executed branch instructions. Usedto predict the outcome of branch instructionswhen these instructions are next encountered.Also more accurately called a branch targetbuffer.

branch instruction an instruction is usedto modify the instruction execution sequenceof the CPU. The transfer of control to an-other sequence of instructions may be uncon-ditional or conditional based on the result ofa previous instruction. In the latter case, ifthe condition is not satisfied, the transfer ofcontrol will be to the next instruction in se-quence. It is equivalent to a jump instruction,although the range of the transfer may be lim-ited in a branch instruction compared to thejump. See alsojump instruction.

branch line coupler coupler comprisedof four transmission lines, each of 90◦ electri-cal length, arranged in a cascaded configura-tion with the end of the last transmission linesection connected to the beginning of the firsttransmission line to form a closed path. Theinput, coupled, direct, and isolated ports arelocated at the connection point of one trans-mission line with the next one.

branch penalty the delay in a pipelineafter a branch instruction when instructionsin the pipeline must be cleared from thepipeline and other instructions fetched. Oc-curs because instructions are fetched into thepipeline one after the other and before theoutcome of branch instructions are known.

branch prediction a mechanism used topredict the outcome of branch instructionsprior to their execution.

branch relation the relationship betweenvoltage and current for electrical compo-nents. Common branch relations are Ohm’sLaw and the lumped equations for capacitorsand inductors. More complex branch rela-tionships would be transistor models.


branch target buffer (BTB) a buffer thatis used to hold the history of previous branchpaths taken during the execution of individ-ual branch instructions. The BTB is usedto improve prediction of the correct branchpath whenever a branch instruction is en-countered.

The branch target bufferor branch tar-get cachecontains the address of each re-cent branch instruction (or the instructionsthemselves), the address of the branch “tar-get” and a record of recent branch directions.The Pentium BTB is organized as an associa-tive cache memory, with the address of thebranch instruction as a tag; it stores the mostrecent destination address plus a two-bit his-tory field representing the recent history ofthe instruction.

branch target cache Seebranch targetbuffer.

branch voltage the voltage across abranch of a circuit.

Branly, Edouard Eugene (1844–1940)Born: Amiens, France

Branly is best known for his work in wire-less telegraphy. Branly invented the coherer,a detection device for radio waves. Branlydid much theoretical work in electrostatics,electrodynamics, and magnetism. He didnot, however, develop the practical side of hiswork, hence Marconi and Braun received theNobel Prize for work Branly had pioneered.

Brattain, Walter (1902–1987) Born:Amoy, China.

Brattain is best known as one of the de-velopers of the transistor. In 1956 Brattain,along with John Bardeen and William Shock-ley, received the Nobel Prize for their devel-opment of the point-contact transistor. It wasBrattain who, along with Bardeen, observedthe significant increase in power output froma metal contact resulting from a small in-crease in current applied through a secondcontact attached to the same germanium sur-

face. This research led to the development ofintegrated circuits.

Braun, Karl Ferdinand (1850–1918)Born: Fulda, Germany

Braun is best known for his invention ofthe oscilloscope and for improvements toMarconi’s telegraph. Braun was to sharethe Nobel Prize in Physics with Marconi in1909. Braun held a number of teaching poststhroughout Germany. His research resultedin the principle of magnetic coupling, whichallowed significant improvements in radiotransmission. He discovered crystal recti-fiers, which were a significant component inearly radio sets.

breadboard a preliminary, experimentalcircuit, board, device or group of them. It isbuilt only to investigate, test, analyze, eval-uate, validate, determine feasibility, developtechnical data, and to demonstrate the tech-nical principles related to a concept, device,circuit, equipment, or system. It is designedin a rough experimental form, only for labo-ratory use, and without regard to final physi-cal appearance of a product.

breadth-first search a search strategy fortree or trellis search where processing is per-formed breadth first, i.e., the processing forthe entire breadth of the tree/trellis is com-pleted before starting the processing for thenext step forward.

break frequency the critical frequencyin a frequency - dependent response: espe-cially that frequency which may separate twomodes of the response, e.g. the frequencythat defines where the low frequency regionends and the midband response begins.

break point Seebreakpoint.

break-out box (BOB) a testing devicethat allows the designer to switch, cross, andtie interface leads. It often has LEDs to per-mit monitoring of the leads. Typical use isfor RS-232 interfaces.


breakaway points of the root loci break-away points on the root loci correspond tomultiple-order roots of the equation.

breakaway torque minimum torqueneeded to begin rotating a stationary load.Breakaway torque represents the absoluteminimum starting torque specification for amotor used to drive the load.

breakdown as applied to insulation (in-cluding air), the failure of an insulator or in-sulating region to prevent conduction, typi-cally because of high voltage.

breakdown strength voltage gradient atwhich the molecules of medium break downto allow passage of damaging levels of elec-tric current.

breakdown torque maximum torque thatcan be developed by a motor operating atrated voltage and frequency without expe-riencing a significant and abrupt change inspeed. Sometimes also called the stall torqueor pull-out torque.

breakdown voltage the reverse biasedvoltage across a device at which the currentbegins to dramatically deviate and increaserelative to the current previously observed atlower voltages close to the breakdown volt-age. This effect is attributed to avalanche orzener breakdown. It is usually specified at apredetermined value of current.

In a diode, applying a voltage greater thanthe breakdown voltage causes the diode tooperate in the reverse breakdown region.

breakpoint (1) an instruction address atwhich a debugger is instructed to suspend theexecution of a program.

(2) a critical point in a program, at whichexecution can be conditionally stopped toallow examination if the program variablescontain the correct values and/or other ma-nipulation of data. Breakpoint techniques areoften used in modern debuggers, which pro-

vide nice user interfaces to deal with them.See alsobreakpoint instruction.

breakpoint instruction a debugging in-struction provided through hardware supportin most microprocessors. When a programhits a break point, specified actions occur thatsave the state of the program, and then switchto another program that allows the user to ex-amine the stored state. The user can suspendthe execution of a program, examine the reg-isters, stack, and memory, and then resumethe program’s execution, which is very help-ful in a program’s debugging.

breath noise the noise that is commonlyproduced when talking at the microphone. Itis due to breathing.

breeder reactor a nuclear reactor inwhich a non-fissile isotopes are converted tofissile isotopes by irradiation. Ideally, sucha reactor produces more fissile products thanit consumes.

Bremsstrahlung electromagnetic radia-tion, usually in the X-ray region of the spec-trum produced by electrons in a collision withthe nucleus of an atom. Bremsstrahlung ra-diation is produced in regions of high electricpotential such as areas surrounding electro-static septa and RF cavities. Bremsstrahlungis German for breaking.

Brewster angle the angle from normal atwhich there is no reflection at a planar inter-face between two media. The Brewster an-gles for perpendicular and parallel polariza-tions are different. For nonmagnetic media,in which the relative permeability is unity,the Brewster angle for perpendicular polar-ization does not exist.

Brewster mode a bound radiative sur-face mode when one of the media is a plasmamedium and has a positive dielectric func-tion.


Brewster window transmission windoworiented at Brewster’s angle with respect toan incident light beam; light polarized in theplane of incident experiences no reflection.

bridge a simple device that connectstwo or more physical local-area networks(LANs). It forwards packets of data fromone LAN segment to another without chang-ing it, and the transfer is based on physicaladdresses only. The separate LAN segmentsbridged this way must use the same protocol.

bridge balance condition represents therelationship between bridge circuit compo-nents when the current in the balance indica-tor is absent. Most of the technically usefulbridges include a regular connection (series,parallel, series-parallel, or parallel-series) oftwo two-ports. The condition of balance canbe reformulated in terms of two-port param-eters, so that depending on structure, the sumof two forward transfer parameters or the sumof one forward and another backward transferparameter is equal to zero.

bridge calibration used in bridge trans-ducer applications. It is achieved connect-ing two auxiliary circuits to the bridge. Onecircuit including two resistors and a poten-tiometer is connected in parallel to the bridgepower supply diagonal, and the potentiome-ter tap and one end of detector are connectedto the same bridge node. Sliding the tap,one can eliminate the bridge offset. An-other circuit, usually including a constantand a variable resistor, is connected in serieswith power supply. This circuit allows oneto change the voltage applied to the bridge,and to establish the correspondence betweenthe maximal deflection of the detector andmaximum of the physical variable applied tothe bridge resistors playing the role of activegauges.

bridge circuit the circuit that includesfour lateral impedances,Z1, Z2, Z3, Z4, adiagonal impedanceZo, and a voltage sourceEg of the output impedanceZg is an exam-

ple of the so-called bridge circuit. This andother similar circuits are characterized by thebridge balance condition, which representsa relationship between the bridge elementswhen the current in the diagonal impedanceis absent (in the case shown this condition isZ1Z3 = Z2Z4). The bridge circuits find ap-plication in instrumentation and transducers.

bridge linearization necessary designconcern in transducer application of thebridge circuits. It is achieved by reductionof the bridge sensitivity in the bridges whereonly one arm is a transducer. Linearizationcan also be achieved with two transducersproviding the signals of opposite signs andconnected in the opposite arms of the bridgeor using a current source instead of the volt-age source as a bridge power supply.

bridge rectifier a full-wave rectifier toconvert ac to dc, that contains four rectifyingelements for single phase, and six elementsfor three phase, connected as the arm of abridge circuit.

bridge sensitivity the ratio of the vari-ation of the voltage or the current throughthe detector to the variation of the compo-nent that causes the disbalance of the bridgecircuit.

bridge-controlled multivibrators usingswitches in a two-operational amplifier or inan amplifier-comparator multivibrator so thatthe bridge is “rotated” each half of the period,one can obtain control of the oscillation fre-quency by detuning a resistive bridge. Thecircuit can be applied in sensors with limitednumber of access wires.

bridging using bridges for local-area net-works.

brightness the perceived luminance or ap-parent intensity of light. This is often differ-ent from the actual (physical) luminance, asdemonstrated by brightness constancy, Machband, and simultaneous contrast.


brightness adaptation the ability of thehuman visual system (HVS) to shift the nar-row range in which it can distinguish differ-ent light intensities over a large span of lumi-nances. This permits the overall sensitivityof the HVS to gray levels to be very largeeven though the number of gray levels thatit can simultaneously differentiate is fairlysmall. See alsogray level, human visualsystem (HVS), luminance.

brightness constancy the perception thatan object has the same brightness despitelarge changes in its illumination. Thus apiece of paper appears to be approximatelyas white in moonlight as in sunlight, eventhough the illumination from the sun maybe one million times greater than that fromthe moon.See alsobrightness, human visualsystem (HVS), illumination, simultaneouscontrast.

Brillouin flow a stream of electron beamemitted from an electron gun that is not ex-posed to a focusing magnetic field.

Brillouin frequency shift the frequencyshift that a wave experiences in undergoingBrillouin scattering. The shift can be to eitherlower or higher frequency, and typically hasa value in the range 0.1 to 10 GHz.See alsoStokes scattering, anti-Stokes scattering.

Brillouin laser acoustic maser in whichthe amplification mechanism is considered tobe Brillouin scattering.

Brillouin scattering the scattering oflight from sound waves. Typically in Bril-louin scattering the sound waves have fre-quencies in the range 0.1 to 10 GHz, whereasin acousto-optics the sound waves have fre-quencies<0.1 GHz. Brillouin scattering canbe either spontaneous or stimulated.Seealso acousto-optic effect, spontaneous lightscattering, stimulated light scattering.

broadband a service or system requiringtransmission channels capable of supportingbit rates greater than 2 Mbit/s.

broadband antenna an antenna whosecharacteristics (such as input impedance,gain, and pattern) remain almost constantover a wide frequency band. Two such typesof antennas are the log periodic and the bi-conical.

broadband emission an emission havinga spectral distribution sufficiently broad incomparison to the response of a measuringreceiver.

broadband integrated services digital net-work (B-ISDN) a generic term that gen-erally refers to the future network infrastruc-ture that will provide ubiquitous availabilityof integrated voice, data, imagery, and videoservices.

broadband system a broadband commu-nication system is one that employs a highdata transmission rate. In radio terminologyit implies that the system occupies a wide ra-dio bandwidth.

broadcast (1) the transfer of data to multi-ple receiver units simultaneously rather thanto just one other subsystem.

(2) a bus-write operation intended to berecognized by more than one attached device.

broadcast channel a single transmit-ter, multiple receiver system in which iden-tical information is transmitted to each re-ceiver, possibly over different channels.Seealso interference channel, multiple accesschannel.

broadcast channel allocations a fre-quency of a width prescribed by a nation’scommunications governing agency that arestandardized throughout the country for usein one-way electronic communication.


broadcast picture quality the accept-able picture performance for NTSC terres-trial telecast signals. A panel of untrainedobservers subjectively evaluates the NTSCreceived picture and sound quality as signalimpairments are inserted into the broadcastsignal. The evaluation scores are used todetermine the values for objectionable sig-nal impairment levels. The signal impair-ments tested are the video and audio signal-to-noise ratios, the interference due to adja-cent channel signals, the interference due toco-channel signals, and the echoes (ghosts)caused by multipath signals effects.

broadcasting sending a message to mul-tiple receivers.

broadside when the pattern factor ismaximum in the H plane (for a dipole an-tenna along the z axis this is the plane wheretheta=90 degrees).

broadside array an array where the mainbeam of the array is directed perpendicularto the array axis. In many applications it isdesirable to have the maximum radiation ofan array directed normal to the axis of thearray.

broadside coupled microstrip lines mi-crostrip lines that share the same groundplane but separated from each other in normaldirection to the ground plane. Both the mi-crostrip lines are aligned at their centers alongthe normal direction to the ground plane.

Brown book SeeIEEE Color Books.

Brownian motion a stochastic processwith independent and stationary increments.The derivative of such a process is a whitenoise process. A Brownian motion processXt is the solution to a stochastic differentialequation of the form

dX

dt= b(t, Xt )+ σ(t, Xt ) ·Wt,

whereWt is a white noise process.

brownout an intentional lowering of util-ity voltage to reduce loading on the system.

brush a conductor, usually carbon or acarbon–copper mixture, that makes slidingelectrical contact to the rotor of an electri-cal machine. Brushes are used with slipringson a synchronous machine to supply the DCfield and are used with a commutator on a DCmachine.

brush rigging the components used tohold the brushes of a rotating machine inplace, and to insure proper brush tension isapplied.

brush tension the force required on thebrushes of a rotating machine to insure propercontact between the brush and the commuta-tor or slipring. Proper brush tension is usu-ally provided by springs, and is specified inthe manufacturer’s technical manual of themachine.

brushless DC motor Seeelectronicallycommutated machine.

brushless exciter Seerotating-rectifierexciter.

brushless rotary flux compressor a ro-tating machine designed to deliver pulsedoutput (1 MJ in 100µs). The stator coils areexcited by an external capacitor bank. Therotor is a salient structure that compressesthe flux resulting in amplification of the elec-tric pulse, by converting the rotating kineticenergy of the rotor to electrical energy.

BSO abbreviation for bismuth silicon ox-ide, Bi4SiO20. A photoconductive insulat-ing crystal that exhibits photorefractive ef-fects. Useful in applications such as multi-beam coupling and phase conjugation.

BTB Seebranch target buffer.

BTC Seeblock truncation coding.


Buck converter.

Buck-boost converter.

BTDF Seebidirectional transmissiondis-tribution function.

BTMA Seebusy tone multiple access.See alsoISMA.

bubble chamber an instrument for ren-dering visible the tracks of ionizing particles.It is characterized by a vessel filled with a su-perheated transparent liquid, commonly hy-drogen or deuterium. The passage of an ion-izing particle through this liquid is marked bythe appearance of a series of bubbles alongthe particle trajectory. If the liquid is sub-jected to a magnetic field, as is usually thecase, the charged particle trajectories will becurved, the curvature providing informationabout the particles’ charge and momentum.

buck converter a transistor is switchedby PWM trigger pulses and a diode providesa current continuation path when the transis-tor is off, thus the input voltage is chopped.A lowpass LC filter is used to attenuate theswitching ripple at the output. The input cur-rent to a basic buck converter is discontinu-ous; therefore, in many applications an LCprefilter is applied to reduce EMI. The out-put voltagevo is related to the input volt-agevi by vo = vid and it can be controlledby varying the duty ratiod. Isolated versionof a buck converter include forward, push-pull, halfbridge, and bridge converters. Alsocalled chopper or step-down converter.

buck-boost converter Seebuck-boosttransformer.

buck-boost transformer a special pur-pose 2- or 4-coil transformer used to producemodest increases or decreases in the utiliza-tion voltage at a load site. The low-voltagecoil(s), which typically have rated voltagesof 5% to 15% of the high-voltage coils, andin use, the high- and low-voltage, coils, areconnected in series to produce an autotrans-former arrangement. If primary voltage is ap-plied to the high voltage coil and load voltageis taken from the series coil combination, thelow-voltage coil adds to, or boosts, the loadutilization voltage. Conversely, reductionsin load utilization voltage occur when theseprimary and secondary connections are re-versed causing the low-voltage coil to buckthe supply voltage. A typical 4-coil buck-boost transformer would have two 120 V pri-mary coils and two 12 V secondary coils,which could be used to produce voltage ra-tios of (120/132), (120/144), (240/252), and(240/264).

In a basic buck-boost converter, the induc-tor accumulates energy from the input volt-age source when the transistor is on and re-leases energy to the output when the transis-tor is off. It can be viewed as a buck converterfollowed by a boost converter with topologicsimplification. In a buck-boost converter, theoutput voltagevo is related to the input volt-


agevi by vo = vid/(1 − d) and it can becontrolled by varying the duty ratiod. Notethat the output voltage is opposite polarity tothe input. Also called a buck-boost converter,up-down transformer or up-down converter.See alsoflyback converter.

bucket a stable phase space area wherethe particle beam may be captured and accel-erated. An RF bucket is the stable region inlongitudinal phase space. The bucket widthgives the maximum phase error or timing er-ror at the RF cavity, which a particle mayhave, and still complete the whole accelera-tion cycle. The bucket height is the corre-sponding limit on momentum error.

bucket truck a motor truck equipped witha shell or bucket at the end of a hydraulically-operated insulated arm. A line worker standsin the bucket and is thus raised to gain accessto overhead conductors.

bucking fields Seedifferentially com-pounded.

buddy memory allocation a memory al-location system based on variable sized seg-ments will usually allocate space for a newsegment from a free area somewhat largerthan necessary, leaving an unallocated frag-ment of the original space. In “buddy” allo-cation, this fragment cannot be used until itsadjacent allocated space is released. Buddyallocation reduces memory fragmentation byensuring that available areas cannot be re-peatedly subdivided.

Buff book SeeIEEE Color Books.

buffer a temporary data storage area inmemory that compensates for the differentspeeds at which different elements are trans-ferred within a system. Buffers are usedwhen data transfer rates and/or data process-ing rates between sender and receiver vary,for instance, a printer buffer, which is neces-sary because the computer sends data to the

printer faster than the data can be physicallyprinted.See alsobuffered input/output.

buffered input/output input/output thattransfers data through a “buffer,” or tempo-rary storage area. The main purpose of thebuffer is to reduce time dependencies of thedata and to decouple input/output from theprogram execution. Data may be preparedor consumed at an irregular rate, whereas thetransfer to or from disk is at a much higherrate, or in a burst.

A buffer is used in “blocked files,” wherethe record size as seen by the user does notmatch the physical record size of the device.

buffering (1) the process of moving datainto or out of buffers or to use buffers todeal with input/output from devices.See alsobuffer, buffered input/output.

(2) in optics, material surrounding theoptical fiber that provides the first layer ofprotection from physical and environmentaldamage. The buffering is usually surroundedby one or more layers of jacketing materialfor additional physical protection of the fiber.

bug (1) an error in a programmed imple-mentation (may be either hardware or soft-ware). Bugs may refer to errors in correct-ness or performance.

(2) a syntactical or logical error in a com-puter program. A name attributed to earlycomputers and electronic testing.

built-in logic block observer techniquethat combines the basic features of scan de-signs, pseudo-random test pattern genera-tion, and test result signature analysis.

built-in self-test (BIST) special hard-ware embedded into a device (VLSI chip ora board) used to perform self testing. On-line BIST assures testing concurrently withnormal operation (e.g., accomplished withcoding or duplication techniques). Off-lineBIST suspends normal operation and is car-ried out using built-in test pattern generator


and test response analyzer (e.g., signature an-alyzer).

bulb generator a free-standing genera-tor contained in a streamlined, waterproofbulb-shaped enclosure and driven by a water-wheel resembling a ship’s propeller on a shaftwhich extends from one end of the enclosure.They are used in tidal power installations.

bulk power a term inclusive of the gener-ation and transmission portions of the powersystem.

bulk scattering scattering at the volumeof an inhomogeneous medium, generally alsopossessing rough boundaries. It is due to in-homogeneities in the refractive index.

bulk substation a substation located ona high-voltage transmission line which sup-plies bulk power to a non-generating utility.

bulldog an attachment for a wire or hoist.

bump a localized orbit displacement cre-ated by vertical or horizontal correction el-ement dipoles used to steer beam throughavailable aperture or around obstacles.

bunch a group of particles captured in aphase space bucket.

bundle the practice of paralleling severalconductors per phase in an overhead trans-mission line for the purpose of increasing am-pacity and decreasing inductive reactance.

bundle spacer a rigid structure which isused to maintain the spacing of wires in abundled conductor on an overhead electricpower transmission lineSeebundle.

bundled services utility services whichare sold together, like power transmissionand distribution services in non-deregulatedelectric utilities.

buried via a via connected to neither theprimary side nor the secondary side of a mul-tilayer packaging and interconnecting struc-ture, i.e., it connects only internal layers.

burndown breakage of an overhead elec-tric power line due to heating from excesscurrent.

burnup a measure (e.g., megawattt-days/ ton) of the amount of energy extracted fromeach unit of fissile material invested in a nu-clear reactor.

burn-in component testing where infantmortality failures (defective or weak parts)are screened out by testing at elevated volt-ages and temperatures for a specified lengthof time.

burst refresh in DRAM, carrying outall required refresh actions in one continu-ous sequence—aburst. See alsodistributedrefresh.

burst transfer the sending of multi-ple related transmissions across an intercon-nect, with only one initialization sequencethat takes place at the beginning of the burst.

burstiness factor used in traffic descrip-tion, the ratio of the peak bit rate to the aver-age bit rate.

bus (1) a data path connecting the dif-ferent subsystems or modules within a com-puter system. A computer system will usu-ally have more than one bus; each bus willbe customized to fit the data transfer needsbetween the modules that it connects.

(2) a conducting system or supply point,usually of large capacity. May be composedof one or more conductors, which may bewires, cables, or metal bars (busbars).

(3) a node in a power system problem(4) a heavy conductor, typically used with

generating and substation equipment.

bus admittance matrix SeeY-bus.


bus acquisition the point at which abus arbiter grants bus access to a specific re-questor.

bus arbiter (1) the unit responsible forchoosing which subsystem will be given con-trol of the bus when two or more requestsfor control of the bus happen simultaneously.Some bus architectures, such as Ethernet, donot require a bus arbiter.

(2) the device that performs bus arbitra-tion. See alsobus arbitration.

bus arbitration the process of determin-ing which competing bus master should begranted control of the bus. The act of choos-ing which subsystem will be given control ofthe bus when two or more requests for controlof the bus happen simultaneously. The ele-ment that make the decision is usually calledthe bus arbiter.See alsobus priority.

bus architecture a computer system ar-chitecture in which one or more buses areused as the communication pathway betweenI/O device controllers, the CPU, and memory.See alsochannel architecture.

bus bandwidth (1) the data transfer ratein bits per second or bytes per second. Insome instances the bandwidth average rateis given and in others the maximum rate isgiven. It is approximately equal to the widthof the data bus, multiplied by the transfer ratein bus data words per second. Thus a 32 bitdata bus, transferring 25 million words persecond (40 ns clock) has a bandwidth of 800Mb/s.

The useful bandwidth may be lowered bythe time to first acquire the bus and possiblytransfer addresses and control information.

(2) the transfer rate that is guaranteed thatno user will exceed.

bus bar a heavy conductor, typically with-out insulation and in the form of a bar of rect-angular cross-section.

bus broadcast Seebroadcast.

bus controller the logic that coordinatesthe operation of a bus.

A device connected to the bus will issuea bus request when it wishes to use the bus.The controller will arbitrate among the cur-rent requests and grant one requester access.The bus controller also monitors possible er-rors, such as use of an improper address, adevice not releasing the bus, and control er-rors.

Bus control logic may reside in multiplesubsystems, distributed control, or may becentralized in a subsystem.See alsobuscycle, bus master.

bus cycle the sequence of steps involvedin a single bus operation. A complete bus cy-cle may require that several commands andacknowledgments are sent between the sub-systems in addition to the actual data that issent.

For example,1. the would-be bus master requests

access to the bus2. the bus controller grants the requester

access to the bus as bus master3. the bus master issues a read command

with the read address4. thebus slaveresponds with data5. the master acknowledges receipt of the

data6. the bus master releases the bus.

The first two steps may be overlapped withthe preceding data transfer.

See alsobus controller, bus master.

bus differential relay a differential relayspecifically designed to protect high powerbuses with multiple inputs.

bus driver the circuits that transmit asignal across a bus.

bus grant an output signal from a pro-cessor indicating that the processor has relin-quished control of the bus to a DMA device.

bus hierarchy a network of busses linkedtogether (usually multiple smaller busses


connected to one or more levels of largerbusses), used to increase the number of el-ements that may be connected to a high-performance bus structure.

bus idle the condition that exists whenthe bus is not in use.

bus impedance matrix SeeZ-bus.

bus interface unit in modern CPU im-plementations, the module within the CPUdirectly responsible for interactions betweenthe CPU and the memory bus.

bus line one of the wires or conductorsthat constitute a bus. A bus line may be usedfor data, address, control, or timing.

bus locking the action of retaining con-trol of a bus after an operation which wouldnormally release the bus at completion. Inthe manipulation of memory locks, a memoryread must be followed by a write to the samelocation with a guarantee of no interveningoperation. The bus must be locked from theinitial read until after the update write to givean indivisible read/write to memory.

bus master a bus device whose requestis granted by the bus controller and therebygains control of the bus for one or more cy-cles or transfers. The bus master may alwaysreside with one subsystem, or may be trans-ferred between subsystems, depending on thearchitecture of the bus control logic.See alsobus controller, bus cycle.

bus owner the entity that has exclusiveaccess to a bus at a given time.

bus phase a term applying especially tosynchronous buses, controlled by a centralclock, with alternating “address” and “data”transfers. A single transfer operation re-quires the two phases to transfer first the ad-dress and then the associated data. Bus ar-bitration may be overlapped with precedingoperations.

bus priority rules for deciding the prece-dence of devices in having bus requests hon-ored.

Devices issue requests on one of severalbus request lines, each with a different buspriority. A high priority request then “wins”over a simultaneous request at a lower prior-ity.

The request grant signals then “daisychains” through successive devices along thebus or is sent directly to devices in appropri-ate order. The requesting device closest tothe bus controller then accepts the grant andblocks its propagation along the bus.

Buses may have handle interrupts and di-rect memory accesses with separate prioritysystems.

bus protocol (1) a set of rules that twoparties use to communicate.

(2) the set of rules that define preciselythe bus signals that have to be asserted by themaster and slave devices in each phase of abus operation.

bus request an input signal to a processorthat requests access to the bus; a hold signal.Competing bus requests are resolved by thebus controller.See alsobus controller.

bus slave a device that responds to a re-quest issued by the bus master.See alsobusmaster.

bus snooping the action of monitoringall traffic on a bus, irrespective of the ad-dress. Bus snooping is required where thereare several caches with the same or overlap-ping address ranges. Each cache must then“snoop” on the bus to check for writes to ad-dresses it holds; conflicting addresses may beupdated or may be purged from the cache.

Bus snooping is also useful as a diagnostictool.

bus state triggering a data acquisitionmode initiated when a specific digital code isselected.


bus tenure the time for which a devicehas control of the bus, so locking out otherrequesters. In most buses, the bus priority ap-plies only when a device completes its tenure;even a low priority device should keep itstenure as short as possible to avoid interfer-ence with higher priority devices.See alsobus priority.

bus transaction the complete sequenceof actions in gaining control of a bus, per-forming some action, and finally releasingthe bus.See alsobus cycle.

bus watching Seebus snooping.

bus width the number of data lines in agiven bus interconnect.

bus-connected reactor Seeshunt reactor.

bus yard an area of a generating station orsubstation in which bus barscf and switchesare located.

Bush, Vannevar (1890–1974) Born: Ev-erett, Massachusetts, U.S.A.

Bush is best know as the developer of earlyelectromechanical analog computers. His“differential analyzer,” as it was called, arosefrom his position as a professor of powerengineering at the Massachusetts Instituteof Technology. Transmission problems in-volved the solution of first- and second-orderdifferential equations. These equations re-quired long and laborious calculations. Hisinterest in mechanical computation arosefrom this problem. Bush’s machines wereused by the military during World War II tocalculate trajectory tables for artillery. Van-nevar Bush was also responsible for inventingthe antecedent of our modern electric meter.He was also scientific advisor to PresidentRoosevelt on the Manhattan Project.

bushing a rigid, hollow cylindrical insu-lator which surrounds a conductor and whichextends through a metal plate such as a the

wall of a transformer tank so as to insulatethe conductor from the wall.

bushing transformer a potential trans-former which is installed in a transformerbushing so as to take advantage of the in-sulating qualities of that bushing.

busway a specialized raceway whichholds un-insulated bus bars in a building.

busy tone multiple access (BTMA) syn-onym for idle tonemultipleaccess.

busy waiting a processor state in whichit is reading a lock and finding it busy, so itrepeats the read until the lock is available,without attempting to divert to another task.The name derives from the fact that the pro-gram is kept busy with this waiting and is notaccomplishing anything else while it waits.The entire “busy loop” may be only 2 or 3instructions.

“Busy waiting” is generally deplored be-cause of the waste of processing facilities.

Butler matrix a feed system (also calledbeam-forming system), that can excite an an-tenna array so that it produces several beams,all offset from each other by a finite angle.The system makes use of a number of inputports connected through a combination of hy-brid junctions and fixed phase shifters.

Butterworth alignment a common filteralignment characterized by a maximally flat,monotonic frequency response.

Butterworth filter an IIR (infinite im-pulse response) lowpass filter with a squaredmagnitude of the form:

|H(ω)|2 = 1

1 + (jωjωc

2N)

buzz stick a tester for insulators, espe-cially strain insulators in a string. It consistsof a pair of probes connected to each side


of a small sphere gap. When the probes aretouched to each terminal of a good insulator,the gap will break down and emit a buzzingsound.

BVGD common notation for FET gate-to-drain reverse breakdown voltage.

BVGS common notation for FET gate-to-source reverse breakdown voltage.

bw common notation for radian bandwithin radians per second.

bwa common notation for fractional arith-metic mean radian bandwidth in radians persecond.

bwg common notation for fractional geo-metric mean radian bandwidth in radians persecond.

BWO Seebackward wave oscillator.

BX cable a flexible, steel-armored cableused in residential and industrial wiring.

bypass Seeforwarding.

bypass switch a manually-operatedswitch used to connect load conductors whenan automatic transfer switch is disconnected.

byte in most computers, the unit of mem-ory addressing and the smallest quantity di-rectly manipulated by instructions. The term“byte” is of doubtful origin, but was usedin some early computers to denote any fieldwithin a word (e.g., DEC PDP-10). Sinceits use on the IBM “Stretch” computer (IBM7030) and especially the IBM System/360 inthe early 1960s, a byte is now generally un-derstood to be 8 bits, although 7 bits is alsoa possibility.

byte multiplexer channel an I/O chan-nel that can be assigned to more than onedata transfer at a time and can be releasedfor another device following each byte trans-fer. (In this regard, it resembles a typicalcomputer bus.) Byte multiplexing is par-ticularly suited to lower speed devices withminimal device buffering. (IBM terminol-ogy) See alsoselector channel, multiplexerchannel.

byte serial a method of data transmissionwhere bits are transmitted in parallel as bytesand the bytes are transmitted serially. For ex-ample, the Centronics-style printer interfaceis byte-serial.


Cc common symbol for speed of light infree space.c = 3 × 1010 cm/s.

CGD common notation for FET gate-to-drain capacitance.

CGS common notation for FET gate-to-source capacitance.

C-band microwave frequency range,3.95-5.85 Ghz.

C-element a circuit used in an asyn-chronous as an interconnect circuit. Thefunction of this circuit is to facilitate thehandshaking communication protocol be-tween two functional blocks.

cable an assembly of insulated conduc-tors, either buried or carried on poles (aerialcable).

cable limiter a cable connector that con-tains a fuse. Cable limiters are used to pro-tect individual conductors that are connectedin parallel on one phase of a circuit.

cable tray a specialized form of racewayused to hold insulated electric power cablesin a building.

cache an intermediate memory store hav-ing storage capacity and access times some-where in between the general register set andmain memory. The cache is usually invisi-ble to the programmer, and its effectivenesscomes from being able to exploit program lo-cality to anticipate memory-access patternsand to hold closer to the CPU: most accessesto main memory can be satisfied by the cache,thus making main memory appear to be fasterthan it actually is.

A hit occurs when a reference can be sat-isfied by the cache; otherwise a miss occurs.The proportion of hits (relative to the totalnumber of memory accesses) is the hit ratio ofthe cache, and the proportion of misses is themiss ratio.See alsocode cache, data cache,direct mapped cache, fully associative cache,set associative cache, and unified cache.

cache aliasing a situation where two ormore entries (typically from different virtualaddresses) in a cache correspond to the sameaddress(es) in main memory. Considered un-desirable, as it may lead to a lack of consis-tency (coherence) when data is written backto main memory.

cache block the number of bytes trans-ferred as one piece when moving data be-tween levels in the cache hierarchy or be-tween main memory and cache). The termline is sometimes used instead of block. Typ-ical block size is 16-128 bytes and typicalcache size is 1-256 KB. The block size ischosen so as to optimize the relationship ofthe “cache miss ratio,” the cache size, and theblock transfer time.

cache coherence the problem of keepingconsistent the values of multiple copies of asingle variable, residing either in main mem-ory and cache in a uniprocessor, or in dif-ferent caches in a multiprocessor computer.In a uniprocessor, the problem may arise ifthe I/O system reads and writes data into themain memory, causing the main memory andcache data to be inconsistent, or if there isaliasing. Old (stale) data could be outputif the CPU has written a newer value in thecache, and this has not been transported tothe memory. Also, if the I/O system has in-put a new value to main memory, new datawould reside in main memory, but not in thecache.

cache hit when the data referenced by theprocessor is already in the cache.


cache line a block of data associated witha cache tag.

cache memory Seecache.

cache miss a reference by the processorto a memory location currently not housed inthe cache.

cache replacement when a “cache miss”occurs, the block containing the accessed lo-cation must be loaded into the cache. If thisis full, an “old” block must be expelled fromthe cache and replaced by the “new” block.The “cache replacement algorithm” decideswhich block should be replaced. An exampleof this is the “Least Recently Used (LRU)”algorithm, which replaces the block that hasgone the longest time without being refer-enced.

cache synonym Seecache aliasing.

cache tag a bit field associated witheach block in the cache. It is used to de-termine where (and if) a referenced block re-sides in the cache. The tags are typicallyhoused in a separate (and even faster) mem-ory (the “tag directory”) which is searchedfor in each memory reference. In this search,the high order bits of the memory addressare associatively compared with the tags todetermine the block location. The numberof bits used in the tag depends on the cacheblock “mapping function” used: “Direct-mapped,” “Fully associative,” or the “Block-set-associative” mapped cache.

CAD Seecomputer-aided design.

cage-rotor induction motor an inductionmotor whose rotor is occupied by copper oraluminum bars, known as rotor bars, insteadof windings. Also commonly referred to asa squirrel-cage induction motor.

calculating board a single-phase scalemodel of a power system that was used to

calculate power flows before the advent ofelectronic computers.

calibration the procedure of character-izing the equipment in place for a particularmeasurement set-up relative to some knownquantity, usually a calibration standard trace-able to the National Institute for Standardsand Technology (NIST).

calibration kits designed for use withvector network analyzers. With these kitsyou can make error-corrected measurementsof devices by measuring known devices(standards) over the frequency range of in-terest. Calibration standards include shorts,open, sliding, and fixed loads.

calibration standards a precision de-vice used in the process of calibrating an EMmeasurement system. It can be a standardgain horn, an open, a short, a load, sphere,etc., used to characterize an RCS, antenna,or transmission line measurement system.Most calibration standards are provided withdocumentation that can be traced to a set ofstandards at the NIST.

call instruction (1) command within acomputer program that instructs the com-puter to go to a subroutine.

(2) an instruction used to enter a subrou-tine. When a call instruction executes, thecurrent program counter is saved on stack,and the address of the subroutine (providedby the call instruction) is used as the new pro-gram counter.

calorimeter a device used to determineparticle energies by measuring the ionizationof a particle shower in a heavy metal, usuallyiron and lead.

CAM acronym for content-addressablememory or computer-aided manufacturing.See associative memory, computer-aidedmanufacturing.


CAMAC acronym for computer auto-mated monitor and control — an internation-ally accepted set of standards for electronicinstrumentation, which specifies mechanical,electrical, and functional characteristics ofthe instrument modules.

camera a device for acquiring an image,usually in a photographic or electronic form— in the latter case typically as a TV camera.Cameras may operate in optical, infra red, orother wavelength bands.

camera calibration a process in whichcertain camera parameters, or equivalentlysome quantities that are required for determi-nation of the perspective projection on an im-age plane of a point in the 3-D world, are cal-culated by using the known correspondencebetween some points in the 3-D world andtheir images in the image plane.

camera model (1) the representation ofthe geometric and physical features of a stere-ovision system, with relative references be-tween the two camera coordinate systems,and absolute references to a fixed coordinatesystem.

(2) a mathematical model by which theperspective projection on an image plane ofa point in the 3-D world can be determined.

can slang for a pole-top distribution trans-former.

candela (cd) unit of measurement for lu-minous intensity (illuminating power in lu-mens/sr). The luminous intensity of 1/60 of1 cm2 of projected area of a blackbody radi-ator operating at the temperature of solidifi-cation of platinum (2046K). Historically, theunit of measurement for the light emitted byone flame of a specified make of candle.

candle Seecandela.

candle power Seecandela.

candlepower distribution a curve, gen-erally polar, representing the variation of lu-minous intensity of a lamp or luminaire in aplane through the light center.

canned magnet a magnet that is com-pletely encased in its own vacuum jacket.

Canny edge detector an edge detectorthat uses an approximation to the optimal fil-ter. Seeinfinite symmetric exponential filter.

Canny operator an edge detector devisedby John Canny as the optimal solution to avariational problem with three constraints.The general solution obtained numericallycan be approximated in practical contexts bythe first derivative of a Gaussian. Canny op-erator usually refers to the extension to twodimensions of this approximation, i.e., to useof a set of oriented operators whose orthog-onal cross sections are a Gaussian and thederivative of a Gaussian. Its advantage is itscapability for allowing edges and their orien-tations to be detected to sub-pixel accuracy.It uses a convolution with a Gaussian to re-duce noise and a derivative to enhance edgesin the resulting smoothed image. The twoare combined into one step — a convolutionwith the derivative of a Gaussian. A hystere-sis thresholding stage is included, to allowclosed contours to remain closed.

CAP Seecarrierless amplitude/phasemodulation.

capability an object that contains both apointer to another object and a set of accesspermissions that specify the modes of accesspermitted to the associated object from a pro-cess that holds the capability.

capability curve Seecapability diagram.

capability diagram also called capabil-ity curve. Graphical representation of thecomplex power limits for safe operation ofa synchronous machine. The vertical axis isaverage powerP and the horizontal axis is


reactive powerQ. The region of allowableoperation is determined by factors such as ro-tor thermal limit, stator thermal limit, ratedpower of prime mover (alternator operation),and stability torque limit.

capability list a list of capabilities, usu-ally associated with a process, defining a setof objects and the modes of access permit-ted to those objects. Computer systems havebeen designed to use capability lists to definethe memory environment for process execu-tion.

capacitance the measure of the electri-cal size of a capacitor, in units of farads.Thus a capacitor with a large capacitancestores more electrons (coulombs of charge)at a given voltage than one with a smallercapacitance.

In a multiconductor system separated bynonconductive mediums, capacitance (C)is the proportionality constant between thecharge (q) on each conductor and the volt-age (V ) between each conductor. The totalequilibrium system charge is zero. Capac-itance is dependent on conductor geometry,conductor spatial relationships, and the ma-terial properties surrounding the conductors.

Capacitors are constructed as two metalsurfaces separated by a nonconducting elec-trolytic material. When a voltage is appliedto the capacitor the electrical charge accumu-lates in the metals on either side of the non-conducting material, negative charge on oneside and positive on the other. If this mate-rial is a fluid then the capacitor is electrolytic;otherwise, it is nonelectrolytic.

capacitance bridge a circuit that includestwo branches which form a balanced drive(two sinusoidal voltage sources connectedin series with common point grounded) andtwo capacitances connected in series betweenfree ends of the voltage sources. The detectorof current (virtual ground of an operationalamplifier is a suitable choice) is connectedbetween the common point of the capacitors

and ground. The circuit finds application incapacitive sensors.

capacitive reactance the opposition of-fered to the flow of an alternating or pulsatingcurrent by capacitance measured in ohms.

capacitively coupled current Seecapacitively coupled field.

capacitively coupled field field applied tothe affected limb by electrodes touching theskin (the current from the electrodes has bothdisplacement and conduction components).

capacitor bank (1) an assembly at onelocation of capacitors and all necessary ac-cessories, such as switching equipment, pro-tective equipment, and controls, required fora complete operating installation.

(2) a group of (typically 3) capacitorsmounted on an electric power line for volt-ageboosting or power factor correction.

capacitor-start induction motor (CSIM)a single-phase induction motor with a ca-

pacitor in series with its auxiliary winding,producing nearly a 90◦ phase difference be-tween the main winding and the auxiliarywinding currents at starting. This results ina high starting torque, so this motor is usedfor hard-to-start loads. The auxiliary wind-ing and capacitor are removed from the cir-cuit by a centrifugal switch as the machineapproaches operating speed.

capacity miss a category of cache missesdenoting the case where the cache is not largeenough to hold all blocks needed during ex-ecution of a program.See alsoconflict miss,cold start miss.

capacity region for a multiple termi-nal communications system. The entire setof rate-vectors for which there exist chan-nel codes such that the probability of mak-ing a decoding error can be made arbitrar-


ily small. See alsoachievable rate region,multiple access channel.

capture effect a phenomenon found inpacket switched networks in which nonequalpowers in packet radio networks using con-tention protocols lead to higher throughputs.In contention protocols used in packet ra-dio networks, the transmitted packets are al-lowed to collide. If two packets collide andone is significantly stronger in power, thispacket is more likely to be captured (de-tected) by the receiver.

capture range the range of input frequen-cies over which the PLL can acquire phaselock.

capture register internal register which,triggered by a specified internal or externalsignal, store or “capture” the contents of aninternal timer or counter.

carbon brush a block of carbon used tomake an electrical contact to a rotating coilvia the commutator of a DC machine or theslip rings of a synchronous machine.

carbon dioxide (CO2) linear gasmolecule consisting of one carbon and twooxygen atoms, medium for an important classof lasers.

carbon dioxide laser laser in which theamplifiying medium is carbon dioxide gas;efficient, powerful, and commercially impor-tant laser that is pumped and configured inmany ways and has its principal output linesin the mid-infrared.

carbon resistor thermometer a carbonresistor whose temperature sensitivity pro-vides good temperature resolution.

carcinotron a forward radial travelingwave amplifier in which microwave signalsare fed to the radial slow wave structure.

card a printed circuit board that can beplugged into a main board to enhance thefunctionality or memory of a computer.

card cage mechanical device for holdingcircuit cards into a backplane.

cardinal series the formula by whichsamples of a bandlimited signal are interpo-lated to form a continuous time signal.

cardinal vowel according to English pho-netician Daniel Jones, a vowel correspondingto one of the extreme positions of the voweldiagram.

carrier amplitude amplitude of the ra-dio frequency sinusoid used as a vehicle fortransporting intelligence from the sendingend of a communications link to the receiv-ing end. For an AM, FM, or PM wave, thepeak amplitude of the spectral component inthe frequency domain about which symme-try exists. The carrier amplitude (as a func-tion of time) contains a portion of the intelli-gence for angle modulation (Seefrequencymodulationandphase modulation). In con-trast, the carrier amplitude contains no in-formation for AM or any of the SSB varia-tions (Seeamplitude modulationandsinglesideband modulation), but is merely used asa frequency marker.

carrier concentration the number of mo-bile charge carriers per unit volume, positive(holes) or negative (electrons). In a semicon-ductor, both concentrations are present andare modifiable by externally applied electricfields.

carrier current communication the useof electric lines to carry communication sig-nals.

carrier frequency in pulse-width-modu-lated (PWM) switching schemes, the switch-ing frequency that establishes the frequencyat which the converter switches are switched.In sine-triangle PWM, the carrier frequency


is the frequency of the triangle waveform thatthe control or modulating signal is comparedto.

carrier lifetime the average duration anelectron or a hole stays in a certain state.

carrier phase the phase of a sinusoidalsignal that is the carrier in a modulationscheme such as AM, FM, SSB, etc. The car-rier may be defined in the formA cos(ωct +φ). The carrier is specified by the parametersA (amplitude),ωc (carrier frequency), andφ(carrier phase).

carrier shift the difference in frequencybetween the steady state, mark, and space infrequency shift keying (FSK) systems.

carrier signal the RF signal in a com-munications system that has the modulatingsignal superimposed on it. This signal mayhave its frequency, amplitude, or phase variedto form a modulated signal. Without modu-lation it is a simple RF signal.

Many communication systems rely on theconcept of sinusoidal amplitude modulation,in which a complex exponential signal c(t)has its amplitude multiplied (modulated) bythe information-bearing signalx(t). Thissignalx(t) is typically referred to as the mod-ulating signal and the signalc(t)as the carriersignal. The modulated signaly(t) is then theproduct of these two signals.

y(t) = x(t)c(t)

carrier suppression in SSB communi-cations, the degree to which the carrier am-plitude is reduced from its original valueout of the modulator. (See alsobalancedmodulator. ) Carrier suppression is gen-erally used as a method to significantly re-duce the amount of unnecessary transmittedpower, based upon the fact that no informa-tion is contained within the carrier amplitudein an AM waveform. It is sometimes desir-able to only partly suppress the carrier, leav-

ing what is termed a “pilot tone” at the carrierfrequency.

carrier synchronization a synchroniza-tion technique used in radio receivers. In allradio receivers some sort of carrier frequencysynchronization is required; the phase syn-chronization is needed only if phase coherentdemodulation is desired. Can be categorizedas open and closed loop carrier synchroniza-tion.

carrier-sense multiple access (CSMA) arandom-access method of sharing a bus-typecommunications medium in which a poten-tial user of the medium listens before begin-ning to transmit. The channel sensing signif-icantly reduces the probability of collisions.Compare withALOHA.

carrier-to-interference ratio (CIR) sim-ilar to signal-to-interference ratio but usu-ally used in cellular communication systemswhere the carrier refers to the signal of inter-est and the interference refers to interferencefrom other transmitters in the system.Seealsosignal-to-interference ratio.

carrier-to-noise ratio the ratio of the am-plitude of the carrier signal to that of the noisein the IF bandwidth measured at any pointin the receiver before any nonlinear processsuch as amplitude limiting or detection. Thecarrier to noise ratio is typically expressed indecibels.

carrierless amplitude/phase modulation(CAP) an implementation of a quadratureamplitude modulation transmitter in whichthe passband in-phase and quadrature signalsare generated directly via quadrature digi-tal filters. A recent application for CAP ishigh-speed digital subscriber lines.See alsoquadrature amplitude modulation.

carry overflow signal that occurs whenthe sum of the operands at the inputs of theadder equals the base. A binary adder, adding1+ 1 will produce a sum of 0 and carry of 1.


carry bit Seecarry.

carry flag Seecarry.

carry look-ahead adder high-speedadder that uses extra combinational logic togenerate all carries in an m-bit block in paral-lel. A method of generating the signals corre-sponding to the carries (borrows) in an addi-tion (subtraction) circuit that does not requireall the lower order carries to be determined;a high-speed carry.

Cartesian-based control the system de-picted in the figure. Notice that inverse kine-matics is embedded into the feedback controlloop. Due to the inverse kinematics calcula-tions, Cartesian-based control is of greatercomputational complexity. HereXd , Xd ,andXd denote position, velocity, and accel-eration of the desired trajectory in Cartesianspace.τ is a vector of generalized forces andq is a vector of generalized positions.

Cartesian-based control scheme.

Cartesian product a mathematical oper-ation on two sets. The Cartesian product oftwo sets, sayA andB, denotedA×B, is theset of all ordered pairs with the first elementof each pair an element ofA and the other anelement ofB. That is,

A× B = {(a, b) | a ∈ A andb ∈ B}

Cartesian space Seeexternal space.

cartridge fuse replaceable electricalsafety device in which metal melts and in-terrupts the circuit when the current exceedsa preset limit in duration and magnitude.

cascade connection a series connectionof amplifier stages or networks in which theoutput of one feeds the input of the next.

cascade system a 3-level systemcontaining high-, intermediate-, and low-energy states. Resembling a cascade,these states are coupled, in that sequence,by two electromagnetic fields. See alsocathodoluminescent.

cascode a circuit technique in which thecurrent output of the collector (drain) of aBJT (FET) is buffered by a common base(common gate) amplifier stage. The purposeis to increase the bandwidth and/or outputresistance.

cascode amplifier an amplifier consistingof a grounded-emitter input stage that drivesa grounded-base output stage; advantages in-clude high gain and low noise; widely usedin television tuners.See alsocascode.

CASE See computer-aided softwareengineering.

castellation recessed metallized featureon the edges of a chip carrier that interconnectconducting surfaces or planes within or on thechip carrier.

casual filter a filter of which the transitionfrom the passband to the stopband is gradual,not ideal. This filter is realizable.

catadioptric an optical system made upof both refractive elements (lenses) and re-flective elements (mirrors).

catastrophic code a convolutional codein which a finite number of code symbol er-rors can cause an unlimited number of de-coded symbol errors.

catastrophic encoder a convolutionalencoder with at least one loop in the state-transition diagram with zero accumulatedcode symbol weight, at least one nonzero in-


formation symbol and not visiting the zerostate. After decoding, a finite number of(channel) errors can result in an infinite num-ber of errors (catastrophic error propagation).

catastrophic error propagation whenthe state diagram contains a zero distancepath from some nonzero state back to thesame state, the transmission of a 1 causesan infinite number of errors. See alsocatastrophic encoder.

catastrophic thermal failure an immedi-ate, thermally induced total loss of electronicfunction by a component or system.

catcher a cavity resonator of a multi-cavity klystron proximate to the collector tocatch microwave energy from the bunchedelectrons.

categoric input a nonnumeric (symbolic)input, e.g., gender, color, which is usually fedto a network using one-out-of-N coding.

catenation symbols strung together toform a larger sequence, as the characters in aword and the digits in a number.

cathode the negative electrode of a de-vice. Contrast withanode.

cathode ray tube (CRT) a vacuum tubeusing cathode rays to generate a picture on afluorescent screen. These cathode rays are infact the electron beam deflected and modu-lated, which impinges on a phosphor screento generate a picture according to a repeti-tive pattern refreshed at a frequency usuallybetween 25 and 72 Hz.

cathodoluminescent the property of lu-minescent crystals (phosphors) to emit visi-ble light with bombarded electrons.

catoptric an optical system made up ofonly reflective elements (mirrors).

CATV Seecommunity-antenna television.

Cauchy distribution the density functionfor a Cauchy distributed random variableXis

fX(x) = 1

1 + x2

Note that the moments for this random vari-able do not exist, and that the cumulative dis-tribution function is not defined.See alsoprobability density function, moment.

Caurer filter Seeelliptic filter.

causal system a system whose outputdoes not depend on future input; the outputat timet may depend only on the input sig-nal {f (τ) : τ ≤ t}. For example, the volt-age measured across a particular element in apassive electric circuit does not depend uponfuture inputs applied to the circuit, and henceis a causal system.

If a system is not causal, then it is non-causal. An ideal filter which will filter in realtime all frequencies present in a signalf (t)

requires knowledge of{f (τ) : τ > t}, and isan example of a noncausal system.

causality a systemH : Xe → Xe, orequivalently, an operator that maps inputsfrom the extended spaceXe into outputs fromthe same space where the output at timet

is not a function of future inputs. This canbe expressed using truncations as follows: AsystemH is causal if

[Hx(·)]T = [HxT (·)]T ∀x ∈ XeSee alsoextended spaceandtruncation.

cavity (1) a fully enclosed, hollow con-ductor, in which only time-harmonic electro-magnetic fields of specific frequencies (i.e.,resonant frequencies) exist. Each resonantfrequency is identified by a collection ofnumbers in conjunction with a mode designa-tor of the transverse electric, transverse mag-netic, or transverse electromagnetic type.


(2) in optics, region of space that is par-tially or totally enclosed by reflecting bound-aries and that therefore supports oscillationmodes.

cavity dumping fast removal of energystored in a laser cavity by switching the effec-tive transmission of an output coupling mir-ror from a low value to a high value.

cavity lifetime one of several names usedto indicate the time after which the energydensity of an electromagnetic field distribu-tion in a passive cavity maybe expected tofall to 1/e of its initial value; the name pho-ton lifetime is also common.

cavity ratio (CR) a number indicatingcavity proportions calculated from length,width, and height. It is further defined intoceiling cavity ratio, floor cavity ratio, androom cavity ratio.

cavity short a grounded metal rod con-necting the body of an RF cavity. By ground-ing the cavity, it is kept from resonating.

Cayley–Hamilton theorem for 2-D generalmodel let Tpq be transition matrices de-fined by

ETpq =

A0T−1,−1 + A1T0,−1 + A2T−1,0+In for p = q = 0A0Tp−1,q−1 + A1Tp,q−1+A2Tp−1,qf or p 6= 0 and/orq 6= 0

and

[Ez1z2 − A0 − A1z1 − A2z2]−1

=∞∑

p=−n1

∞∑q=−n2

Tpqz−(p+1)1 z

−(q+1)2

Exi+1,j+1 = A0xij + A1xi+1,j + A2xi,j+1

+ B0uij + B1ui+1,j + B2ui,j+1

i, j ∈ Z+ (the set of nonnegative integers)where xi,j ∈ Rn is the semistate vector,ui,j ∈ Rm is the input, andE, Ak, Bk(k = 0,1,2) real matrices withE possibly

singular or rectangular. A pair(n1, n2) ofpositive integersn1, n2 such thatTpq = 0for p < −n1 and/orq < −n2 is called theindex of the model. Transition matricesTpqof the generalized 2-D model satisfy

n1∑p=0

n2∑q=0

dpqTp−k−1,q−t−1 = 0

for {k < 0 andm1 < k ≤ 2n1 − 1t < 0 andm2 < t ≤ 2n2 − 1

wheredpq are coefficients of the polynomial

det [Ez1z2 − A0 − A1z1 − A2z2]

=n1∑p=0

n2∑q=0

dpqzp

1zq

2

andm1, m2 are defined by the adjoint matrix

adj [Ez1z2 − A0 − A1z1 − A2z2]

=m1∑i=0

m2∑j=0

Hij zi1zj

2

(m1 ≤ n− 1, m2 ≤ n− 1)

Cayley–Hamilton theorem for 2-D Roessermodel letTij be transition matrix definedby

Tij =

I (the identity matrix)for i, j = 0

T10 :=[A1 A20 0

], T01 :=

[0 0A3 A4

]T10Ti−1,j + T01Ti,j−1 for i, j ∈ Z+0 for i < 0 or/andj < 0

(Z+ is the set of nonnegative integers) of the2-D of the Roesser model[xhi+1,jxvi,j+1

]=[A1 A2A3 A4

][xhijxvij

]+[B1B2

]uij

i, j ∈ Z+ wherexhij ∈ Rn1 andxvij ∈ Rn2

are the horizontal and vertical state vectors,respectively,uij ∈ Rm is the input vector,andA1, A2, A3, A4, B1, B2 are real matri-ces. The transition matricesTij satisfy the


equation

n1∑i=0

n2∑j=0

aij Ti+h,j+k = 0

for h = 0,1, . . . and k = 0,1, . . . whereaij are coefficients of the 2-D characteristicpolynomial

det

[In1z1 − A1 −A2

−A3 In2z2 − A4

]=

n1∑i=0

n2∑j=0

aij zi1zj

2

(an1n2 = 1

)

CB SeeCitizen’s band.

CBE Seechemical beam epitaxy.

CBR Seeconstant bit rate.

CCD Seecharge-coupled device.

CCD memory Seecharge-coupled-devicememory.

CCDF Seecomplementary cumulativedistribution function.

CCI Seecochannel interference.

CCIR SeeInternational Radio ConsultativeCommittee.

CCITT SeeComite Consulatif InternationalTelegraphique et Telephonique.

CCITT two-dimensional a modifiedrelative element address designate scheme.The position of each changing element onthe present line is coded with respect to thechanging element on the reference line or thepreceding changing element on the presentline. The reference line lies immediatelyabove the present line.

CCR Seecochannel reuse ratio.

CCS Seecommon channel signaling.

CCVT Seecoupling capacitor voltagetransformer.

CD Seecompact disk, critical dimension.

CD Seecritical dimension.

cd Seecandela.

CD-I Seecompact disk-interactive.

CD-ROM a read-only compact disk.Seealsocompact disk.

CDF Seecumulative distribution function.

CDMA See code division multipleaccess.

CdS abbreviation for cadmium sulfide,a photoconductor with good visible light re-sponse.

CEL Seecontrast enhancement layer.

CELL Seesurface-emitting laser logic.

cell (1) in mobile radio communica-tions, the area serviced by one base sta-tion. One way of categorizing the cells isaccording to their size. Cell sizes may rangefrom a few meters to many hundred kilome-ters. See alsopicocell, nanocell, nodal cell,microcell, macro cell, large cell, megacell,satellite cell.

(2) in ATM systems a small packet of fixedlength. The CCITT chose a cell size of 53-bytes comprising a 5-byte header and 48-bytepayload for their ATM network.

cell library a collection of simple logicelements that have been designed in accor-dance with a specific set of design rules andfabrication processes. Interconnections ofsuch logic elements are often used in semi-custom design of more complex IC chips.

cell switching means of switching dataamong the ports (inputs and outputs) of a


switch such that the data is transferred in unitsof a fixed size.

cell-cycle-specific control a type of con-trol arising in scheduling optimal treatmentprotocols in the case when treated popula-tion is sensitive to the therapy only in cho-sen phases of its cell cycle. A mathematicalmodel of population dynamics used to solvethe problem of such control should be com-posed of subsystems sensitive and insensitiveto the drug. It may be achieved by the useof compartmental models. Typical types ofperturbations of the cell cycle considered ascel-cycle-specific control are as follows: cellarrest, cell killing, and alteration of the transittime. They could be applied for representa-tion of cell synchronization, cell destruction,and cell recruitment from the specific phase.

cellular automaton system designedfrom many discrete cells, usually assem-bled in one- or two-dimensional regular ar-rays, each of which is a standard finitestate machine. Each cell may change itsstate only at fixed, regular intervals, andonly in accordance with fixed rules that de-pend on cells’ own values and the values ofneighbors within a certain proximity (usuallytwo- for one-dimensional, and four- for two-dimensional cellular automata). Cellular au-tomata are a base of cellular computers; finegrain systems that are usually data-driven andused to implement neural networks, systolicarrays, and SIMD architectures.

cellular communications traditionally,an outside-of-building radio telephone sys-tem that allows users to communicate fromtheir car or from their portable telephone.

cellular manufacturing grouping ofparts by design and/or processing similaritiessuch that the group (family) is manufacturedon a subset of machines which constitute acell necessary for the group’s production.

cellular spectral efficiency the cellularspectral efficiency of a system is defined as

the sum of the maximum data rates that can bedelivered to subscribers affiliated to all basestations in a re-use cluster of cells, occupyingas small a physical area as possible. Mathe-matically, the cellular spectral efficiency,η,is defined as

η =∑rj=1

∑Ki=1Rij

BAclusterbit/s/Hz/km2

wherer denotes the number of cells in a re-use cluster,Rij denotes the data rate mea-sured in bits/s at some predefined BER avail-able to subscriberI in cell j of the re-usecluster,B denotes the total bandwidth mea-sured in hertz allocated to all cells in the re-use cluster, andAcluster denotes the physicalarea, measured in square kilometers occupiedby the re-use cluster.

CELP Seecode excited linear prediction.

center frequency (1) the frequency ofmaximum or minimum response for a band-pass or a bandstop filter, respectively; oftentaken as the geometric mean of the lower andupper cutoff frequencies.

(2) the frequency at the center of a spec-trum display.

(3) the average frequency of the emittedwave after modulation by a sinusoidal signal.

(4) the frequency of a non-modulatedwave.See alsochannel.

center of average an approach of defuzzi-fication that takes the weighted average of thecenters of fuzzy sets with the weights equalto the firing strengths of the correspondingfuzzy sets.

center of gravity method Seecentroidmethod.

center of projection the point within aprojector from which all the light rays ap-pear to diverge; the point in a camera towardwhich all the light rays appear to converge


before they cross the imaging plane or pho-tographic plate.

central absolute moment for randomvariablex, thepth central absolute momentis given byE[|x − E[x]|p. See alsocentralmoment, absolute moment, expectation.

central limit theorem (CLT) in proba-bility, the theorem that the density function ofsome function ofn independent random vari-ables tends towards a normal curve asn tendsto infinity, as long as the variances of the vari-ables are bounded: 0< σ ≤ vi ≤ γ < ∞.Hereσ andγ are positive constants, andvi isthe variance of theith random variable.SeealsoGaussian distribution.

central moment for random variableXthenth central moment is given by

E[(X −m)2] =∫ ∞

−∞(x −m)2fX(x)dx

wherefX(x) is the probability density func-tion ofX. See alsocentral absolute moment,absolute moment. See alsoexpectation.

central processing unit (CPU) a partof a computer that performs the actual dataprocessing operations and controls the wholecomputing system. It is subdivided into twomajor parts:

1. The arithmetic and logic unit (ALU),which performs all arithmetic, logic, andother processing operations,

2.The control unit (CU), which sequencesthe order of execution of instructions, fetchesthe instructions from memory, decodes theinstructions, and issues control signals to allother parts of the computing system. Thesecontrol signals activate the operations per-formed by the system.

centralized arbitration a bus arbitrationscheme in which a central bus arbiter (typ-ically housed in the CPU) accepts requestsfor and gives grants to any connected device,wishing to transmit data on the bus. The con-nected devices typically have different prior-

ities for bus access, so if more than one de-vice wants bus access simultaneously, the onewith the highest priority will get it first. Thisprioritization is handled by the bus arbiter.

centrifugal switch a speed-sensitiveswitch operated by centrifugal force, mountedon the shaft of a motor, used to perform a cir-cuit switching function. Used in single-phaseinduction motors to disconnect the startingwinding when the motor approaches operat-ing speed.

centripetal force force that is presentduring the robot motion. The force dependsupon the square of the joint velocities of therobot and tend to reduce the power availablefrom the actuators.

centroid (1) a region in the pattern spaceto which a remarkable number of patterns be-long.

(2) the center of a mass.(3) description of the center of a particle

beam profile.

centroid defuzzification a defuzzifica-tion scheme that builds the weighted sum ofthe peak values of fuzzy subsets with respectto the firing degree of each fuzzy subset. Alsocalled height defuzzification.

centroid method a widely used methodof defuzzification whereby the centroid of themembership function of the fuzzy set is usedas the defuzzified or crisp value. It is alsoknown as the center of gravity method or thecomposite moments method.

centroidal profile a method for charac-terizing and analyzing the shape of an objecthaving a well defined boundary. The cen-troid of the shape is first determined. Thena polar(r, θ) plot of the boundary is com-puted relative to this origin: this plot is thecentroidal profile, and has the advantage ofpermitting template matching for a 2-D shapeto be performed relatively efficiently as a 1-Dprocess.


centrosymmetric medium a materialthat possesses a center of inversion symme-try. Of importance because, for example,second-order nonlinear optical processes areforbidden in such a material.

cepstrum inverse Fourier transform of thelogarithm of the Fourier power spectrum of asignal. The complex cepstrum is the inverseFourier transform of the complex logarithmof the Fourier transform of the signal.

ceramic ferrite a relatively inexpensivepermanent magnet material with decent co-ercivity and low energy product that is com-posed of strontium or barium oxide and ironoxide. Also called hard ferrite.

cerebellar model articulation (CMAC)network a feedforward network devel-oped originally as a model of the mam-malian cerebellum. Several variants nowexist, but basic operation involves the firstlayer of the network mapping the input into ahigher-dimensional vector and a second layerforming the network output by means of aweighted sum of the first layer outputs. Theweights can be trained using the LMS rule.Developed mainly for application in robotics,it has also been used in pattern recognitionand signal processing. Often called a CMACnetwork.

Cerenkov counter a detector for chargedparticles. It consists essentially of a trans-parent medium such as a gas, which emitsCerenkov radiation when a charged particlepasses through at a velocity greater than thevelocity of light in the medium. The mass ofa particle in a beam of known momentum canbe determined with such a counter by mea-suring the characteristic angle at which theCerenkov radiation is emitted.

Cerenkov radiation light emitted whena charged particle traverses a medium with avelocity greater than the velocity of light inthe medium. The Cerenkov light is emittedin a cone centered on the particle trajectory.

The opening angle of this cone depends onthe velocity of the particle and on the veloc-ity of light in the medium. The phenomenoninvolved is that of an electromagnetic shockwave and is the optical analogue of sonicboom. Cerenkov radiation provides an im-portant tool for particle detection.

certainty equivalence principle a designmethod in which the uncertainties of processparameters are not considered. Found in self-tuning regulators where the controller param-eters or the process parameters are estimatedin real-time and are then used to design thecontroller as if they were equal to the true pa-rameters. Although many estimation meth-ods could provide estimates of parameter un-certainties, these are typically not used in thecontrol design.

CFD Seecrossed field devices.

CFIE See combined field integralequation.

CGA Seecolor graphics adapter.

chain code a method for coding thin con-tours or lines, for example, in a bilevel pic-ture, which encodes the direction of move-ment from one point to the next. For 8-connected contours, a three-bit code may beused at each point to indicate which of itseight neighbors is the succeeding point.

chain matrix SeeABCD matrix.

chain parameters SeeABCD parameters.

chain reaction a process in which high-energy neutrons emitted from fissile radioac-tive material are directed into more fissilematerial such that more neutrons are emit-ted. The process creates heat which is usedto power thermal power plants.

chaining when the output stream of onearithmetic pipeline is fed directly into another


arithmetic pipeline; used in vector computersto improve their performance.

chaining of fuzzy rules a reasoning strat-egy which searches the knowledge base andchain from rule to rule to form inferences anddraw conclusions. In forward chaining, achain of data-driven rules are evaluated forwhich the conditional parts are satisfied toarrive at the conclusion. Backward chain-ing is goal-driven in which subgoals are es-tablished, where necessary, through which achain of rules are selected, eventually satis-fying the goal.

chamfer distance a digital distance basedon a chamfer mask, which gives the distancebetween a pixel and those in its neighbor-hood; then the chamfer distance between twonon-neighboring pixels (resp., voxels) is thesmallest weighted length of a digital pathjoining them. The word “chamfer” comesfrom the fact that with such a distance a circleis in fact a polygon. Then-dimensional Man-hattan and chessboard distances are chamferdistances; the Euclidean distance is not. Inthe 2-D plane, the best chamfer distances aregiven by the(3,4) and (5,7,11) Chamfermasks: in the(3,4) mask, a pixel is at dis-tance 3 from its horizontal/vertical neighborsand at distance 4 from its diagonal neigh-bors, while in the(5,7,11) mask, it is atdistance 5 from its horizontal/vertical neigh-bors, at distance 7 from its diagonal neigh-bors, and at distance 11 from its neighborsdistant by 1 and 2 respectively along the twoaxes. Seechessboard distance, Euclideandistance, Manhattan distance.

channel (1) the medium along which datatravel between the transmitter and receiver ina communication system. This could be awire, coaxial cable, free space, etc.See alsoI/O channel.

(2) the conductivity path between thesource and the drain of a field effect tran-sistor.

(3) a single path for transmitting electri-cal signals. Example 1: The band of fre-

quencies from 50 Hz to 15 KHz (Channel A)and 15 KHz to 75 KHz (Channel B) whichfrequency modulate the main carrier of anFM stereo transmitter. Example 2: A portionof the electromagnetic spectrum assigned foroperation of a specific carrier from the FMbroadcast band (88 to 108 MHz) of frequen-cies 200 KHz wide designated by the centerfrequency beginning at 88.1 MHz and con-tinuing in successive steps to 107.9 MHz.

channel allocation the act of allocatingradio channels to cells, base stations, or cellsectors, in a radio network, also referred toas frequency allocation, or frequency plan-ning. The allocation typically follows an al-gorithm that attempts to maximize the num-ber of channels used per cell and minimizethe interference in the network.

channel architecture a computer sys-tem architecture in which I/O operations arehandled by one or more separate processorsknown as channel subsystems. Each chan-nel subsystem is itself made up of subchan-nels, in which control unit modules controlindividual I/O devices. Developed by IBM,and used primarily in mainframe systems, thechannel architecture is capable of a very highvolume of I/O operations.

channel capacity a fundamental limit onthe rate at which information can be reliablycommunicated through the channel. Also re-ferred to as “Shannon capacity,” after ClaudeShannon, who first formulated the concept ofchannel capacity as part of the noisy channelcoding theorem.

For an ideal bandlimited channel with ad-ditive white Gaussian noise, and an input av-erage power constraint, the channel capacityisC = 0.5 log(1+S/N) bit/Hz, whereS/Nis the received signal-to-noise ratio.

channel code a set of codewords used torepresent messages, introducing redundancyin order to provide protection against errorsintroduced by transmission over a channel.See alsosource code.


channel coding the process of intro-ducing controlled redundancy into an infor-mation sequence mainly to achieve reliabletransmission over a noisy channel. Channelcoding can be divided into the areas of blockcoding and trellis coding. Also called errorcontrol coding.See alsoblock coding, trelliscodingandconvolutional coding.

channel command word an “instruc-tion” to an I/O channel. The commands con-sists of parameters (e.g., “operation,” “dataaddress,” “count”) giving the channel pro-cessor information on type of I/O operationrequested (e.g., “read” or “write”), where thedata is to be read or written, and the numberof bytes involved in the data transfer.

In the IBM mainframe architecture thereare six different types of channel con-trol words: READ, READ BACKWARD,WRITE, CONTROL, SENSE, and JUMP.

channel control word See channelcommand word.

channel encoder a device that convertssource-encoded digital information into ananalog RF signal for transmission. The typeof modulation used depends on the particu-lar digital audio broadcasting (DAB) system,although most modulation techniques em-ploy methods by which the transmitted sig-nal can be made more resistant to frequency-selective signal fading and multipath distor-tion effects.

channel estimation estimation of the ra-dio channel parameters in the receiver. Typi-cally delays, amplitudes, carrier phases, anddirection-of-arrivals need to be estimated de-pending on the receiver configuration. Chan-nel estimation is a modern way to look at re-ceiver synchronization based mainly on feed-back control loops, since in principle anymethod known to estimation theory can beapplied to achieve synchronization of the re-ceiver over an unknown radio channel.

channel I/O an approach to I/O process-ing in which I/O operations are processed in-dependent from the CPU by a channel sys-tem. See alsochannel architecture.

channel matched VQ See channeloptimized vector quantization.

channel measurement See channelsounding.

channel modeling the act of describingthe effect of the (radio) channel on the trans-mitted signal in a form suitable for mathemat-ical analysis, computer simulation or hard-ware simulation. A channel model is a col-lection of channel characteristics essential tothe performance of the communication sys-tem under study, organized in such a way thatthe basic performance trade-offs of the sys-tem can be analyzed or simulated withouthaving to construct prototypes or carry outfield tests.

channel optimized vector quantization(COVQ) a combined source-channelcode for block-based source coding (vectorquantization) and block channel coding. Achannel optimized vector quantizer can bedesigned using a modified version (takingchannel induced distortion into account) ofthe generalized Lloyd algorithm). Also re-ferred to as channel matched VQ.See alsonoisy channel vector quantization.

channel program the set of channel con-trol words that make up the instruction se-quence that controls an I/O channel.See alsochannel control word.

channel reliability function the ratefunction with infinitesimal error probabilityexpressed by

E(R) ={ 1

2C∞ 0 ≤ R ≤ 14C∞(√

C∞ − √R)2

14C∞ ≤ R ≤ C∞

for transmission of orthogonal or simplexsignal over infinite bandwidth AWGN chan-


nel. C∞ is the capacity of the infinite band-width white Gaussian noise channel, definedas

C∞ = Pav

Noln2bit/s .

channel robust vector quantizationa vector quantizer that has been made ro-bust against channel errors.See alsonoisychannel VQ.

channel robust VQ Seechannel robustvector quantization.

channel sounding the act of recordingfrom a real, physical channel a selected set ofcharacteristics describing the channel. Thisinvolves transmitting a known signal to thechannel under study, and at the receiver pro-cessing the received signal to extract the re-quired information, typically the impulse re-sponse or the transfer function of the channel.

channel spill leakage of RF energy froma radio channeln into a radio channeln± i,i ≥ 1, due to finite channel filter attenuationoutside of the bandwidth ofn.

channel step Seefrequency synthesizer.

channel subsystem the I/O processingcomponent of a computer system conformingto the channel architecture model.

channel waveguide a light guide that iseither raised above or diffused into a sub-strate.

channel-to-case thermal resistance theproportionality constant (denotedθcc) at ther-mal equilibrium between the temperature dif-ference of the FET channel (Tchannel) and aspecified case surface (Tcase) to the dissipatedpower in the channel (Pw), in units of◦C/W .The specified surface is usually the most ef-fective in reducing the temperature. It in-cludes the thermal resistance of the chip, dieattach material (solder or adhesive), packag-

ing and mounting medium, as applicable.

θcc = Tchannel− Tcase

Pw

∣∣∣∣equilibrium

channelizer a system that decomposes anRF signal into narrow-band output channels;term often applied to acousto-optic spectrumanalyzers that are driven by RF frequencysignals. See alsoacousto-optic spectrumanalyzer.

chaos (1) erratic and unpredictable dy-namic behavior of a deterministic system thatnever repeats itself. Necessary conditions fora system to exhibit such behavior are that it benonlinear and have at least three independentdynamic variables.

(2) in microelectronics, deterministic mo-tion, in which the statistics are essentiallythose of a Gaussian random process.

chaotic behavior a highly nonlinear statein which the observed behavior is very depen-dent on the precise conditions that initiatedthe behavior. The behavior can be repeated(i.e., it is not random), but a seemly insignif-icant change, such as voltage, current, noise,temperature, rise times, etc., will result indramatically different results, leading to un-predictability. The behavior may be chaoticunder all conditions, or it may be well be-haved (linear to moderately nonlinear) un-til some parametric threshold is exceeded, atwhich time chaotic behavior is observed. In amildly chaotic system, noticeable deviationsresulting from small changes in the initialconditions may not appear for several cyclesor for relatively long periods. In a highlychaotic system, the deviations are immedi-ately apparent.

character (1) letter, number or symbolas used on a computer keyboard.

(2) data type that represents an alphanu-meric character as a group of bits, usually asan eight-bit byte.


character recognition Seeoptical characterrecognition.

character string (1) a series of continu-ous bytes in memory, where each byte repre-sents one character.

(2) data structure corresponding to or-dered sequence of characters.

characteristic equation the polynomialequation that results when the characteristicfunction is equated to zero. Its roots gives thesingularities of the transfer function model,which in turn determine its transient behav-ior. Specifically, any root of the character-istic equation that has a negative real partindicates a stable decaying transient, whileany root with a positive real part indicates anunstable growing transient. Any root withzero real part indicates a marginally stabletransient that neither decays nor grows. Theimaginary part of the root gives the frequencyof oscillation of the transient signal.See alsocharacteristic function.

characteristic function (1) the namegiven to the denominator polynomial of atransfer function model. Through partialfraction expansion of a transfer function andsubsequent inverse Laplace transformation,it is obvious that the characteristics of thesystem dynamics are defined by this func-tion. For example, the transfer function

g(s) = 9

6 + 5s + s2

has characteristic function

φ(s) = 6 + 5s + s2 = (s + 2)(s + 3)

so its output response will contain terms like

y(t) = αe−2t + βe−3t + . . .

that are characteristic of the system itself.(Other terms in the response are attributedto the forcing input signal.) See alsocharacteristic equation.

(2) a transformed probability densityfunction,

8x(ω) = E[exp(jωT x)

]

useful in the analytic computation of higherorder moments and convolutions of probabil-ity densities.

characteristic impedance inherent prop-erty of a transmission line that defines theimpedance that would be seen by a signalif the transmission line were infinitely long.If a signal source with a “source” or “refer-ence” impedance equal to the characteristicimpedance is connected to the line there willbe zero reflections.

characteristic loci the plots of the eigen-values of transfer function matrices, eval-uated over a range of frequencies. Thesetraces, which are parametrized by frequency,are shown on a single Nyquist plot andused to predict the closed loop stability ofmultiinput-multioutput systems, by applica-tion of the principle of the argument for com-plex variable functions. Unlike the Nyquistplots for single-input-single-output systems,an individual eigenvalue might not encirclethe plane an integral number of times, yet thetotal encirclements of all the eigenvalues willbe an integral number.

characteristic polynomial and equation ofgeneralized 2-D model the determinant

p (z1, z2)

= det[Ez1z2 − A0 − A1z1 − A2z2]

=n1∑i=0

n2∑j=0

aij zi1zj

2

(n1, n2 ≤ rank E) is called the 2-D char-acteristic polynomial of the generalized 2-Dmodel



i, j ∈ Z+ (the set of nonnegative integers)wherexij ∈ Rn is the semistate vector,uij ∈Rm is the input vector, andE,Ak, Bk (k =0,1,2) are real matrices withE possibly sin-gular or rectangular.


p(z1, z2) = 0 is called the 2-D character-istic equation of the model.

characteristic polynomial assignment of2-D Roesser model consider the 2-DRoesser model[xhi+1,jxvi,j+1

]=[A1 A2A3 A4

][xhijxvij

]+[B1B2

]uij

i, j ∈ Z+ (the set of nonnegative integers)with the state-feedback

uij = K

[xhijxvij

]+ vij

wherexhij ∈ Rn1, andxvij ∈ Rn2 are the hori-zontal and vertical state vectors, respectively,uij ∈ Rm is the input vector, andA1,A2,A3,A4, B1, B2 are real matrices of the model,

K = [K1,K2] ∈ Rm×(n1+n2)

andvij ∈ Rm is a new input vector. Given themodel and a desired 2-D characteristic poly-nomial of the closed-loop systempc(z1, z2),find a gain feedback matrixK such that

det

[In1z1 − A1 − B1K1 −A2 − B1K2

−A3 − B2K1 In2z2 − A4 − B2K2

]

= pc (z1, z2) =n1∑i=0

n2∑j=0

dij zi1zj

2

(dn1n2 = 1

)

characteristic polynomial of 2-D Fornasini–Marchesini model the determinant

p (z1, z2) = det[Inz1z2 − A1z1 − Az2z2

]=

n1∑i=0

n2∑j=0

aij zi1zj

2 (ann = 1)

is called the 2-D characteristic polynomial ofthe 2-D Fornasini–Marchesini model

xi+1,j+1 = A1xi+1,j + A2xi,j+1

+ B1ui+1,j + B2ui,j+1

i, j ∈ Z+ (the set of nonnegative integers)wherexij ∈ Rn is the local state vector,uij ∈

Rm is the input vector, andAk,Bk (k = 1,2)are real matrices.p(z1, z2) = 0 is called the 2-D character-

istic equation of the model.

characteristic polynomial of 2-D Roessermodel the determinant

p (z1, z2) = det

[In1z1 − A1 −A2

−A3 In2z2 − A4

]=

n1∑i=0

n2∑j=0

aij zi1zj

2

(an1n2 = 1

)is called the 2-D characteristic polynomial ofthe 2-D Roesser model[xhi+1,jxvi,j+1

]=[A1 A2A3 A4

][xhijxvij

]+[B1B2

]uij

i, j ∈ Z+ (the set of nonnegative integers)wherexhij ∈ Rn1, andxvij ∈ Rn2 are the hori-zontal and vertical state vectors, respectively,uij ∈ Rm is the input vector, andA1,A2,A3,A4, B1, B2 are real matrices.p(z1, z2) = 0 is called the 2-D character-

istic equation of the model.

characterization the process of cal-ibrating test equipment, measuring, de-embedding and evaluating a component orcircuit for DC RF and/or digital performance.

charge a basic physical quantity that is asource of electromagnetic fields.

charge carrier a unit of electrical chargethat when moving, produces current flow. Ina semiconductor two types of charge carri-ers exist: electrons and holes. Electronscarry unit negative charge and have an ef-fective mass that is determined by the shapeof the conduction band in energy-momentumspace. The effective mass of an electron in asemiconductor is generally significantly lessthan an electron in free space. Holes haveunit positive charge. Holes have an effectivemass that is determined by the shape of thevalence band in energy-momentum space.The effective mass of a hole is generally sig-nificantly larger than that for an electron. For


this reason, electrons generally move muchfaster than holes when an electric field is ap-plied to the semiconductor.

charge conservation physical law (de-rived from Maxwell’s equations) indicatingthat no change in the total charge within acertain volume can exist without the properflow of charge (current) through that volume.

charge density describes the distributionof charge along a line, on a surface or in avolume. May be discrete or continuous.

charge-coupled device (CCD) a solid-state device used to record images. A CCD isa digital device which counts the photons thatstrike it by making use of the photoelectriceffect. In a typical CCD array, a large numberof such devices is collected into a 2-D grid.Each device corresponds to a single pixel,and the number of electrons in the device islinearly related to the brightness or intensityvalue at that point in the CCD.

charge-coupled device detector acharge-coupled device (CCD) connected tophotodetectors, where the photocharge is putinto the CCD potential wells for transport andprocessing.

charge-coupled-device memory large-capacity shift registers making use of charge-coupled devices (CCD), i.e., MOS devicesin which data bits are stored dynamically ascharge between a gate and the substrate. Thisforms a multigate MOS transistors with thesource and drain terminals “stretched” apart,and a number of gate terminals in between.The first gate terminal (closest to the source)inserts bits (charge) into the register, andthe following gates are controlled with over-lapping clocks allowing the charge to movealong the array. At the far (drain) end, the bitunder the final gate terminal is detected as achange in current.

charge-spring model See electronoscillator model.

charging current that portion of an elec-tric power line’s current which goes to chargethe capacitance of the line. The charging cur-rent is not available for power transmission.

chattering fast switching. The termcomes from the noise generated by the sus-tained rapid opening and closing of a switch-ing element.See alsodiscontinuous control.

Chattuck coil a finely wound solenoidabout a flexible, nonmagnetic core that isusually used in conjunction with a fluxmeterto measure magnetic potential between twopoints; a magnetic analog of a voltmeter.

CHDL Seecomputer hardware descrip-tion language.

Chebyshev alignment a common filteralignment characterized by ripples of equalamplitude within the pass-band and a steeprolloff in the vicinity of cutoff frequency.

Chebyshev filter one of a class of com-monly used low pass, high pass, band passand band stop filters with an equiripplecharacteristic, designed to achieve relativelyrapid rolloff rates (See alsoButterworthfilter) near cutoff frequencies, at the expenseof a loss of monotonicity in either the pass-bands or the stopbands.

checkerboarding Seefragmentation.

checkpoint time in the history of exe-cution at which a consistent version of thesystem’s state is saved so that if a later eventcauses potential difficulties, the system canbe restarted from the state that had been savedat the checkpoint. Checkpoints are impor-tant for the reliability of a distributed sys-tem, since timing problems or message losscan create a need to “backup” to a previousstate that has to be consistent in order for theoverall system to operate functionally.

checkpointing method used in rollbacktechniques in which some subset of the sys-


tem states (data, program, etc.) is saved atspecific points (checkpoints), during the pro-cess execution, to be used for recovery if afault is detected.

checksum checksum is a value used todetermine if a block of data has changed. Thechecksum is formed by adding all of the datavalues in the block together, and then find-ing the 2’s complement of the sum. Thechecksum value is added to the end of thedata block. When the data block is examined(possibly after being received over a serialline), the sum of the data values and check-sum should be zero.

checksum character in data communi-cation and storage devices, an extra characteris often added at the end of the data so that thetotal number of ones in a block, including thechecksum character is even. The checksumcharacter is used to detect errors within thedata block.

chemical beam epitaxy (CBE) a mate-rial growth technique that uses metal organicmolecules in high vacuum growth cham-ber and a controlled chemical reaction on aheated substrate to grow a variety of II-VI,III-V, and group IV materials with atomiclayer control. Used to create material struc-tures for a variety of electronic and opticaldevices using quantum wells, heterostruc-tures, and superlattices. This growth tech-nique combined aspects of both MBE andMOCVD growth.

chemical laser a laser in which the am-plification results from one or more chemi-cal reactions; potentially very powerful withprincipal output lines in the mid-infrared.

chemical sensor the interface device foran instrumentation system that determinesthe concentration of a chemical substance.

chemical vapor deposition (CVD) a pro-cess used in the manufacture of integratedcircuits or optical fibers whereby a thin solid

film of one material is deposited on the sur-face of another by using a radio frequency orother electrical energy source to dissociate areactive gas.

chemically amplified resist a type ofphotoresist, most commonly used for deep-UV lithography, which, upon post-exposurebake, will multiply the number of chemicalreactions through the use of chemical catal-ysis.

chemiluminescence light emitted as aresult of a chemical reaction.

Chernobyl typically refers to a fire at anuclear power plant near Kiev in the Republicof the Ukraine.

chessboard distance the distance be-tween discrete points arising from theL∞ norm. Given two discrete pointsx = (x1, . . . , xn), y = (y1, . . . , yn)

on an n-dimensional integer lattice, thechessboard distancebetweenx and y ismax{|x1 − y1|, . . . , |xn − yn|}. So calledbecause it equals the number of moves madeby a King when going from one position toanother in the game of chess.Seenorm.

chi-squared distribution a probabilitydistribution withn degrees of freedom andprobability density function

f (x) = xn2−1e− x

2u(x)

2n20(n2)

.

chip (1) a small piece of semiconductormaterial upon which miniaturized electroniccircuits can be built.

(2) an individual MMIC circuit or subsys-tem that is one of several identical chips thatare produced after dicing up an MMIC wafer.

(3) in direct-sequence spread-spectrumtransmission, the high bandwidth symbols,or pulses making up the signature sequence.They are used to spread the bandwidth of thedata in frequency. Usually the time duration


of these pulses are many times smaller thanthat of the information symbols leading tosignificantly greater spreading of the signalbandwidth.

chip carrier a low-profile rectangularcomponent package, usually square, whosesemiconductor chip cavity or mounting areais a large fraction of the package size andwhose external connections are usually on allfour sides of the package.

chip chart this term is often used forthe “gray scale” chart used in the process ofaligning television camera systems. The grayscale provides logarithmic reflectance rela-tionships.

chip select a control signal input to, e.g.,a memory chip, used to make this particularchip “active” in reading or writing the databus. Read or write is determined by anothercontrol input signal: the “R/W-signal.” Typ-ically, some of the high order bits from theCPU’s address bus are decoded to form thechip select signals.

chip-to-chip optical interconnect opti-cal interconnect in which the source and thedetector are connected to electronic elementsin two separate chips.

chirp the varying in time of a carrier fre-quency signal.See alsochirp function.

chirp function a signal whose frequencyvaries monotonically with time, e.g., a lin-ear chirp possesses a linear-frequency or aquadratic-phase variation.

chirp signal Seechirp function.

chirping a shifting of the optical fre-quency often observed in modulated semi-conductor lasers where the laser gain is mod-ulated at high bandwidth; arises due to thelater portions of the modulating signal seeinga different refractive index, or carrier density,than the earlier portions.

Cholesky decomposition a matrix-alge-braic theorem that states that, for any positivedefinite square matrixA, there exists a lower-left triangular matrixG such thatA = G GT .

chopper Seebuck converterand DCchopper.

chopper - depth of modulation a markernormally associated with the monitoring ofthe depth of modulation of a television broad-cast signal on a waveform monitor. Thechopper reference is used to set the 0% mod-ulation point relative to the video signal. Thesync signal is typically at the 100% modula-tion level.

chroma the portion of the video sig-nal defining the color information in the im-age. The chroma signal is defined by changesin the 3.579545 Mhz interlaced sinewave.Phase changes create changes in color, peak-to-peak changes in the sinewave alter the sat-uration of the color while changes in the DClevel of the chroma signal alter the luminance(brightness).

chromatic aberration (1) beam spread-ing due to different momentum of the par-ticles that are being bent by the quadrupolefields at different angles.

(2) the failure of a lens to simultaneouslyfocus all colors of light. It arises since therefractive index of a material depends on thewavelength of light.

chromaticity (1) the ratio of tune spreadto momentum spread of the beam. Chro-maticity affects the focusing and bendingproperties of magnets by making them sen-sitive to particle momentum. This results infocusing and bending dispersion of the beamin a manner analogous to an optical system.

(2) specification of color stimuli. Thechromaticity coordinates are relative RGBvalues correlated with hue and saturation.


chrominance (1) the color informationin the video signal that is defined in terms ofhue and saturation.

(2) the component of color which is in-dependent of, and complementary tolumi-nance; chrominanceis 2-D: for example, itcan be decomposed intohueandsaturation.Seehue, intensity, luminance, saturation.

chronaxie the minimum duration of aunidirectional square-wave current needed toexcite a nerve when the current magnitude istwice rheobase.

CIE See Commision Internationald’Eclairage.

CIE diagram the projection of the plane(X+ Y +Z) = 1 onto theXY plane, whereX, Y,Z are the respective tristimulus valuesas defined by the CIE (Seetristimulus valueand Commision International d’Eclairage).The CIE diagram shows all of the visiblechromaticity values and maps all colors withthe same chromaticity but different value (lu-minances) onto the same points.

CIM Seecomputer-integrated manufac-turing.

CIR Seecarrier-to-interference ratio.

circle detection the location of circles inan image by a computer. Often accomplishedwith the Hough transform.

circle diagram (1) graphical representa-tion of the operation of an induction machine.It is based on the approximate equivalent cir-cuit and expresses stator and rotor current re-lations for all operating modes (motor, brak-ing, generator) and all values of slip. Severalvariations of the diagram exist.

(2) graphical representation of the powerflow through a transmission line. The maxi-mum power flow through the line can be de-termined by the impedance of the line.

circuit a physical device consisting ofan interconnection of elements, or a topolog-ical model of such a device. For example,an electric circuit may be constructed by in-terconnecting a resistor and a capacitor to avoltage source. A representation of this cir-cuit is shown by the diagram in the figure.

Circuit example.

circuit (STM) switching technology thatprovides a direct connection between twoendpoints; data is transferred directly be-tween the endpoints of a circuit without beingstored in any intermediate nodes.

circuit breaker a circuit breaker is a de-vice that makes and breaks the electrical con-tact between its input and output terminals.The circuit breaker is capable of clearing faultcurrents (tripping) as well as load currents.The circuit breaker consists of power con-tacts with arc clearing capability and associ-ated control and auxiliary circuits for closingand tripping the breaker under the requiredconditions.

circuit protection devices or controlmeasures used to safeguard electrical circuitsfrom unsafe operating regions, such as over-currents and overvoltages.

circuit switching a method of commu-nication in which a physical circuit is estab-lished between two terminating equipmentsbefore communication begins to take place.This is analogous to an ordinary phone call.


circuit-set a closed path where all verticesare of degree 2, thus having no endpoints inthe path.

circuit-switched service a telecommu-nications service, where the communicationsresource is retained for the whole period ofcommunication. For example wired tele-phone services or mobile telephone servicesin the first- and second-generation systemscan be classified to circuit-switched services.

circulant matrix a squareN × N

matrix M = {mi,j } such thatmi,j =m(i+n)modN,(j+n)modN ; that is, that eachrow of M equals the previous row rotatedone element to the right. All circulant matri-ces are diagonalized by the discrete Fouriertransform.

circular cavity a section of the circularwaveguide closed at both ends by conductingplates.

circular convolution Seeperiodic convolu-tion.

circular mil the area of a circle whichmeasures 0.001 inch in diameter.

Used to specify the cross-sectional area ofa wire.

circular polarization a polarization stateof a radiated electromagnetic field in whichthe tip of the electric field vector traces acircle as a function of time for a fixed po-sition. The sense of rotation of the electricfield vector is either right-hand or left-hand(clockwise or counter-clockwise).

circular register buffer a set of generalpurpose CPU registers organized to providea large number of registers, which may be ac-cessed a few at a time. The group of registersaccessible at any particular time may be read-ily changed by incrementing or decrementinga pointer, with wraparound occurring fromthe highest numbered registers to the lowestnumbered registers, hence the name circular

register buffer. There is overlap between thegroups of adjoining registers that are accessi-ble when switching occurs. The overlappingregisters can be used for passing argumentsduring subroutine calls and returns. The cir-cular register buffer is a feature of the SPARCCPU architecture. In the SPARC CPU thereare 256 registers, available 32 at a time, withan overlap of eight registers above and eightregisters below the current group.

circular self-test path a BIST techniquebased on pseudorandom testing assured byarranging flip-flops of a circuit (during test)in a circular register in which each flip-flopoutput is ex-ored with some circuit signaland feeds the input of the subsequent flip-flop. This register simultaneously providestest pattern generation and test result com-paction.

circularity measure the size invariant ra-tio of area divided by perimeter squared forsmall shapes and much used as a preliminarydiscriminant or measure of shape, so-calledbecause it is a maximum for circular objects.

circularly polarized light light com-posed of two orthogonal polarizations thatare 90 degrees out of phase; the resultant lightamplitude vector thus rotates about the direc-tion of propagation at the optical frequency.

circulator a multiport nonreciprocal de-vice that has the property that an electromag-netic wave incident in port 1 is coupled to port2 only, an electromagnetic wave incident inport 2 is coupled into port 3 only and so on.

CIRF Seecochannel interference reduc-tion factor.

CISC processor Seecomplex instructionset computer.

Citizen’s band (cb) 40 channels wherethe carrier frequency is between 26.965 MHzand 27.405 MHz established by the FCC for


short-distance personal or business commu-nication.

city-block distance a distance measurebetween two real valued vectors(x1, x2, . . . ,

xn) and(y1, y2, . . . , yn) defined as

Dcity block =n∑i=1

|xi − yi |

City-block distance is a special case ofMinkowski distance whenλ = 1. See alsoMinkowski distance. Also called Manhat-tan distance.

cladding the optical material that concen-trically surrounds the fiber core and providesoptical insulation and protection for the core.The refractive index of the cladding must belower than that of the core material so thatoptical power is guided through the fiber bytotal internal reflection at the core-claddingboundary.See alsototal internal reflection,Snell’s Law.

clamping the process of fixing either theminima or maxima of a voltage.

Clapp oscillator an oscillator whose fre-quency is determined by a tuned parallel LCcircuit with a split capacitance, i.e., two seriescapacitances, in the capacitive branch and anadditional series tuning capacitance in the in-ductive branch. The Clapp oscillator is a vari-ation of the Colpitts oscillator.

class (1) in general, patterns are com-monly discriminated into different categoriesaccording to certain properties they share.The categories in which a given set of patternsare partitioned are referred to as classes.

(2) in object orientation, is an entity thatdefines a set of objects which share the sameattributes and processes.

class fuse SeeUL classes.

class A amplifier an amplifier in whichthe active device acts as a modulated cur-rent source biased midway between satura-

tion and conduction cutoff. In a class A am-plifier, as the amplitude of an applied sinu-soidal signal is increased, the output will startto clip at both ends simultaneously. This isequivalent to a conduction angle of 360 de-grees as long as the output signal is not clip-ping, which is avoided. This term is oftenused to include any amplifier operating withsignal levels low enough such that signal clip-ping is not present (i.e., small signal condi-tions).

class A-B amplifier most current sourceamplifiers fall into this category, which in-cludes all amplifiers biased somewhere be-tween class A and class B. As the amplitudeof a sinusoidal signal is increased, the outputwill start to cut off first. Further increaseswill cause clipping due to saturation. Thusthe conduction angle is between 180 and 360degrees, dependent on applied signal ampli-tude. Device saturation is usually avoided.

class B amplifier an amplifier in whichthe active device acts as a modulated cur-rent source biased at conduction cutoff. Ina class B amplifier, an applied sinusoidal sig-nal will result in only half of the sinusoidbeing amplified, while the remaining half iscut off. Further increases in the signal am-plitude will eventually cause the remaininghalf of the signal to saturate and clip, whichis usually avoided. This is equivalent to aconduction angle of 180 degrees, regardlessof signal amplitude.

class B-D amplifier switched mode am-plifier where the device is biased at cutoff,and the input signal is large enough to drivethe amplifier into heavy saturation such thatonly a small percentage of time is spent intransition. The amplifier is literally switchedbetween cutoff and saturation, and thus thesaturation angle is a significant percentageof the conduction angle, which is 180 de-grees. The unfiltered, broadband output cur-rent waveform of a class B-D amplifier re-sembles a stepped squarewave. It is impor-tant to note that only frequency related infor-


mation (FM) is preserved in a class B-D am-plifier, while all amplitude information (AM)is lost. Usually, class B-D power amplifiersare designed in a push-pull configuration totake advantage of both halves of a cycle.

class B-E amplifier transient switchedmode amplifier where the device is biasedat cutoff, the input signal is large enoughto drive the amplifier into heavy saturationsuch that only a small percentage of time isspent in transition, and the design is suchthat during saturation the waveform is de-termined by the switch circuit transient re-sponse, while the waveform during cutoff isdetermined by the transient response to theentire circuit, including the load. The ampli-fier is literally switched between cutoff andsaturation, the transient responses are wellcontrolled, and thus the saturation angles ap-proach the conduction angle, which is 180degrees. The final tuned output current waveform of a class B-E amplifier resembles anideal squarewave. It is important to note thatonly frequency related information (FM) ispreserved in a class B-E amplifier, while allamplitude information (AM) is lost.

class C amplifier a current source ampli-fier biased beyond the conduction cutoff suchthat operation will not begin until the inputsignal reaches a specific amplitude, and re-sults in less than half of an input sinusoidbeing amplified. If the signal amplitude isincreased sufficiently, saturation and the as-sociated clipping will occur. Thus the con-duction angle is between 0 and 180 degrees,regardless of amplitude. Device saturation isusually avoided.

class D amplifier switched mode ampli-fier where the device is biased somewherebetween class A and class B cutoff, and theinput signal is large enough to drive the am-plifier from cut-off to heavy saturation suchthat only a small percentage of time is spent intransition. The amplifier is literally switchedbetween cutoff and saturation, and thus thesaturation angle is a significant percentage

of the conduction angle, which is 180 de-grees. The unfiltered, broadband output cur-rent waveform of a class D amplifier resem-bles a stepped squarewave. It is importantto note that only frequency related informa-tion (FM) is preserved in a class D ampli-fier, while all amplitude information (AM) islost. Usually, class D power amplifiers aredesigned in a push-pull configuration to takeadvantage of both halves of a cycle.

class E amplifier a transient switchedmode amplifier where the device is biasedsomewhere between class A and class B cut-off, the input signal is large enough to drivethe amplifier into heavy saturation such thatonly a small percentage of time is spent intransition, and the design is such that dur-ing saturation the waveform is determined bythe switch circuit transient response, whilethe waveform during cutoff is determined bythe transient response to the entire circuit,including the load. The amplifier is literallyswitched between cutoff and saturation, thetransient responses are well controlled, andthus the saturation angles approach the con-duction angle, which is 180 degrees. The fi-nal tuned output current waveform of a classE amplifier resembles an ideal squarewave.It is important to note that only frequency re-lated information (FM) is preserved in a classE amplifier, while all amplitude information(AM) is lost.

class E-F amplifier a harmonic tuned orharmonic reaction amplifier (HRA) in whichdevices, biased for class B operation, arearranged in a push/pull configuration, andare utilized to inject each other with largeharmonic currents in order to modulate theamplitude of the fundamental output currentthrough the device, resulting in improvedswitching efficiency. The even order har-monics must be shorted at the output, whilethe odd order harmonics must be provided anopen at the output.

class F amplifier a high-efficiency opera-tion in amplifiers. The class F amplifier has a


load impedance optimized not only for a fun-damental wave but also for harmonic wavesto improve efficiency. An efficiency of theclass F amplifier is 100% under an ideal con-dition, where the optimum load impedancefor even harmonic waves is short and that forodd harmonic waves is open.

class G amplifier a frequency multiply-ing or harmonic amplifier biased somewherebetween class A and class C, in which theinput is tuned to the fundamental input fre-quency and the output is tuned to a frequencymultiple of the input.

class H amplifier frequency mixing am-plifier biased somewhere between class Aand class C, in which the inputs are tuned tothe input frequencies and the output is tunedto either the sum or difference frequency.

class S amplifier sampling or pulse widthmodulation amplifier in which a samplingcircuit (or pulse width modulator), pulse am-plifier and a low pass filter are cascaded.The input signal is sampled at a significantlyhigher rate than the input frequency (this re-quires a high frequency sampling signal),and the original signal is transformed intoa constant amplitude pulse chain in whichthe pulse widths are proportional to the orig-inal signal’s amplitude. The resulting pulsechain is amplified using any of the highly effi-cient switching methods desired. The outputis then demodulated using a low pass filter,replicating the original signal. It is impor-tant to note that rapid variations in the inputsignal amplitude relative to the sampling sig-nal will cause significant distortion or loss ofinformation.

classified VQ See classified vectorquantization.

classified vector quantization (CVQ)a vector quantization technique where dif-ferent codebooks are developed based imageedge features. The codebook used to encodea particular block is determined by a clas-

sifier with differentiating capability betweenthe types of features. A number of codebooksare developed each to encode blocks of pix-els containing specific types of features.Seealsovector quantization.

classifier (1) given a set of patterns ofdifferent classes, a classifier is a system ca-pable of determining the membership of eachpattern.

(2) a method of assigning an object to oneof a number of predetermined classes.

clean cache block a cache block (or“line”) is clean if it is a copy of the informa-tion stored in memory. A clean block can beoverwritten with another block without anyneed to save its state in memory.

clear (1) to set the value of a storagelocation to zero (often used in the context offlip-flops or latches).

(2) clearing a bit (register) means writinga zero in a bit (register) location. Opposite to“set.”

clearing time the total time required tomelt and clear, and thus totally open, a fuse-type overcurrent device.

cleaved coupled cavity semiconductorlaser configuration in which the amplifyingregion has been cleaved to introduce a mid-cavity reflecting boundary; added reflector isintended to improve mode-selectivity char-acteristics.

click noise in a fading channel, the noiseassociated with a threshold crossing. In afading radio channel situation, the movinguser crosses the standing wave patterns inthe propagation environment. As the usercrosses the minima, the service quality tem-porarily downgrades and, in analogue sys-tems, is noticeable as clicking.

climbers two metal spikes, each of whichis strapped to the inside of a line worker’slegs, pointing down near the ankle. Plunged


into the sides of a wooden utility pole, theyprovide purchase for the worker to scale thepole.

clipping nonlinear distortion that occurswhen the input to an amplifier exceeds theamplifier’s linear range. The amplifier out-put saturates at its limit, giving a “clipped”appearance to the output waveform.

clock (1) the oscillator circuit that gener-ates a periodic synchronization signal.

(2) a circuit that produces a series of elec-trical pulses at regular intervals that can beused for timing or synchronization purposes.

clock cycle one complete event of a syn-chronous system’s timer, including both thehigh and low periods.

clock doubling a technique in which theprocessor operates internally at double theexternal clock frequency.

clock duty cycle the percentage of timethat the electronic signal remains in the trueor 1 state.

clock pulse a digital signal that, via itsrising edge or falling edge, triggers a digi-tal circuit. Flip-flops and counters typicallyrequire clock pulses to change state.

clock recovery in synchronous systems,the act of extracting the system clock signalfrom the received sequence of informationsymbols.See alsosymbol synchronization.

clock replacement algorithm a page re-placement algorithm described as follows: Acircular list of page entries corresponding tothe pages in the memory is formed. Each en-try has a use bit which is set to a 1 when thecorresponding page has been referenced. Apointer identifies a page entry. If the use bitof the page entry is set to a 1, the use bit is re-set to a 0 and the pointer advances to the nextentry. The process is repeated until an entryis found with its use bit already reset, which

identifies the page to be replaced. The pointeradvances to the next page entry for the nextoccasion that the algorithm is required. Theword “clock” comes from viewing the pointeras an arm of a clock. Also known as a first-in-not-used-first-out replacement algorithm.

clock skew the phenomenon where differ-ent parts of the circuit receive the same stateof clock signal at different times because ittravels in wires with different lengths. Thisskew of the signals causes a processing ele-ment to generate an erroneous output. Distri-bution of the clock by means of optical fibers,waveguides, a lens, or a hologram, eliminatesclock skew.

clock speed the rate at which the tim-ing circuit in a synchronous system generatestiming events.

closed convex set a set of vectorsC suchthat ofx, y ∈ C thenλx + (1 − λ)y ∈ C forall 0 ≤ λ ≤ 1.

closed kinematic chain in vision engi-neering, a sequence of links which forms aloop.

closed-loop control control action achievedby a closed feedback loop, i.e., by measuringthe degree to which actual system responseconforms to desired system response and ap-plying the difference to the system input todrive the system into conformance.

closed-loop DC motor acceleration theuse of sensors to provide feedback to the mo-tor control circuit indicating the motor is ac-tually accelerating before the starting resis-tors are removed from the armature circuit.Two popular methods to sense motor accel-eration are CEMF coils, and current sensingcoils.

closed-loop gain the gain of an opera-tional amplifier circuit with negative feed-back applied (with the negative feedbackloop “closed”).


closed-loop optimal control operationor structure of the controller with the deci-sion mechanism which, under uncertainty,uses in the best possible way — with respectto a givena priori criterion — all informa-tion available to the controller; in particular,the closed-loop optimal control takes into ac-count all future time instants at which newdecisions will be made; the best known ex-ample of a closed-loop optimal control rule isthe solution of the linear-quadratic-Gaussianproblem (LQG problem).

closed-loop system any system havingtwo separate paths inside it. The first pathconducts the signal flow from the input of thatsystem to the output of that same system. Thesecond path conducts the signal flow from theoutput to the input of the system, thus estab-lishing a feedback loop for the system. Seethe figure below for a general description of aclosed-loop system. The forward and back-

A closed-loop system.

ward signal path construct the closed loop,which conduct the signal flow from the inputstage to the output stage and then back to theinput stage of that system.

closing a basic morphological operation.Given a structuring elementB, the closingbyB is the composition of thedilation byBfollowed by the erosion byB; it transformsX intoX • B = (X ⊕ B) B. The closingby B is what one calls an algebraic closing;this means that:(a) it is a morphologicalfilter; (b) it is extensive, in other words it can

only increase an object.Seedilation, erosion,morphological filter, structuring element.

closing/opening filter one of an impor-tant class of morphological filters. Letγandϕ be opening and closing operators re-spectively. The following operators can beobtained by composingγ andϕ (i.e., by ap-plying them in succession):γ ϕ, ϕγ , γ ϕγ ,andϕγϕ. These are all morphological fil-ters, and collectively they are called clos-ing/opening filters or opening/closing filters.No further operators can be obtained by com-posingγ andϕ. Seeclosing, morphologicalfilter, opening.

clothes pin slang for a wood or plastic clipused to secure a blanket to conductors.

CLT Seecentral limit theorem.

cluster a group of data points on a spaceor a group of communicating computer ma-chines. A cluster of computers on a localnetwork can be installed to provide their ser-vice as a unique computer. This is frequentlyused for building large data storage and Webservers. In computer disks, a cluster consistsof a fixed number of sectors. Each sectorcontains several bytes, for example 512.

cluster analysis in pattern recognition,the unsupervised analysis of samples to clus-ter them into classes based on (a) a distancemetric and (b) a clustering algorithm. Typ-ical algorithms minimize a cluster criterion(e.g., representation error) by grouping sam-ples hierarchically or by iteratively reassign-ing samples to clusters. The K-means algo-rithm is an example of the latter. In the case of2-D measurements, cluster analysis becomesa method of image segmentation.

clustering (1) any algorithm that createsthe major clusters from a given set of patterns.

(2) a method of unsupervised learning thataims to discover useful structure in unlabeleddata by grouping similar patterns.


Seehierarchical clustering. See alsodistance measure, similarity measure.

clutter the name given to backgroundsignals which are currently irrelevant to a de-tection system; clutter is a form of structurednoise.

CMA Seeconstant modulus algorithm.

CMAC network Seecerebellar modelarticulation network.

CMMR Seecommon mode rejectionratio.

CMOS Seecomplementary metal oxidesemiconductor.

CO abbreviation for central switchingoffice.

CO2 Seecarbon dioxide.

CO2 laser Seecarbon dioxide laser.

co-linear array a phased array of straightelements in which the axes of the elements liealong a common line. The elements are typ-ically center fed half-wave dipoles or foldeddipoles.

co-occurrence matrix an array of num-bers that relates the measured statistical de-pendency of pixel pairs. Co-occurrence ma-trices are used in image processing to identifythe textural features of an image.

co-prime polynomials polynomials thathave no common factors. For example, poly-nomials(s2+9s+20)and(s2+7s+6)are co-prime, while(s2+5s+6) and(s2+9s+14)are not, since they have a common(s + 2)factor.

co-tunneling a cooperative process,whereby electrons can tunnel through two se-ries connected tunneling barriers. In this pro-cess, the tunneling of an electron through oneof the barriers causes a self-consistent shift of

potential onto the second barrier, which re-sults in the immediate inducing of a particletunneling through this latter barrier.

coarticulation the transient process cor-responding to the utterance of two phonemes.It is due to the movement of the articulatoryorgans between the different positions corre-sponding to the two phonemes.

coax Seecoaxial cable.

coaxial cable A transmission line formedby two concentric conductors separated by adielectric designed to confine the fields andtheir energy in the medium between said con-ductors. It is often used in applications wheresignal interference between the cable and itssurrounds must be kept to a minimum. Alsocalled coax.

coaxial magnetron a radial magnetronwhere the anode and cathode are graduallytransformed into a coaxial line.

cochannel interference (CCI) interfer-ence caused by radio transmitters operatingon the same radio frequency as that of a par-ticular wanted radio frequency signal.

cochannel interference reduction factor(CIRF) a key factor used to design a cel-lular system to avoid the cochannel interfer-ence.

cochannel reuse ratio (CRR) the reuseratio between radio communication cells us-ing the same radio channels.

cochannels radio channels occupying thesame radio frequency allocationn.

cochlea a snail-shaped passage commu-nicating with the middle ear via the round andoval window. Its operation consists of trans-ducing the acoustical vibration to nerve im-pulses, subsequently processed in the brain.


Cockroft–Walton circuit a cascadingvoltage multiplier invented in 1932 by JohnCockroft and Ernest Walton.

code (1) a technique for representing in-formation in a form suitable for storage ortransmission.

(2) a mapping from a set of messages intobinary strings.

code acquisition the process of ini-tial code synchronization (delay estimation)between the transmitter and receiver in aspread-spectrum system before the actualdata transmission starts. It usually requiresthe transmission of a known sequence.Seealsocode tracking.

code cache a cache that only holds in-structions of a program (not data). Codecaches generally do not need a write policy,but see self-modifying code. Also called aninstruction cache.See alsocache.

code combining an error control codetechnique in which several independently re-ceived estimates of the same codeword arecombined with the codeword to form a newcodeword of a lower rate code, thus providingmore powerful error correcting capabilities.This is used in some retransmission protocolsto increase throughput efficiency.

code converter a device for changingcodes from one form to another.

code division multiple access (CDMA)a technique for providing multiple access tocommon channel resources in a communica-tion system. CDMA is based on spread spec-trum techniques where all users share all thechannel resources. Multiple users are distin-guished by assigning unique spreading codesto each user. Traditionally, individual detec-tion is accomplished at the receiver throughcorrelation or matched filtering.

code efficiency the unitless ratio of the av-erage amount of information per source sym-

bol to the code rate, where the amount of in-formation is determined in accordance withShannon’s definition of entropy. It is a funda-mental measure of performance of a codingalgorithm.

code excited linear prediction (CELP) aclass of linear predictive speech coding meth-ods where the excitation is composed of sam-ple vectors from VQ codebooks.

code hopping the use of a new spread-ing code for each transmitted bit in a spread-spectrum system.Compare withfrequencyhopping.

code letter SeeNEMA code letter.

code rate in forward error control and linecodes, the ratio of the source word length tothe code word length, which is the averagenumber of coded symbols used to representeach source symbol.

code segment area in a process’ virtualaddress space used to contain the program’sinstructions.

code tracking the process of continu-ously keeping the code sequences in the re-ceiver and transmitter in a spread-spectrumsystem synchronized during data transmis-sion. See alsocode acquisition.

code V a widely employed computer codefor design of optical systems by Optical Re-search Associates.

codebook a set of codevectors (or code-words) that represent the centroids of a givenpattern probability distribution. See alsovector quantization.

codebook design a fundamental prob-lem in vector quantization (VQ). The mainquestion addressed by codebook design ishow the codebook should be structured toallow for efficient searching and good per-formance. Several methods (tree-structured,


product codes, M/RVQ, I/RVQ, G/SVQ,CVQ, FSVQ) for codebook design are em-ployed to reduce computational costs low.See alsovector quantization, tree structuredVQ.

codebook generation a fundamentalproblem in vector quantization. Codebooksare typically generated by using a training setof images that are representative of the im-ages to be encoded. The best training imageto encode a single image is the image itself.This is called a local codebook. The mainquestion addressed here is what codevectorsshould be included in the codebook.See alsovector quantization.

codebook training the act of designing acodebook for a source coding system. TheLBG algorithm is often used to design thecodebook for vector quantizers.

codec word formed from encoder anddecoder. A device that performs encodingand decoding of communications protocols.

coded modulation an integrated modu-lation and coding approach for bandwidth-constrained channel where the redundancyintroduced by the code is compensated byincreasing the number of signals, for per-formance improvement without additionalbandwidth or transmission power.

codeword the channel symbol assignedby an encoder to a source symbol. Typicallythe codeword is a quantized scalar or vector.

coding the process of programming, gen-erating code in a specific language. The pro-cess of translating data from a representationform into a different one by using a set ofrules or tables.See alsoASCII, EBCDIC,binary.

coding at primary rates for videoconfer-encing Seeimage coding for videoconfer-encing.

coding gain (1) the reduction in signal-to-noise ratio required for a specified errorperformance in a block or convolutional for-ward error control system as compared to anuncoded system with the same informationrate, channel impairments, and modulationand demodulation techniques. In a trelliscoded modulation system, it is the ratio ofthe squared free distance in the coded systemto that of the uncoded system.

(2) the difference between the SNR/bit(dB) required for an uncoded and a codedsystem to attain the same arbitrary error prob-ability. Depends on the code parameters andalso on the SNR per bit.

coding of graphics use of a representa-tion scheme for graphics. Graphics codingis typically a two-level coding scheme. Bothexact and approximate methods are applica-ble to this type of coding. Run-length coding,predictive coding, line-to-line predictive dif-ferential coding, and block coding are typicalfor graphic coding.

coding of line drawings use of a rep-resentation scheme for line drawings. Linedrawings are typically coded using chaincodes where the vector joining two succes-sive pixels are assigned a codeword. Higherefficiency is obtained by differential chaincoding in which each pixel is represented bythe difference between two successive abso-lute codes.

coding redundancy Seeredundancy.

coefficient of thermal expansion (CTE)mismatch the difference between the co-efficients of thermal expansion of two com-ponents, i.e., the difference in linear thermalexpansion per unit change in temperature.(This term is not to be confused with ther-mal expansion mismatch).

coefficient of utilization (CU) the ra-tio of the lumens reaching the working planeto the total lumens generated by the lamp.This factor takes into account the efficiency


and distribution of the luminaire, its mount-ing height, the room proportions, and the re-flectances of the walls, ceiling, and floor.

coefficient sensitivity let a transfer func-tion be a ratio of polynomials

F(s) = N(s)

D(s)= a0 + a1s + . . .+ ams

m

d0 + d1s + . . .+ dnsn

in which the coefficientsai anddi are realand can be functions of an arbitrary circuitelementx. For such an elementx one maydefine the relative coefficient sensitivities asfollows:

Sai

x = ∂ai

∂x

x

aiSdi

x = ∂di

∂x

x

di

The relationship between the function sensi-tivity and coefficient sensitivities can be es-tablished as well.

coercive field the electric field value atwhich the polarization is zero when a ferro-electric material is cycled through the hys-teresis loop. A material has a negative and apositive coercive field and these are usually,but not always, equal in magnitude to eachother.

coercive force the demagnetizing fieldapplied to a permanent magnet that reducesits magnetic induction to zero; the x-interceptof the normal demagnetization curve. Acommonly listed material property that in-dicates magnet performance in static condi-tions.

coercivity Seecoercive force. See alsointrinsic coercive force.

COGEN Seecogeneration.

cogeneration (1) any of a number ofenergy generation systems in which two (ormore) forms of energy are produced in formspractical for use or purchase by an end user.Typical systems produce electrical energy forsale to a utility and process steam for localspace heating or other process uses. Cogen-

eration designs are generally adopted to in-crease the overall efficiency of a power gen-eration process.

(2) typically, the production of heat en-ergy, e.g. to heat buildings, as an adjunct tothe production of electric power.

cognitive map the cognitive map, intro-duced by R. Axelrod to study decision mak-ing processes, consists of points, or nodes,and directed links between the nodes. Thenodes correspond to concepts.See alsofuzzycognitive map.

coherence (1) measure of the extent towhich knowledge of a field at one point inspace permits prediction of the field at an-other point.

(2) in an optical fiber.(3) the coherence between two wide-sense

stationary random processes is equal to thecross power spectrum divided by the squareroot of the product of the two auto-powerspectra. The magnitude of the coherence sodefined is thus between 0 and 1.

coherence bandwidth the bandwidthover which the effect of communicationchannel can be assumed constant. Signals ofbandwidth less than this can be transmittedwithout significant distortion.

coherence distance See coherencelength.

coherence length distance over whichthe amplitude and phase of a wave can bepredicted.

coherence time the time over which theeffect of communication channel can be as-sumed constant. Signals of duration less thanthis can be transmitted without significantdistortion.

coherency a property of a control areathat has stiff interconnections between gen-erators. Such an area thus may be described


with the use of only a single frequency statevariable.

coherent integration where magnitudeand phase of received signals are preservedin summation.

coherent acousto-optical processoracousto-optical (AO) signal processor wherethe light is amplitude-modulated by theacoustic wave in the AO device as opposedto intensity or power modulated.

coherent detection detection techniquein which the signal beam is mixed with alocally generated laser beam at the receiver.This results in improved receiver sensitivityand in improved receiver discrimination be-tween closely spaced carriers.

coherent illumination a type of illumi-nation resulting from a point source of lightthat illuminates the mask with light from onlyone direction. This is more correctly called“spatially coherent illumination.”

coherent light light having a relativelylong coherence length; laser light.

coherent optical communication opti-cal communication approaches where infor-mation is conveyed in the phase of the opticalsignal, therefore requiring that the phase ofthe optical sources be well controlled.Seealsooptical communications.

coherent population trapping a tech-nique for creating a quantum mechanical co-herence in a lambda system by a dissipativeprocess. Ideally, the intermediate excitedstate in the lambda system decays rapidlycompared to the other two states. The coher-ence arises because a particular linear com-bination of ground states is not coupled tothe excited state. Atoms accumulate into thisuncoupled superposition state by a processanalogous to optical pumping, thereby creat-ing a quantum mechanical coherence. Thissystem often has counterintuitive properties,

because it uses a dissipative process to create,rather than destroy, a quantum coherence.

coil a conductor shaped to form a closedgeometric path. Note that the coil willnot be a closed conducting path unless thetwo ends of the coil are shorted together.Coils may have multiple turns, and may havevarious constructions including spool, pre-formed, and mush-wound. The coil may bewrapped around an iron core or an insulatingform, or it may be self-supporting. A coiloffers considerable opposition to AC currentbut very little to DC current.

coil pitch Seecoil span.

coil side that portion of a motor or gen-erator winding that cuts (or is cut by) linesof magnetic flux and, thus, contributes to theproduction of torque and Faraday EMF in thewinding.

coil span the distance, measured eitherin number of coil slots or in spatial (me-chanical) degrees, between opposite sides ofa winding of an electric machine. A full-span (full-pitch) winding is one in which thewinding span equals the span between adja-cent magnetic poles. Windings with span lessthan the distance between adjacent magneticpoles are called short-pitch, fractional-pitch,or chorded windings. Also called coil pitch.

cold plasma a simplified model of theplasma state where the effects that depend onelectron temperature are neglected. The par-ticles are assumed to have no kinetic thermalmotion of their own. The particles are at restexcept for their induced velocities throughthe action of the self consistent electromag-netic fields.

cold reserve the state of an idle thermalgenerating plant whose boilers and turbineare cold and must be brought up to operatingtemperature before power can be generated.


cold start (1) a complete reloading of thesystem with no reassumption. All executedprocesses are lost.

(2) the starting of a computer system froma power-off condition.

(3) the state from which a thermal genera-tion unit must be brought after being in coldreserve.

cold start miss in a cache, a cold startmiss occurs when a computer program is ref-erencing a memory block for the first time,so the block has to be brought into the cachefrom main memory. Also called first refer-ence misses or compulsory misses. Whenthe cache is empty, all new memory blockreferences are cold misses.See alsocapacitymiss, conflict miss.

collapsible reel a take-up reel used in linework which fits on the power-take-off of aline truck.

collector wall the collector of a bipo-lar transistor is located below the surface ofthe substrate. The wall, or sidewall, is thevertical boundary of the collector that meetsthe substrate material. The boundary usuallyforms a p-n junction that provides isolationfrom.

collet a circular spring fingerstock con-nection element for a power vacuum tube.

colliding-pulse-modelocked (CPM) lasera dye laser resonator design for producing

femtosecond pulses; right and left travellingpulses collide in a thin intracavity absorber.

collimated beam with nearly flat phasefronts and slow longitudinal variations of thetransverse amplitude distribution.

collinear geometry acousto-optical tun-able filter acousto-optical tunable filterdevice where the acoustic and light wavespropagate in the same direction. Also abbre-viated collinear geometry AOTF.

collinear geometry AOTF Seecollineargeometry acousto-optical tunable filter.

collision (1) in a pipeline, a situation whentwo or more tasks attempt to use the samepipeline stage at the same time.

(2) in a hash table, whenn + 1 differentkeys are mapped by a hashing function tothe same table index (wheren entries can bestored).

collision broadening broadening ofthe spectral profile of an amplifying or ab-sorbing transition due to inelastic or phase-interrupting elastic collisions.

collision vector a binary number in whichthe ith bit is a 1 ifsubmitting a task into thepipelinei cycles after a task will cause a col-lision.

color visual sensation associated with thewavelength or frequency content of an opticalsignal.

color blooming phenomenon where theexcess charge at a photo receptor can spreadto neighboring receptors, and change theirvalues in proportion to the overload. ForRGB cameras, this effect can modify notonly the luminance but also the chrominanceof pixels. Seecolor clipping, chrominance,luminance.

color burst burst of eight to ten cycles ofthe 3.579545 MHz (3.58 MHz) chrominancesubcarrier frequency that occurs during thehorizontal blanking of the NTSC compositevideo signal. The color burst signal synchro-nizes the television receiver’s color demodu-lator circuits.

color clipping phenomenon where the in-tensity of the light on a photoreceptor exceedssome threshold, the receptor becomes satu-rated and its response is no longer linear, butlimited to some bound. For RGB cameras,this effect can modify not only the luminance


but also the chrominance of pixels.Seecolorblooming, chrominance, luminance.

color coding the process of identifyingcomponents’ values and tolerances by meansof a set of colored bands or dots.

color correction in practical photome-try it is known that the system used to mea-sure luminescence will not possess the stan-dard eye spectral response as specified by the1931 International Commission on Illumina-tion (CIE). The measurement “system” willundoubtedly consist of a photodetector, anoptical filter and associated lenses. Unfortu-nately, system output is highly dependent onits spectral response.

color difference signals the chromi-nance signal component that results fromsubtracting the luminance (Y) componentfrom a primary color. The luminance sig-nal corresponds to the changes in brightnessas from a monochrome video signal. Thethree primary colors for color television sig-nals are located in Maxwell’s chromaticitydiagram. Red is at a wavelength of 0.7 mi-crometer, green at 0.546 micrometer and blueat 0.436 micrometer. The luminance signalcomponent results from the matrix additionof the primary colors. The matrix propor-tions are 30% red, 59% green, and 11% blue.Two color difference signals, (R-Y) and (B-Y) are sufficient to convey all the informationnecessary to reproduce full color at the TV re-ceiver. The color difference signal (G-Y) canbe determined by proper proportions of the(R-Y), (B-Y) and Y signals at the receiver.

color graphics adapter (CGA) a videoadapter proposed by IBM in 1981. It is ca-pable of emulating MDA. In graphic mode,it allows one to reach 640× 200 (wide perhigh) pixels with 2 colors or 320× 200 with4 colors.

color image coding compression of colorimages is usually done by transforming RGBcolor space into aYC1C2 space, whereY rep-

resents luminance andC1 andC2 are colordifference signals. TheC1 andC2 signals arethen subsampled, but coded with the samealgorithm as theY signal. Standard algo-rithms do not attempt to exploit correlationsbetween the three signals.

color matching the process of mixingthree fixed and independent primary colorsso that an observer (trichromat) interprets theformulation as being the same as a speci-fied but arbitrary color. In color television,the three primary colors are fixed at specificwavelengths bandsλR, λG, andλB corre-sponding to colors red, green, and blue.

color preference index (CPI) measureappraising a light source for appreciativeviewing of colored objects or for promotingan optimistic viewpoint by flattery.

color representation a method of defin-ing a signal or an image pixel value to beassociated with a color index.

color saturation a color with the dom-inant wavelength located at the periphery ofMaxwell’s chromaticity diagram. A fullysaturated color is pure because it has not beencontaminated by any other color or influence.

color signal the portion of a modulatedsignal that determines the colors of the in-tended output display.

color space the spaceC within whichcolors are represented in the image functionI : R2 → C.

color temperature the color a black ob-ject becomes when it is heated. The stan-dard color “white” occurs when a tungstenfilament is heated to a temperature of 6800degrees Kelvin. The temperature of 6800 Kcorresponds to a standard white raster as de-fined by the NTSC. The color temperature forwhite is useful for comparing color match-ing and color decoding among different dis-plays that use different color phosphors. The


standard “white” is obtained by mixing the30% red, 59% green, and 11% blue color sig-nals. Differences in the color saturation forthe different phosphors found in televisionCRTs will modify the required proportionsof red, green, and blue to produce the stan-dard “white.”

color-bar patterns a standard color-barpattern for an NTSC video signal consistingof a composite video signal containing a 77%and a 100% white chip, and yellow, cyan,green, magenta, red, blue, and black chips.These patterns represent ideal color and lu-minance levels that can be input in a videosystem for setting levels and verifying sys-tem performance.

Colpitts oscillator a particular case of anLC-oscillator whenX1 andX2 are capacitors(hence,Xm = 0),X3 an inductance.

column decoder logic used in a direct-access memory (ROM or RAM) to select oneof a number of rows from a given columnaddress.See alsotwo-dimensional memoryorganization.

column distance the minimum Hammingdistance between sequences of a specifiedlength encoded with the same convolutionalcode that differ in the first encoding interval.

column-access strobe Seetwo-dimen-sional memory organization.

comb filter an electric wave filter that ex-hibits an amplitude versus frequency plot ofperiodically spaced pass bands interspersedwith periodic stop bands. This plot resem-bles the teeth of an ordinary hair comb, fromwhich the filter derives its name.

comb function a function made of evenlyspaced, equal amplitude time or frequencycomponents (the Fourier transform of theComb function is another Comb function).The comb function is useful for discretizingcontinuous signals and can be represented

as the infinite sum of delta functions evenlyspaced through time or frequency.

comb-line filter filter consisting of par-allel coupled transmission line resonatorswhere all resonators are grounded on oneside and capacitively loaded to ground on theother. Adjacent resonators are grounded onthe same side. When fabricated as strip con-ductors in microstrip or stripline form, themetalized patterns have the appearance of acomb.

combination tone various sum and dif-ference frequency that are generated whentwo intense monochromatic fields interactwith the same semiclassically described lasermedium.

combinational lock interconnections ofmemory-free digital elements.

combinational logic a digital logic, inwhich external output signals of a device aretotally dependent on the external input sig-nals applied to the circuit.

combined cycle plant a gas-turbinepower plant in which the exhaust gases areused to heat water in a boiler to provide steamto run a turbogenerator.

combined field integral equation (CFIE)a mathematical relationship obtained by

combining the electric field integral equation(EFIE) and magnetic field integral equation(MFIE). It is normally used in electromag-netic scattering calculations from a conduct-ing body to avoid non-physical interior res-onances that appear by using either EFIE orMFIE alone.

combined source-channel coding a gen-eral term for approaches to source-channelcoding, where the source and channel codesare combined into one overall code. In the lit-erature, the term is also used, more loosely,for approaches where (any kind of) joint opti-mization of the source and channel coding is


utilized. Also commonly referred to as jointsource-channel coding.

combo trouble-shooter in combinedelectric and gas utilities, a practice which isgrowing in popularity is the use of combotroubleshooters. The combo troubleshooteris cross trained in both electric and gas ser-vice practices. The cross functionality per-mits more efficient deployment of resources.

come-a-long a ratcheted winch or block-and-tackle for pulling conductors into place.

command (1) directives in natural lan-guage or symbolic notations entered by usersto select computer programs or functions.

(2) instructions from the central processorunit (CPU) to controllers and other devicesfor execution.

(3) a CPU command, or a single instruc-tion, ADD, LOAD, etc.

Commision International d’Eclairage(CIE) International standards body forlighting and color measurement. Known inEnglish as the International Commission onIllumination.

commit the phase of a transaction inwhich the new states are written to the globalmemory or database. The commit phaseshould not be started until it has been veri-fied that performing the commit will not vi-olate the system’s consistency requirements.In most designs, the commit phase itself mustbe performed under more strict locking thanthe remainder of the transaction.

Comite Consulatif International Telegra-phique et Telephonique (CCITT) Inter-national Consultative Committee for Teleg-raphy and Telephony. This institution, basedin Geneva, Switzerland, issues recommenda-tions concerning all fields related to telecom-munications.

common base amplifier a single transis-tor BJT amplifier in which the input signalis applied to the emitter terminal, the outputis taken from the collector terminal, and thebase terminal is connected to a constant volt-age.

common centroid a technique inthe physical design of integrated circuitsin which two transistors, which must bematched, are actually composed of multipledevices connected in parallel. By appropri-ately connecting the multiple devices, the ef-fective center (“centroid”) of the two transis-tors can be located at the same point, thusimproving the matching in the presence ofnonidealities in the integrated circuit fabri-cation process.See alsocross-quad.

common channel signaling (CCS) atechnique for routing signaling informationthrough a packet-switched network.

common collector amplifier a single-transistor BJT amplifier in which the inputsignal is applied to the base terminal, the out-put is taken from the emitter terminal, and thecollector terminal is connected to a constantvoltage. Also referred to as an emitter fol-lower, since the voltage gain of this config-uration is close to unity (the emitter voltage“follows” the base voltage).

common drain amplifier a single tran-sistor FET amplifier in which the input signalis applied to the gate terminal, the output istaken from the source terminal, and the drainterminal is connected to a constant voltage.Also referred to as a source follower, sincethe voltage gain of this configuration is closeto unity (the source voltage “follows” the gatevoltage).

common emitter a basic transistor am-plifier stage whose emitter is common to bothinput and output loops. It amplifies voltage,current, and hence power.


common emitter amplifier a single-transistor BJT amplifier in which the inputsignal is applied to the base terminal, the out-put is taken from the collector terminal, andthe emitter terminal is connected to a constantvoltage.

common gate amplifier a single-transistor FET amplifier in which the inputsignal is applied to the source terminal, theoutput is taken from the drain terminal, andthe gate terminal is connected to a constantvoltage.

common mode gain for a differentialamplifier, the ratio of the output signal am-plitude to the amplitude of a signal appliedto both the amplifier input terminals (in com-mon). For an ideal differential amplifier, thecommon mode gain would be zero; the devi-ation of a real differential amplifier from theideal is characterized by the common moderejection ratio (CMRR).

common mode noise undesired electri-cal signals in lines that are equal in amplitudeand phase with respect to a reference ground.Common mode voltages and currents can begenerated by power electronic switching cir-cuits and can interfere with control or otherelectronic equipment. Common mode cur-rents will also sum into neutrals and ground-ing conductors, which may cause sensitivefault current detection relays to trip.

common mode rejection ratio (CMRR)a measure of quality of an amplifier with dif-ferential inputs, defined as the ratio betweenthe common-mode gain and the differentialgain.

common source amplifier a single-transistor FET amplifier in which the inputsignal is applied to the gate terminal, the out-put is taken from the drain terminal, and thesource terminal is connected to a constantvoltage.

common-channel interoffice signalingthe use of a special network, dedicated to sig-naling, to establish a path through a commu-nication network, which is dedicated to thetransfer of user information.

common-mode coupling pick-up froman electromagnetic field that induces achange in potential on both signal leads ofequal magnitude and phase relative to theground reference potential.

communication link a point-to-pointcommunication system that typically in-volves a single information source and a sin-gle user. This is in contrast to a communica-tions network, which usually involves manysources and many users.

communication theory Seeinformationtheory.

community-antenna television (CATV)a television receiving and distribution systemin which signals from television stations andsometimes FM stations are received by anten-nas, amplified, and then distributed to com-munity subscribers via coaxial or fiber-opticcable. The system is known as cable TV.

commutating inductance in switchedcircuits (converters, inverters, etc.), the in-ductance that is in series with the switchingelements during the process of commutationfrom one topological state to another. Thisinductance results in noninstantaneous com-mutation due to the fact that current in aninductor cannot change instantaneously.

commutating pole Seeinterpole.

commutating winding Seeinterpole.

commutation the process by which al-ternating current in the rotating coil of a DCmachine is converted to unidirectional cur-rent. Commutation is accomplished via aset of stationary electrical contacts (brushes)sliding over multiple, shaft-mounted electri-


cal contacts that turn with the machine ro-tor. The contacts are the connection pointsin a series-connected loop of the coils thatmake up the rotor winding. The brushes,sliding over these contacts, continually di-vide the loop into two parallel electrical pathsbetween the brushes.

The brushes are positioned such that theymake contact with those commutator seg-ments that are connected to coils that aremoving through a magnetic neutral point be-tween poles of the machine’s field flux. As aresult, all coils making up one parallel pathare always moving under a north magneticpole, and the others are always moving undera south magnetic pole. The movement of thecommutator contacts underneath the brushesautomatically switches a coil from one pathto the other as it moves from a north pole re-gion to a south pole region. Since the coilsin both paths move in the same direction, butthrough opposite flux regions, the voltagesinduced in the two paths are opposite. Con-sequently, the positive and negative ends ofeach path occur at the same points in the se-ries loop, which are at the points where thebrushes contact the commutator. The brushpositions, thus, represent a unidirectional (orDC) connection to the rotating coil.See alsocommutator.

commutation angle time in electrical de-grees from the start to the completion of thecommutation process. Also called overlapangle.

commutativity a property of an opera-tion; an operation is commutative if the re-sult of the operation is not affected by anyreordering of the operands of the operation.Additions and multiplication are commuta-tive, whereas subtraction and division are not.

commutator a cylindrical assembly ofcopper segments, insulated from each an-other, that make electrical contact with sta-tionary brushes, to allow current to flowfrom the rotating armature windings of aDC machine to the external terminals of

the machine. It also, enables reversal ofcurrent in the armature winding.See alsocommutation.

commutator film an oxide layer onthe commutator surface, indicated by a darkcolor or a “film,” that is required for propercommutator action and full loading of themachine. On a new DC machine commu-tator, or on a commutator that has just beenstoned, there is no “film” on the commutator.It is advisable to refer to the manufacturer’stechnical manual for the proper procedure to“break in” the commutator and develop thefilm so the machine can be operated at ratedconditions.

compact disk (CD) a plastic substrateembossed with a pattern of pits that encodeaudio signals in digital format. The disk iscoated with a metallic layer (to enhance itsreflectivity) and read in a drive (CD player)that employs a focused laser beam and mon-itors fluctuations of the reflected intensity inorder to detect the pits.

compact disk-interactive (CD-I) a spec-ification that describes methods for providingaudio, video, graphics, text, and machine-executable code on a CD-ROM.

compact range an electromagnetic mea-surement facility in which far-field condi-tions are achieved by the use of an offsetparabolic reflector. The reflector is fed usinga source antenna or other subreflector sys-tem located at its focus. The term “compact”range is used to describe the relative differ-ence in its size compared to a true far-fieldrange requiring a large separation distancebetween the source antenna and the deviceunder test to achieve the same far-field con-ditions.

compactness measure an alternativename forcircularity measure.

compander a point operation that loga-rithmically compresses a sample into fewer


bits before transmission. The inverse loga-rithmic function is used to expand the codeto its original number of bits before convert-ing it into an analog signal. Typically usedin telecommunications systems to minimizebandwidth without degrading low-amplitudesignals.

companding a process designed to min-imize the transmission bit rate of a signal bycompressing it prior to transmission and ex-panding it upon reception. It is a rudimen-tary data compression technique that requiresminimal processing.

companion matrix the coefficient ma-trix in the state-equation representation of thenetwork describable by a linear differentialequation.

comparator (1) a logic element that com-pares two binary numbers (A andB) to de-termine ifA = B, A < B, or A > B. Anexclusive NOR gate operates like a 1-bit com-parator.

(2) a software tool that compares two com-puter programs, files, or sets of data to iden-tify commonalities or differences. Typicalobjects of comparison are similar versions ofsource code, object code, data base files, ortest results.

compare instruction an instruction usedto compare two values. The processor flagsare updated as a result. For example, the in-struction CMP AL,7 compares the contentsof register AL with 7. The zero flag is set ifAL equals 7. An internal subtraction is usedto perform the comparison.

compartmental model a dynamicalmodel used in analysis of biomedical, phar-macokinetic, and ecological systems. Themain idea in compartmental modeling is to“lump” in reality distributed system into afinite number of homogeneous, well-mixedsubsystems called compartments or pools,which exchange materials with each otherand with the environment. Usually the com-

partments are described by the first-order dif-ferential or difference equations, and in thissense they form a system of state equations.Although the most popular application ofcompartmental models is in modeling pop-ulation dynamics and other biomedical phe-nomena, they may be also used to describesome engineering processes, e.g., distillationcolumns. The case of linear time-invariantcompartments in which exchange rates areproportional to the state of the donor com-partments may be treated by Laplace trans-forms and transfer function analysis.

compatibility (1) two different imple-mentations of the same component wherebythey may both be used in a system with nomodification (often used in the context of newmicroprocessors running software compiledfor older microprocessors).

(2) the capability of a functional unit tomeet the requirements of a specified inter-face.

compensated pulsed alternator (CPA)Seecompulsator.

compensating winding a winding foundin DC machines that is placed in the facesof the main field poles, and connected in se-ries with the armature winding, to producean mmf equal and opposite to the mmf ofthe armature, thereby reducing the effect ofarmature reaction.

compensation (1) operations employedin a control scheme to counteract dynamiclags or to modify the transformation betweenmeasured variables and controller output toproduce prompt stable response.

(2) the alteration of the dynamic behav-ior of a process by the addition of sys-tem blocks. These are usually connectedin cascade with the original process on ei-ther its input or its output variables, or both.See alsocompensator, pre-compensatorandpost-compensator.


compensator a system block added toan existing system (or process) to producea combined transfer function that improvesits performance when connected in a closedloop configuration.See alsocompensation,pre-compensatorandpost-compensator.

compensatory behavior human dy-namic behavior in which the operator’s ac-tions are conditioned primarily by the closed-loop man-machine system errors.

compensatory display for the simplestcase, a display which shows only the differ-ence between the desired input command andthe system output.

compiler a program that translates ahigh level language program into an exe-cutable machine instruction program or otherlower-level form such as assembly language.See also linker, assembler, interpreter,cross-assembler, cross-compiler.

complement (1) to swap 1’s for 0’s and0’s for 1’s in a binary number.

(2) opposite form of a number system.

complement of a fuzzy set the membersoutside of a fuzzy set but within the universeof discourse. Represented by the symbol¬.

Let A be a fuzzy set in the universe of dis-courseX with membership functionµA(x),x ∈ X. The membership function of thecomplement ofA , for all x ∈ X, is

µ¬A(x) = 1 − µA(x)

See alsocomplement, fuzzy set,membershipfunction.

complement operator the logical NOToperation. In a crisp (non-fuzzy) system, thecomplement of a setA is the set of the ele-ments that are not members ofA. The fuzzycomplement represents the degree to whichan element is not a member of the fuzzy set.

complementary arithmetic a method ofperforming integer arithmetic within a com-puter, in which negative numbers are repre-sented in such a way that the arithmetic maybe performed without regard to the sign ofeach number.

complementary cumulative distributionfunction (CCDF) a function describ-ing the probabilityp(x) of achieving all out-comes in an experiment greater thanx.

complementary metal oxide semiconduc-tor (CMOS) (1) refers to the process thatcombines n-channel and p-channel transis-tors on the same piece of silicon (complemen-tary). The transistors are traditionally madeof layers of metal, oxide, and semiconductormaterials, though the metal layer is often re-placed by polysilicon. There are a numberof variations such as HCMOS, high-speedCMOS which scales down the elements com-pared to the standard MOS process and thusincreases the speed and reduces the powerconsumption for each transistor in the CPU.

(2) a CMOS memory device used in com-puters to store information that must be avail-able at startup. The information is main-tained in the device by a small battery.

complete statistic a sufficient statisticTwhere every real-valued function ofT is zerowith probability one whenever the mathe-matical expectation of that function ofT iszero for all values of the parameter. In otherwords, ifW is real-valued function, then T iscomplete if

EθW(T ) = 0∀θ ∈ 2 ⇒ Pθ [W(T ) = 0]= 1∀θ ∈ 2

completion unit Seeretire unit.

complex amplitude magnitude of anearly harmonic function, complex to in-clude phase deviations from a referencewave.


complex amplitude transmittancetransmittance of the complex amplitude,square root of the intensity transmission.

complex beam parameter Seebeamparameter.

complex envelope a low-pass complexvalued signal used to represent a real band-pass signal. The complex envelope is ob-tained from the analytic signal with centerfrequencyωc by multiplying the analytic sig-nal bye−jωct .

complex exponential signal a signal ofthe form x(t) = C expjωt , whereC is aconstant andω is the frequency in radiansper second.

complex frequency a complex numberused to characterize exponential and dampedsinusoidal motion in the same way that anordinary frequency characterizes simple har-monic motion; designated by the constants

corresponding to a motion whose amplitudeis given byAest , whereA is a constant andtis time, ands = σ + jω whereσ is the realpart ofs andω is the imaginary part ofs. ωis also known as the real angular frequency.

complex instruction set computer (CISC)a processor with a large quantity of instruc-tions, some of which may be quite compli-cated, as well as a large quantity of differentaddressing modes, instruction and data for-mats, and other attributes. The designationwas put forth to distinguish CPUs such asthose in the Motorola M68000 family and theIntel Pentium from another approach to CPUdesign that emphasized a simplified instruc-tion set with fewer but possibly faster exe-cuting instructions, called RISC processors.One CISC processor, the Digital VAX, hasover 300 instructions, 16 addressing modes,and its instruction formats may take up 1 to51 bytes.

A CISC processor usually has a relativelycomplicated control unit. Most CISC pro-cessors are microprogrammed.

One of the benefits of a CISC is that thecode tends to be very compact. When mem-ory was an expensive commodity, this was asubstantial benefit. Today, speed of execu-tion rather than compactness of code is thedominant force.

See also microprogramming, reducedinstruction set computer processor.

complex number a number consistingof a real part and an imaginary part, usuallyexpressed in the forma+bi, where the “i” isused to distinguish thatb represents the com-plex part of the number.i is mathematicallydefined as the positive root of−1.

complex power a complex number thatrepresents electric power flow for an AC cir-cuit. When expressed in rectangular form itsreal part is average powerP in watts and itsimaginary part is reactive powerQ in reac-tive volt-amperes. When expressed in polarform its magnitude is apparent powerS involt-amperes and its angle is the power fac-tor angle (the same angle as the impedanceangle for a passive load).See alsoapparentpower.

complex process (system) term usedrather colloquially to denote controlled pro-cess (control system) possessing such charac-teristic features which, separately or jointly,allow to treat this process (system) as a com-plex entity; the features worth considerationare:

1. The process is large in a physical sense— it occupies large space and there are largedistances among its different elements,

2. The model of the process is compli-cated and involves many variables, in partic-ular control inputs,

3. The process is composed of several in-teracting subprocesses and there are identifi-able local objectives and local sets of decisionvariables,

4. The control problem is seen as compli-cated due to the nature of the control objec-tives and the way by which free inputs areformed and influence the controlled process


— even when the underlying physical pro-cess does not seem to be complex.

complex propagation constant propa-gation constant or wave number in a mediumwith gain or loss.

complex system Seecomplex process.

complex transmittance the effect of amedium on both optical phase and amplitudeof light traversing the medium.

complexity-constrained maximum-likeli-hood the maximum of the likelihoodfunction given some quantifiable complex-ity constraint, M, i.e., max{p(y|x,M)}.Breadth-first search algorithms can performcomplexity-constrained maximum-likelihooddetection on tree and trellis structured prob-lems.

compliance matrix for the arm end pointis defined formally by the following expres-sion: JAK−1J TA whereJA is an analyticalJacobian of the manipulator andK a positivedefinite matrix describing joint stiffness ofthe manipulator. MatrixK is invertible. No-tice that the compliance matrix depends onthe structure of the manipulator and changeswith its position in Cartesian space.See alsostiffness of a manipulator armandanalyticalJacobian.

compliant motion motion of the manip-ulator (robot) when it is in contact with its“environment,” such as writing on a chalk-board or assembling parts.

component mounting site a location ona packaging and interconnecting structure,consisting of a land pattern and conductorfan-out to additional lands for testing or vias,used for mounting a single component.

composite a material usually consistingof a resin supporting fibers of a lightweightfabric that may be woven and treated in order

to produce certain strength and/or electricalcharacteristics.

composite maximum method a methodof defuzzification in which the defuzzifiedor crisp value is arrived at using the maxi-mum value of the membership function ofthe fuzzy set.

composite moments method Seecentroid method.

composite second order (CSO) ratioof the power in the second-order distortionproducts to power in the carrier in a cabletelevision channel.

composite sync a synchronizing sig-nal consisting of both horizontal and verticalsync information. Composite sync is usedfor providing synchronizing pulses to videoequipment in the studio.

composite transform a transform that canbe factored into two or more transforms.

composite triple beat (CTB) same ascomposite second order but for third-orderdistortion.Seecomposite second order.

composite video (1) a single video sig-nal that contains luminance, color, and syn-chronization information. NTSC, PAL, andSECAM are all examples of composite videoformats.

(2) the complete video signal. For B&W,it consists of the picture signal, blankingpulses, and synchronizing signals. For color,color synchronizing and color picture infor-mation are added.Seefigure.

compositional rule of inference gener-alization of the notion of function. LetXandY be two universes of discourse,A be afuzzy set ofX, andR is a fuzzy relation inX × Y . The compositional rule of inferenceassociates a fuzzy setB in Y to A in threesteps:

1. Cylindrical extension ofA in X × Y ;


Computed torque control.

2. Intersection of the cylindrical extensionwith R;

3.B is the projection of the resulting fuzzyset onY .

If we chooseintersectionas triangularnormandunionastriangular co-norm, thenwe have the so-calledmax-min compositionB = A ◦ R, i.d.

µB(y) =∨x

[µA(x)

∧µR(x, y)

].

If we choosealgebric productfor trian-gular normandunionastriangular co-norm,then we have the so-calledmax-product com-positionB = A◦R, i.d.

µB(y) =∨x

[µA(x)µR(x, y)] .

The compositional rule of inference is theprincipal rationale behindapproximate rea-soning.

See alsoapproximate reasoning,cylindricalextension of a fuzzy set, fuzzy relation,intersection of fuzzy sets, projection of afuzzy set.

compound-connected DC machine adirect current machine with two field wind-ings in which one field winding is connectedin series and one field winding is connectedin parallel (shunt) with the armature wind-ing. The shunt winding may be connectedahead of the series winding (long-shunt con-nection), or behind the series winding (short-shunt connection).

compound-rectifier exciter a source offield current of a synchronous machine de-rived from the phase voltages and currents of

the machine. The phase voltages and currentsof the machine are fed through transformers,then rectified in order to provide DC quanti-ties to the field winding. The components ofthe exciter are the transformers (voltage andcurrent), rectifiers (including possible gate-circuitry), and power reactors; exclusive ofall input control elements.

compression (1) in information theory,the compact encoding (with a smaller numberof bits) Ic of a digital image or signalI ob-tained by removing redundant or nonsignif-icant information, thus saving storage spaceor transmission time. Compression is termedlossless, if the transformation ofI into Ic isreversible, otherwise it is termed lossy.

(2) in signal processing, at given bias lev-els and frequency, the ratio between the smallsignal power gain(poutSS/pincidentSS) undersmall signal conditions and the large signalpower gain(poutLS/pincidentLS)at a given inputpower, expressed in decibels. As the inputamplitude of a signal is increased, the out-put signal will eventually cutoff and/or clipdue to saturation, resulting in compression.If the large signal is insufficiently large tocause cutout and/or clipping, then the com-pression will be at or near 0 dB.

GCR = 10 log10 (poutSS/pincidentSS)

− 10 log10 (poutLS/pincidentLS)

compression coding the lossy (irre-versible) or loseless (reversible) process ofreducing the amount of digital informationrequired to represent a digital signal.


compression ratio the ratio of the num-ber of bits used to represent a signal beforecompression to that used after compression.

Compton laser free-electron laser inwhich the amplification mechanism is con-sidered to be Compton scattering.

compulsator the compulsator (compen-sated pulsed alternator or CPA), is a speciallydesigned rotating electrical alternator witha very low internal impedance that allowsit to produce large, repetitive pulses of cur-rent. These machines produce an alternatingcurrent output whose frequency is dependentupon the rotor speed and number of magneticpoles in the CPA. Typical output voltages ofa CPA are 1,000–10,000 volts with outputcurrents of up to 5,000,000 amperes and fre-quencies of 100–1,000 hertz.

compulsory miss Seecold start miss.

computational cut-off rate Seecut-offrate.

computational electromagnetics the useof modern digital hardware to obtain solu-tions to Maxwell’s equations and to visualizethese solutions.

computational intelligence See softcomputing.

computed tomography (CT) Seetomog-raphy.

computer (1) an electronic, electrome-chanical, or purely mechanical device thataccepts input, performs some computationaloperations on the input, and produces someoutput.

(2) functional unit that can perform sub-stantial computations, including numerousarithmetic operations, or logic operations,without human intervention during a run.

(3) general or special-purpose program-mable system that is able to execute programsautomatically. It has one or more associ-

ated processing units, memory, and periph-eral equipment for input and output. Usesinternal memory for storing programs and/ordata.

computer architecture an image of acomputing system as seen by a most sophis-ticated computer user and programmer. Theabove concept of a programmer refers to aperson capable of programming in machinelanguage, including the capability of writinga compiler. The architecture includes all reg-isters accessible by any instruction (includingthe privileged instructions), the complete in-struction set, all instruction and data formats,addressing modes, and other details that arenecessary in order to write any program. Thisdefinition stems from the IBM program ofgenerating the 360 system in the early 1960s.Contrast withcomputer organization. SeealsoFlynn’s taxonomy.

computer communication network col-lection of applications hosted on differentmachines and interconnected by an infras-tructure that provides intercommunications.

computer generated hologram a holo-gram where the required complex amplitudeand phase functions are generated by com-puter and written onto an optical medium.

computer hardware description language(CHDL) examples include VHDL andVerilog, current work in CHDL includesmainly languages for verification, and exten-sions of existing languages for system de-scription and analog design. CHDL confer-ences are organized every year.

computer model a computer model ofa device consists of a mathematical/logicalmodel of the behavior of the device repre-sented in the form of a computer program.A good computer model reproduces all thebehaviors of the physical device in questionand can be confidently used to simulate thedevice in a variety of circumstances.


computer organization describes the de-tails of the internal circuitry of the computerwith sufficient detail to completely speci-fies the operation of the computer hardware.Contrast withcomputer architecture.

computer relay a protective relay thatdigitizes the current and/or voltage signalsand uses a microprocessor to condition thedigitized signal and implement the operatinglogic. Seedigital relay.

computer simulation a set of computerprograms that allows one to model the im-portant aspects of the behavior of the spe-cific system under study. Simulation can aidthe design process by, for example, allowingone to determine appropriate system designparameters or aid the analysis process by, forexample, allowing one to estimate the end-to-end performance of the system under study.

computer torque control computedtorque control is depicted in figure. The feed-back controller sends its output through theinverse dynamic model. The feedback con-trol law comprises and independent-joint PDcontroller with velocity reference, plus thedesired acceleration. In the figureqd , qd , andqd denote desired position, velocity, and ac-celeration vectors, respectively.q andq de-note measured generalized position and ve-locity vectors. Finally,Kp andKd are posi-tive definite constant PD controller matrices.

computer vision Seerobot vision.

computer word data path of a computer(the size of virtual addresses); (1) datum con-sisting of the number of bits that forms thefundamental registers, etc.; (2) sequence ofbits or characters that is stored, addressed,transmitted, and operated as a unit within agiven computer. Computer words are one toeight bytes long, but can be longer for specialapplications.

computer-aided design (CAD) field ofelectrical engineering concerned with pro-

ducing new algorithms/programs which aidthe designer in the complex tasks associatedwith designing and building an integrated cir-cuit. There are many subfields of electricalCAD: simulation, synthesis, physical design,testing, packaging, and semiconductor pro-cess support.

computer-aided engineering (CAE)software tools for use by engineers.

computer-aided manufacturing (CAM)manufacturing of components and productswhen based heavily on automation and com-puter tools. See alsocomputer-integratedmanufacturing.

computer-aided software engineering(CASE) a computer application automat-ing the development of graphic and docu-mentation of application design.

computer-integrated manufacturing(CIM) manufacturing approach thatmakes substantial use of computers to controlmanufacturing processes across several man-ufacturing cells. See alsocomputer-aidedmanufacturing.

concatenated code (1) a code that is con-structed by a cascade of two or more codes,usually over different field sizes.

(2) the combination of two or more for-ward error control codes that achieve a levelof performance with less complexity than asingle coding stage would require. Seriallyconcatenated coding systems commonly usetwo levels of codes, with the inner code beinga convolutional code and the outer code be-ing a Reed–Solomon code. Parallel concate-nated codes improve performance throughparallel encoding and iterative serial decod-ing techniques.See alsoturbo code.

concentration gradient a difference incarrier concentration.

concentric resonator usually a symmet-ric laser resonator in which the mirror spacing


Computer torque control.

is equal to twice the mirror curvature; mirrorshave a common center of curvature.

concept formation the process of the in-cremental unsupervised acquisition of cate-gories and their intentional descriptions.

The representative concept formation sys-tems include EPAM, CYRUS, UNIMEM,COBWEB, and SGNN.See alsoself-generat-ing neural network.

concurrency the notion of having multi-ple independent tasks available (tasks in thisdefinition means any work to be done, not aformal computational entity).

concurrent processing having one log-ical machine (which may be a multiproces-sor) execute two or more independent taskssimultaneously.

concurrent read and concurrent write(CRCW) shared memory model, inwhich concurrent reads and writes are al-lowed.

concurrent read and exclusive write(CREW) shared memory model, in whichconcurrent reads but only exclusive writes areallowed.

condenser lens lens system in an opticalprojection system that prepares light to illu-minate the mask.

condition code internal flag used in theconstruction of CPUs. Many computers pro-vide a mechanism for saving the character-istics of results of a particular calculation.Such characteristics as sign, zero result, carryor borrow, and overflow are typical of inte-ger operations. The program may reference

these flags to determine whether to branch ornot.

condition code register register that con-tains the bits that are the condition codes forthe CPU arithmetic or compare instructions.

condition variable a variable set as theresult of some arithmetic or logical compar-ison.

conditionability of generalized 2-D modela mathematical relationship of interest in

control systems.The generalized 2-D model

Exi+1,j+1 = A0xij + A1xi+1,j

+ A2xi,j+1 + B0uij

+ B1ui+1,j + B2ui,j+1

is called conditionable if no two distinct so-lutions to the model for the same input se-quenceuij for (i, j) ∈ [0, N1] × [0, N2]coincide in all their boundary values (Seeboundary values of 2-D general model),wherexij ∈ Rn is the semistate vectoruij ∈Rm is the input vector andE,Ak, Bk (k =0,1,2) are real matrices withE possibly sin-gular or rectangular.

conditional coding an approach to the so-lution of the problem of large code words andlookup tables in block coding. In this schemeone assumes that the receiver already knowsthe componentsb1, b2, . . . bN−1 of N -tupleb. Current componentbN can now be codedusing this information. The assumption thatthere is statistical dependence between pixelsis made.

conditional instruction an instructionthat performs its function only if a certain


condition is met. For example, the instruc-tion JNZ TOP only jumps to TOP if the zeroflag is clear (the “not zero” condition).

conditional statistic a statistic premisedon the occurrence of some event. The proba-bility of eventE1 given thatE2 has occurredis denoted byp(E1|E2). See alsoBayes’rule.

conditionally addressed ROM read-only memory in which not every address canbe used to access a valid word. Usually im-plemented from a PLA.

conductance (1) the reciprocal of resis-tance.

(2) a characteristic that describes the avail-ability and the mobility of conduction elec-trons within a material. The values rangefrom zero for a perfect insulator to infinityfor a perfect conductor. The units are sie-mans.

(3) the ability of a substance to carry athermodynamic flow, such as current, heat,energy, etc.

conducted emission an RF current prop-agated through an electrical conductor.

conducted noise unwanted electrical sig-nals that can be generated by power elec-tronic switching circuits. Conducted noisecan travel through the circuit cables ascommon-mode or differential mode currentsand can interfere with control circuits or otherelectronic equipment.

conduction angle the period duringwhich a device is conducting, i.e., carryingcurrent. While the device could be a switchor any other electrical element, such as induc-tor, phase coil, capacitor, resistor, this term isprimarily applied to power electronic switch-ing devices, which are gated to operate forsome fraction of a power cycle.

The conduction angle is the sum of thepositive transition angle (θ+), the saturationangle (θsat), and the negative transition angle

(θ−). Hence, the conduction angleθcond is

θcond = θ+ + θsat + θ−

conduction band the lowest energy bandthat is not completely occupied by electronsin a crystalline solid.See alsovalence band.

conduction current the drift of electronsin a conductor (or of electrons and holes in asemiconductor).

conduction electron a free electron inthe conduction band of a semiconductor.

conductivity (1) the reciprocal of resis-tivity.

(2) a measure of a material’s ability to con-duct electrical current. Conductivityσ is theratio of the conduction current to the electricfield in Ohm’s Law:

Jc = σE

conduit a pipe through which an electricalcables are laid.

cone beam term describing the shape ofthe beam formed with an X-ray source andbeam restricter. Because the source is a finitedistance from the target, beam divergence oc-curs because the photons in the beam are nottravelling along truly parallel paths. Process-ing can be applied to correct for the beamshape during image reconstruction.

cone of protection a method used to de-termine the extent of protection to surround-ing structures afforded by a tall, groundedstructure like a steel tower. Proposed priorto the "rolling ball" model, this method sug-gests that any structure which can fit within aright circular cone whose vertex is at the topof the tower will be protected from lightningstrikes by that tower. The angle of the cone’svertex is a matter of some controversy.Seerolling ball.


confidence interval an interval aroundthe estimator. The interval contains the un-known parameter with this probability.

confidence level a probability that indi-cates the quality of an experiment.

configuration operation in which a set ofparameters is imposed for defining the oper-ating conditions. The configuration of a per-sonal computer regarding low-level featuresis frequently called set-up. At that level, thememory, the sequence of boot, the disk fea-tures, etc., are defined. The configurationof a computer also involves that of its op-erating system. For example, per MS-DOSsee CONFIG.SYS and AUTOEXEC.BAT.The configuration of applicative software de-pends on the software under configuration it-self.

confinement condition according towhich the amplitude of a beam falls to zeroat large distances from the beam axis.

confinement condition Seeconfinement.

confinement diagram diagram show-ing the values of the mirror curvatures of atwo-mirror laser such that the electromag-netic modes are confined (satisfy the confine-ment condition); sometimes called a stabilitydiagram because beam confinement can beassociated with ray stability.

conflict miss a cache miss category usedto denote the case where, if the cache isdirect-mapped or block-set-associative, toomany blocks map to a set leading to thatblocks can be expelled from the cache, even ifthe cache is not full, and later retrieved again.These are also called “collision misses.”Seealsocapacity miss.

conflict-free multiple access protocolclass of multiple access protocols in whichany transmission from a given user is guar-anteed to have exclusive access to the chan-

nel. It will not be interfered with by a trans-mission from another user. One way ofachieving this is to allocate the channel tothe users either statically of dynamically. Instatic channel allocation schemes the channelcan be divided into exclusive sub-channelsin the time domain (TDMA), frequency do-main (FDMA), code domain (CDMA), po-larization domain (PDMA), or in the space(SDMA). In the dynamic channel allocationscheme, the channel is allocated on a demandbasis using a reservation scheme or tokenpassing.

conflicting goals objectives of severaldecision units in charge of given partitionedsystem, for example objectives of local de-cision units in a large-scale system, whichwould lead to mutually conflicting actionsof the decision makers; a conflict may alsoarise between the objectives as perceived bya supremal unit and the local objectives; con-flict between local and, eventually, betweenglobal and local objectives may or may notbe alleviated by using the coordination in-struments.

confocal parameter measure of the waistsize of a Gaussian beam, 2 pi times waist spotsize squared divided by wavelength; twicethe Rayleigh length.

confocal resonator usually a symmetriclaser resonator in which the mirror spacing isequal to the mirror curvature; mirrors have acommon focal point.

conformal mapping a transformationw = f (z) defined on a domainDwith angle-preserving properties. The method of confor-mal mapping finds application, for example,in the quasi-static analysis of several trans-mission lines such as microstrips, coplanarwaveguides, etc.

confusion matrix a matrix describing thelikelihood of misclassification.

Let ci , i = 1, . . . , n be the classes inwhich a given set of patterns can be parti-


tioned. Leteci ,cj be the percentage of pat-terns of classci that are erroneously recog-nized as patterns of classcj . The matrixE

.= [eci ,cj

] ∈ Rn,n is called a confusionmatrix. Of course

n∑j=1

eci ,cj = 1

congestion a state of a packet-based sys-tem where too many packets are present inthe network and the overall performance de-grades. To resolve the congestion, the systemmust employ some form of congestion con-trol. See alsopreventive congestion controlandreactive congestion control.

conical diffraction a scattering phe-nomenon in photorefractive crystals in whichthe scattered beam forms a cone of light.When a laser beam of finite transversecross section passes through a photorefrac-tive crystal, beam fanning often occurs. Thehologram formed by the incident beam andthe fanned light consists of a multitude ofgratings because the fanned light spans awide solid angle in space. When such a mul-titude of gratings is read out by a laser beam,only a subset of these gratings matches theBragg condition with readout beam. Thewave vectors of the Bragg-matched read-out beams form hollow cones in momentumspace. Therefore, conical diffractions are ob-served most of the time when fanning occurs.Conical diffraction is also often referred to asconical scatterings.

conical scattering Seeconical diffraction.

conjugate symmetric transform a prop-erty of a real-valued function that relates toits Fourier transform. Ifx(t) is a real-valuedfunction and its Fourier transform has theproperty thatX(−w) = X∗(w), where∗ de-notes the complex conjugate. The transformX(w) is said to be conjugate symmetric.

conjunction rule of inference a rule ofreasoning which states that if two proposi-tionsA andB are both individually true, thenthe combined proposition “AAND B” is alsotrue.

connect/disconnect bus Seesplit trans-action.

connected component a maximal-sizedconnected region. Also termed “blob.”

connectedness a graph or subgraph issaid to be connected if there is at least onepath between every pair of its vertices.

connection matrix in a network of gen-eral topology, the connection matrix identi-fies how the circuit elements are connectedtogether.

connection weight in neural networks,within the processing element, an adaptivecoefficient associated with an input connec-tion. It is also referred to as synaptic efficacy.

connection-oriented service a mode ofpacket switching in which a call is estab-lished prior to any information exchange tak-ing place. This is analogous to an ordinaryphone call, except that no physical resourcesneed to be allocated.

connectionist model one of many namesgiven to the learning systems. The notionof learning systems has been developed inthe fields of artificial intelligence, cyber-netics and biology. In its most ambitiousform learning systems attempt to describe ormimic human learning ability. Attainmentof this goal is still far away. The learningsystems that have actually been implementedare simple systems that have strong relationsto adaptive control. The learning systemsare also known under the names of neuralnets, parallel distributed processing models,etc. Examples of learning systems most com-monly used are perceptron, Boltzmann ma-chine, Hopfield network. An interesting fea-


ture of the neural networks is that they op-erate in parallel and that they can be imple-mented in silicon. Using such circuits maybe a new way to implement adaptive controlsystems.

connectionless service a mode of packetswitching in which packets are exchangedwithout first establishing a connection. Con-ceptually, this is very close to messageswitching, except that if the destination nodeis not active, then the packet is lost.

connectivity specifies that sets are consid-ered to be connected. Generally it is basedon an adjacency relation between pixels (orvoxels), so that a setX is connected if andonly if for anyp, q ∈ X there is a sequencep0, . . . , pn (n ≥ 0) such thatp = p0,q = pn, and for eachk < n,pk is adjacent topk+1. Seepixel adjacency, voxel adjacency.

connectivity check a computerized pro-cedure applied to a semiconductor chip’sphysical layout database which verifies theactual circuit on the chip is a correct im-plementation of the circuit described on theschematic diagrams of the chip.

consistency a correctness criterion basedon testing whether the result that is achievedby a set of operations being performed in par-allel is identical to the result that would beobtained if the operations were performed se-quentially without any overlap. Weaker testshave been proposed in order to trade hard-ware complexity for software responsibilityand faster execution.

consistency of interests situation inwhich there are several decision units withconsistent goals.Compare withdisagreementof interests.

consistency principle a principle frompossibility theory relating to the consistencybetween probability and possibility whichstates that the possibility of an event is al-ways at least as great as its possibility.

consistent estimator an estimator whosevalue converges to the true parameter valueas the sample size tends to infinity. If theconvergence holds with probability 1, thenthe estimator is called strongly consistent orconsistent with probability 1.

consistent goals objectives of several de-cision units in charge of a controlled parti-tioned system which, when followed, wouldlead jointly to overall optimal decisions (ac-tions) of these units; independent decisionmakers contributing to common objectives,with consistent goals, form a team.

constant angular velocity normally usedin disk storage units where the disk platterrotate at a constant rotational speed. Becauseof this, and to have the same amount of datain each track, sectors on the inner tracks aremore densely recorded than the outer tracks.

constant bit density on a disk, recordingpattern in which the number of bits per unitdistance is the same over all tracks.

constant bit rate (CBR) describes a traf-fic pattern in which the bits are sent at a fixedor constant rate. An 8-bit analog to digitalconverter sampling at 8 kilo-samples per sec-ond produces a CBR traffic stream with a bitrate of 64 kbps.

constant gain circle locus of input andoutput impedance points plots on the Smithchart that provide constant gain to an ampli-fier.

constant linear velocity used for exam-ple, in some optical disks where the platter ro-tates at different speed, depending on the rel-ative position of the referenced track. This al-lows more data to be stored on the outer tracksthan on the inner tracks. Because it takes timeto vary the speed of rotation, the method isbest suited for sequential rather than randomaccess.See alsoconstant angular velocity.


constant modulus algorithm (CMA)one of a number of algorithms (i.e., maxi-mum ratio combining, Bartlett, Capon, andLMS) proposed in the literature for the adap-tation of the weights associated with each ra-diating element and for combining signals re-ceived on radiating elements.

constant-current transformer two-coiltransformer with a moveable secondary coilused to provide constant output current toa variable load. Constant current is main-tained by mounting both the primary and sec-ondary coils on the center element of a shell-type core and allowing the secondary coil tomove up and down with changes in demandfor load current. Increasing current demanddue to a reduction in load impedance causesthe secondary coil to move away from theprimary coil. Increasing the coil separationincreases flux leakage and reduces the sec-ondary output voltage. The reduced outputvoltage counteracts the demand for more cur-rent. Increases in load impedance reverse theprocess. Movement of the secondary coil iscontrolled automatically by attaching the sec-ondary coil to a counterweight and pulley as-sembly and orienting the coil windings suchthat their flux directions oppose. Increasesin secondary current increase the magneticrepulsion between the coils, which, aided bythe counterweight, moves the secondary coilaway from the primary. Reductions in sec-ondary current produce the opposite effect.

constant-horse power drive a variablespeed drive that is operating in a speed regionwhere it is capable of delivering rated power.For DC machines, this region is above basespeed and is achieved by field weakening.For AC induction motors, this region is aboverated speed and is achieved by increasing thefrequency of the applied voltage.

constant-torque drive a variable-speeddrive that is operating in a speed region whereit is capable of maintaining rated torque. ForDC machines, this region is below base speedand is achieved by reducing the applied arma-

ture voltage. For AC induction motors, thisregion is below rated speed and is achievedby reducing the frequency of the applied volt-age.

constitutive relation describes the rela-tion between the intensity vectors and the fluxdensity vectors in a medium.

constitutive relationships a set of equa-tions that couple the electric-field intensityE,magnetic-field intensityH , the electric fluxdensityD, and the magnetic flux densityB toone another. For simple media, the constitu-tive relationships areD = εE, andB = µH ,whereε andµ are the scalar permittivity andpermeability of the medium, respectively.

constraining core an internal support-ing plane in a packaging and interconnectingstructure, used to alter the structure’s coeffi-cient of thermal expansion.

constraint length in convolutional codes,an indication of the number of source wordsthat affect the value of each coded word. Twotypical forms are:

1. A code with constraint lengthK, inwhich the value of each coded word is af-fected by the present source word and up toK − 1 previous source words.

2. The number of shifts over which a sin-gle message bit can influence the output of aconvolutional encoder.

constraint propagation artificial intelli-gence technique in which a hypothesis gener-ates constraints that reduce the search spaceover the rest of the data. If no eventual contra-diction is derived, then a “match” is achieved.

constructive algorithm learning algo-rithm that commences with a small networkand adds neural units as learning proceedsuntil the problem of interest is satisfactorilyaccommodated.


constructive solid geometry method bywhich complex 3-D objects are defined as thecombination of simpler solids.

contact head Seedisk head.

contact potential the internal voltagethat exists across a p-n junction under ther-mal equilibrium conditions, when no externalbias voltage is applied.

contact printing a lithographic methodwhereby a photomask is placed in directcontact with a photoresist coated wafer andthe pattern is transferred by exposing lightthrough the photomask into the photoresist.

contactor electromechanically actuatedspring-loaded relay contacts normally used tocontrol lights, heat, or other non-motor loads.In essence, it is an electromechanically oper-ated switch that usually requires some formof pilot device for its actuation.

containment building (1) a steel and con-crete structure which encloses and isolatesthe radioactive portion of a nuclear powerplant.

(2) a heavily re-inforced structure whichsurrounds the reactor and other radioactiveportions of a nuclear power plant so as to con-tain radioactive gases or debris in the eventof an explosion.

containment vessel the heavy steelcontainer which encloses the core of apressurized-water reactorcf in a nuclearpower plant.

content-addressable memory (CAM)Seeassociative memory.

contention additional latency incurredas the result of multiple requestors needingaccess to a shared resource, which can onlybe used by one at a time.

contention protocol class of a multi-ple access protocol where the users transmis-

sions are allowed to conflict when accessingthe communication channel. The conflict isthen resolved through the use of a static or dy-namic conflict resolution protocol. Static res-olution means that the conflict resolution isbased on some preassigned priority. A staticresolution can be probabilistic if the statisticsof the probabilities are fixed. A common ex-ample is the p-persistent ALOHA protocol.The dynamic resolution allows for changingthe parameters of the conflict resolution algo-rithm to reflect the traffic state of the system.A common example is the Ethernet protocol.

context the privilege, protection andaddress-translation environment of instruc-tion execution.

context switching an operation thatswitches the CPU from one process to an-other, by saving all of the CPU registers forthe first and replacing them with the CPUregisters for the second.

context units a set of memory units addedto a feedforward network that receives infor-mation when an input is presented to the net-work and passes this information to the hid-den layer when the next input is presented tothe network.

contingency analysis a plan for dealingwith any of the probable faults which mightbefall a particular electric power system, thegoal being to maintain power to the maxi-mum number of customers and/or the mostcritical customers.

contingency list in security analysis, alist, necessarily incomplete, of everythingwhich could possibly go wrong in a sectionof an electric power system.

contingency ranking the process of rank-ing the list of probable contingencies in orderof severity.

contingency selection the process of nar-rowing the list of probable contingencies, or


disturbances, that can be further processedand studied to determine the extent of secu-rity violations in the system.

continuity equation axiom that chargeis a conserved quantity. In point-form, thecontinuity equation is stated as

−∂ρ∂t

= ∇ · J,

whereρ is the charge density andJ is thecurrent density.

continuous duty National ElectricalManufacturers Association (NEMA) classi-fication describing an application in which amachine operates for long periods of time atrelatively constant loads.

continuous Hopfield network a Hop-field network with the same structure as thediscrete version, the one difference being thereplacement of the linear threshold units byneurons with sigmoidal characteristics. Anyinitial setting of the neuron outputs leads to amotion in the network’s state space towardsan attractor which, so long as the weights inthe network are symmetric, is a fixed point.This allows the network to be employed forthe solution of combinatorial optimizationproblems (its main application) by arrang-ing the network’s weights so that an optimalsolution lies at a fixed point of the network’sdynamics. Compare withdiscrete Hopfieldnetwork.

continuous rating term often used to re-fer to the manufacturer’s nameplate ratingsfor an electrical machine, which are the ratedoperating conditions guaranteed by the man-ufacturer for continuous-duty operation.Seealsocontinuous duty.

continuous signal a continuous functionof one or more independent variables suchas time, that typically contains informationabout the behavior or nature of some phe-nomenon.

continuous spectrum when an eigen-value problem is defined over an infinite do-main, the eigenvalues bunch together to forma continuum or a continuous spectrum. Thisconcept is of fundamental importance foropen waveguides either of electromagneticor acoustic type.

continuous speech recognition the pro-cess of recognizing speech pronounced natu-rally with no pauses between different words.

continuous system See incrementalgain.

continuous time signal Seesignal.

continuous time system a process thattransforms continuous time input signals tocontinuous time output signals.See alsosystem.

continuous tone image coding a processthat converts a digitized continuous tone im-age to a binary bit stream which has fewerbits than the original image for the purposeof efficient storage and transmission.Seestillimage coding.

continuous wave (CW) periodic and usu-ally sinusoidal wave, in contrast to a pulsedor modulated wave.

continuous-valued logic similarity byusing the definition of the equivalence incontinuous-valued logic, similarity betweentwo variablesx = (x0, . . . , xn) and y =(y0, . . . , yn), all components of which arecontinuous in the open interval(0,1), can bedefined as

SC(x, y) = ρ

√√√√ n∑i=1

(e (xi, yi))ρ

whereρ takes real value.See equivalence in continuous-valued

logicfor definition of e(xi, yi).

contour the edge that separates an objectfrom other objects and the background. It


must consist of one or several closed curves,one for the outer contour, and the others (ifany) for the inner contours surrounding anyholes.Seecontour filling, contour following,edge.

contour filling an object contour is gen-erally built with an edge detector, but sucha contour can be open, because some of itspieces, not recognized by the edge detector,may be missing. To close the contour, miss-ing pieces can be added by an operator fillingsmall holes in a contour.Seecontour, edgedetection.

contour following an operator which,starting from a contour point, follows theclosed curve made by that contour.Seecontour.

contrast (1) a measure of the intensitydifference (ratio) between an object and theimage background.

(2) the difference in the perception of vi-sual energy between picture white and pic-ture black. The ratio between the darkest andlightest portions of a TV picture.

contrast enhancement alteration ofthe contrast in an image to yield moredetails or more information. See alsocontrast, histogram stretching, histogramequalization.

contrast enhancement layer (CEL) ahighly bleachable coating on top of the pho-toresist that serves to enhance the contrast ofan aerial image projected through it.

contrast rendition factor (CRF) the ra-tio of visual task contrast with a given light-ing environment to the contrast with sphereillumination.

contrast sensitivity the responsiveness ofthe human visual system to low contrast pat-terns. In psychophysics, the threshold con-trast is the minimum contrast needed to dis-tinguish a pattern (such as a spatial sinusoid)

from a uniform field of the same mean lu-minance, and the contrast sensitivity is theinverse of the threshold contrast.See alsohuman visual system.

control intervention, by means of appro-priate manipulated inputs, into the controlledprocess in the course of its operation; someform of observation of the actual controlledprocess behavior is usually being used by thecontroller.

control and status register (CSR) aninternal CPU register that contains a packedbit array of I/O control information. CSRscan reside on I/O devices as well.

control bus contains processor signalsused to interface with all external circuitry,such as memory and I/O read/write signals,interrupt, and bus arbitration signals.

control channel the control channelused to transmit network control information.No user information is sent on this channel.Compare withtraffic channel.

control chart plot of data over time in-dicating the fluctuation of the main statisti-cal characteristics applied in statistical qual-ity control. Control charts can be used todetermine if a process is in a state of statis-tical control by examining past data and todetermine control limits that would apply tofuture data in order to check if the processmaintains in this state.

The individual observations are plottedagainst three lines. The center one representsan estimate of the process mean, standarddeviation or other statistic, two others rep-resent the lower control limit (LCL) and theupper control limit (UCL), respectively. Ifthe control charts are being used for the firsttime, it is necessary to determine trial con-trol limits.These limits should be revised ifthe points outside them are traced to a specialcause which can be removed. The most fre-quently used control charts areX, s (mean,standard deviation) andX,R (mean, range)


charts. Control charts are used to aid in iden-tification of special causes of variation, re-duction in product variability, and keepinggood records.

control horizon end time of the controlinterval over which the operation of the con-trol system is considered; if the control inter-val is infinite then the control horizon is alsoinfinite.

control input Seemanipulated input.

control instruction machine instructionthat controls the actions of a processor such assetting flags to enable specific modes of oper-ation. Generally, control instructions do notperform computations. Sometimes controlinstructions include instructions that can ef-fect sequential execution of a program, suchas branch instructions.

control interval time interval over whichthe operation of the control system is consid-ered; control interval can be finite or infinite.The notion of a control interval is essentialwhen the controlled process features variousaccumulation phenomena.

control layer part of the controller re-sponsible for performing tasks associatedwith a particular aspect of the control; par-ticular control layer results from vertical de-composition of the controller into a multi-layer control structure; the layers may differin their function, or the control interval con-sidered, or both. Typical examples of controllayers would be regulation layer of a continu-ous industrial process (see direct control) andset point control layer of such process.Seealsooptimizing control layer.

control line in a bus, a line used in acomputer bus to administrate bus transfers.Examples are bus request (a device wants totransmit on the bus), or bus grant (the busarbiter gives a device transmit access on thebus).

control memory a semiconductor mem-ory (typically RAM or ROM) used to holdthe control data in a microprogrammed CPU.This data is used to control the operation ofthe data path (e.g., the ALU, the data pathbusses, and the registers) in the CPU. If thecontrol memory uses RAM, the CPU is saidto be microprogrammable, which means thatthe CPU’s instruction set can be altered bythe user and the CPU can thus “emulate” theinstruction set of another computer. Same ascontrol store and micromemory.

control policy Seecontrol rule.

control problem a design problem con-cerned with constructing a device called thecontroller whose goal is to force the con-trolled variable of the plant or process to be-have in a desired manner. The elements ofthe control problem are control objective, amodel of the process to be controlled, admis-sible controllers, and a means of evaluatingthe performance of a control strategy.Seealsocontroller, controlled variable.

control rod an assembly of neutron-absorbing material, typically boron, which isextended into the core of a nuclear reactor todampen the chain reaction.

control rule decision mechanism (se-quence of such mechanisms), used — withinthe considered control layer — to specify on-line the values of the control inputs; for ex-ample, the values of the manipulated inputsin case of the direct control layer or the set-point values in case of the optimizing controllayer. Also known as control policy.

control scene initial entity, a given“world,” which is then partitioned into thecontrolled process and its (process) environ-ment; control scene is the initial “world”which is of interest to a control engineer or asystem analyst.

control store Seecontrol memory.


control structure essentially the same asthe control system; this term is used whenone wants to indicate that the controller iscomposed of several decision units, suitablyinterlinked; decision units of a control struc-ture usually differ in their tasks, scope of au-thority and access to information; dependingon the context one speaks of a control struc-ture or of a decision structure; decentralizedcontrol, multilayer control, hierarchical con-trol are examples of control structures.

control surface any of the movable partssuch as tabs, panels, or wings that controlthe depth of a submarine or the attitude ofa flight vehicle moving through the atmo-sphere. For example, the yaw angle of anairplane is controlled by the rudder, the pitchangle by the elevators, while the roll angleby the ailerons. In fuzzy logic community,control surface may mean a plot of a typi-cal fuzzy logic controller output as a func-tion of its two inputs. The inputs to a typicalfuzzy logic controller are the error betweenthe desired and the actual plant output, andthe change-in-error.

control system (1) the entity compris-ing the controlled process and the controller.Control system is influenced by the environ-ment of the process both through the free in-puts to the process itself and through any cur-rent information concerning the behavior ofthese free inputs that is made available to thecontroller.

(2) an arrangement of interconnecting ele-ments that interact and operate automaticallyto maintain a specific system condition orregulate a controlled variable in a prescribedmanner.

control transformer a step-down trans-former used to provide power to the controlportion of a power or motor circuit.

control under uncertainty operation ofthe control system in a situation when there isa significant uncertainty regarding current orfuture values of the free inputs, which leads

to uncertainty in the envisaged behavior ofthe controlled process and, eventually, thereis uncertainty concerning the internal behav-ior of the controlled process; in particularwhen the forecasted scenarios of the futurefree input values tend to appear to be largelydifferent from the actual future free input re-alization.

controllability (1) the property of a sys-tem that ensures the existence of boundedcontrol inputs to drive any arbitrary initialstate to any arbitrary final state in finite time.For linear systems, an algebraic conditionthat involves system and input matrices canbe used to test this property.

(2) the ability to establish the required teststimuli at each node in a circuit by settingappropriate values on the circuit inputs.

controllability at a given time a charac-teristic of some dynamical systems. A lineardynamical system is said to be controllableat a given time if there exists a finite timet1,such that it is controllable in a time interval[t0, t1].

controllability condition for nonstation-ary discrete system a condition foundin some dynamical systems. A linear dy-namical nonstationary discrete-time systemis controllable in an interval[k0, k1] if andonly if the controllability matrix

k1−1∑j=k0

F (k1, j + 1) B(j)BT (j)F T (k1, j + 1)

is nonsingular. See alsodynamical linearnonstationary discrete-time finite-dimensionalsystem.

controllability in a fixed interval a char-acteristic of some dynamical systems. A dy-namical discrete system is said to be control-lable in an interval[k0, k1] if for any initialstatex(k0) ∈ Rn and any vectorx1 ∈ Rn

there exists a sequence of admissible con-trolsu(k), k = k0, k0 +1, . . . , k1 −2, k1 −1such that the corresponding trajectory of the


dynamical system satisfies the condition

x (k1, x (k0) , u) = x1

controllability in a given time intervala characteristic of some dynamical systems.A dynamical continuous system is said to becontrollable in a given time interval[t0, t1] iffor any initial state

x (t0) ∈ Rn

and any vector

x1 ∈ Rn

there exists an admissible control

u ∈ L2 ([t0, t1] , Rm)

such that the corresponding trajectory of thedynamical systemx(t, x(t0), u) satisfies thefollowing condition:

x (t1, x (t0) , u) = x1

controllability of nonstationary systemsa characteristic of some dynamical systems.A linear dynamical nonstationary system iscontrollable in time interval[t0, t1] if andonly if the n × n dimensional controllabil-ity matrix

W (t0, t1) =∫ t1

t0

F (t1, t) B(t)

× BT (t)F T (t1, t) dt

is nonsingular. See alsodynamical linearnonstationary continuous-time finite-dimensional system.

controllability of stationary systems acharacteristic of some dynamical systems. Alinear dynamical stationary system is control-lable if and only if

rank[B|AB|A2B| . . . |AkB| . . . |An−1B

]= n

See also dynamical linear stationarycontinuous-time finite-dimensional system.

controlled process part of the controlscene that can be influenced by the manipu-lated inputs set by the controller and the freeinputs from the (process) environment; ma-nipulated inputs will be set (adjusted) to real-ize specified control objectives. Control ob-jectives can be stated in terms of constraintsimposed on specified quantities, in terms ofperformance indices to be minimized or max-imized, or in other form; free inputs to theprocess from its environment do not depend,by definition, on the process behavior.

controlled process model model, usu-ally stated in form of a set of differential ordifference equations, describing the behaviorof the controlled process as caused by its in-puts; different models of the same controlledprocess may be used for various purposes,for steady-state control or for model-basedpredictive control.

controlled rectifier a rectifier that usesswitching elements that have forward volt-age blocking capability to allow a variablevoltage DC output.See alsothyristor.

controlled source a voltage or currentsource whose intensity is controlled by a cir-cuit voltage or current elsewhere in the cir-cuit. Also called dependent source.

controlled variable (1) the quantity, usu-ally the output of a plant or process, that is be-ing controlled for the purpose of the desiredbehavior, for example, transient response orsteady-state response.

(2) variable associated with the behaviorof the controlled process and such that onewants this variable either to follow a desiredtrajectory over a given time interval or to bekept at a prescribed constant value, i.e., at aspecified set-point; introduction of a set ofcontrolled variables is necessary to define atwo-layer industrial controller with the reg-


ulation direct control layer and the set pointoptimizing control layer.

See alsocontroller.

controller (1) the entity that enforces thedesired behavior — as specified by the con-trol objectives — of the controlled process byadjusting the manipulated inputs. The valuesof these inputs are either predetermined ordecided upon (computed) using on-line, i.e.,real time, decision mechanism of the con-troller — based on the currently available in-formation.See alsocontrolled variable.

(2) a device that generates the input to theplant or process. The role of the controlleris to force the controlled variable of the plantor process to behave in a desired manner.

(3) a unit that directs the operation of asubsystem within a computer. For instance,a disk controller interprets data access com-mands from host computer (via a bus), andsends read/write, track seeking, and othercontrol signals to the drive. During this time,the computer can perform other tasks, un-til the controller signals DATA READY fortransfer via the CPU bus.

convection current a current in whichelectrons are released for movement outsideof a material.

convective heat transfer the process bywhich a moving fluid transfers heat to or froma wetted surface.

convergence the condition when the elec-tron beams from a multi-beam CRT meet at asingle point. For example, the correct regis-tration of the three beams in the color picturetube.

convergence in probability for some se-quences of random numbers, the tendency toa single number.

To wit, for a sequence of numbersxn, anda random variablex, if for all ε > 0,

P {| xn − x |> ε} → 0

for n → ∞, then the sequencexn tends toxin probability.

convergent state the equilibrium state ofa dynamic system described by a first ordervector differential equation is said to be con-vergent if there exists aδ = δ(t0), such that,

‖ x (t0)− xe ‖< δ ⇒ limt→∞ x(t) = xe

See alsostable state.

converter a generic term used in thearea of power electronics to describe a recti-fier, inverter, or other power electronic devicethat transforms electrical power from one fre-quency and voltage to another.

convex fuzzy set (1) a fuzzy set that hasa convex type of membership function.

(2) a fuzzy set in which allα-level sets areconvex.See alsoα-level set.

convolution the mathematical operationneeded to determine the response of a sys-tem from its stimulus signal and its weightingfunction. The convolution operation is de-noted by the symbol “∗.” The convolution oftwo continuous time signalsf1(t) andf2(t)

is defined by

f1(t) ∗ f2(t) =∫ ∞

−∞f1(τ )f2(t − τ)dτ

=∫ t

0f1(τ )f2(t − τ)dτ if f1(t), f2(t)

= 0, t < 0

The integral on the right-hand side of theabove equation is called the convolution in-tegral, and exists for allt ≥ 0 if f1(t) andf2(t) are absolutely integrable for allt > 0.f1 is the weighting function that character-izes the system dynamics in the time domain.It is equivalent to the response of the systemwhen subjected to an input with the shape of aDirac delta impulse function. Laplace trans-formation of the weighting function yieldsthe transfer function model for the system.


The convolution of two discrete time sig-nalsf1[k] andf2[k] is defined by

f1[k] ∗ f2[k] =∞∑

i=−∞f1[i]f2[k − i]

=k∑i=0

f1[i]f2[k − i] if f1[k], f2[k] = 0, k < 0

The summation on the right-hand side of theabove equation is called the convolution sum.Convolution is useful in computing the sys-tem output of LTIL systems.

convolution integral Seeconvolution,superposition integral.

convolutional code (1) a code generatedby passing the information sequence to betransmitted through a linear finite-state shiftregister and the coder memory couples thecurrently processed data with a few earlierdata blocks. Thus the coder output dependson the earlier data blocks that have been pro-cessed by the coder.

(2) a channel code based on the trellis cod-ing principle but with the encoder function(mapping) determined by a linear function(over a finite alphabet). The name “convo-lutional code” comes from the fact that theoutput sequence is a (finite alphabet) convo-lution between the input sequence and theimpulse response of the encoder.

convolutional coding a continuous er-ror control coding technique in which con-secutive information bits are combined al-gebraically to form new bit sequences to betransmitted. The coder is typically imple-mented with shift register elements. Witheach successive group of bits entering theshift register a new, larger set of bits are cal-culated for transmission based on current andprevious bits. If for everyk information bitsshifted into the shift register, a sequence ofn

bits are calculated, the code rate isk/n. Thelength of the shift register used for storinginformation bits is known as the constraintlength of the code. Typically, the longer the

constraint length, the higher the code protec-tion for a given code rate.See alsoblockcoding, error control coding.

cooling tower a reinforced-concrete,spool-shaped structure in a thermal powerplant in which condenser cooling wateris itself cooled by convectively-driven airstreams.

Cooper, Leon Niels (1930– ) Born: NewYork, New York, U.S.A.

Cooper is best known for his work on su-perconductivity, resulting in a Nobel Prizehe shared with John Bardeen and J. RobertSchrieffer. He postulated the idea that is nowknown as Cooper pairs, as part of the expla-nation of why some metals loose their con-ductivity at very low temperatures.

coordinate system a system for definingthe location of a point in space relative tosome reference point and for defining a set ofreference directions at each and every pointin space.

coordinated rotation digital computer(CORDIC) algorithm for calculatingtrigonometric functions using only additionsand shift operations.

coordinating unit Seecoordinator unit.

coordination process of influencing —by the coordinator — the local units in sucha way as to make their behavior or deci-sions consistent with the objectives of thecoordinator unit; coordination can be eitheriterative or periodic; iterative coordinationplays an essential role in multilevel optimiza-tion. See alsodirect method, mixed methodcoordination, price method coordination.

coordination by exception coordinationperformed only in unusual (emergency) situ-ations, otherwise the control or decision mak-ing is designed to be fully decentralized.


coordination instrument any means bywhich the coordinator unit influences behav-ior or decisions of the subordinate (local)units; coordination instruments are often ex-pressed in terms of vectors of values assumedby the coordination variables. In the case ofprice method, the coordination variables arethe prices by which the interaction input andoutput variables are multiplied.

coordination strategy mechanism (algo-rithm) used to generate the values of the co-ordination instruments in the course of eitheriterative or periodic coordination; in case ofiterative coordination, the convergence andthe speed of convergence of the coordinationstrategy is of main concern; in case of peri-odic coordination, other issues are important— these can be stability of the coordinationprocess as well as quantitative aspects of thecontrolled process behavior.

coordinator unit control (decision) agentin a hierarchical control structure, being incharge of decisions (control instruments) in-fluencing operation of the local decisionunits; coordinator unit performs either itera-tive or periodic coordination of the local de-cisions; coordinator unit is often regarded asthe supremal unit of the hierarchical controlstructure. Also called coordinating unit.

copolarized the plane wave whose po-larization is the same as that of the referenceplane wave (e.g., radiated wave from an an-tenna) is said to be copolarized (otherwise itis crosspolarized).

copper jacket timer a magnetic time-offtimer that can be used in definite time DCmotor acceleration starters and controllers.The copper jacket relay functions by slowingthe dissipation of the magnetic field when thecoil is turned off. After a certain amountof time the spring tension on the contac-tor overcomes the strength of the dissipat-ing magnetic field — and causes the con-tacts to change state. Time delays with thecopper jacket timer are adjusted by adding,

or removing, permeable shims between thecoil and copper jacket. The more shims inplace the slower the magnetic field dissipates,hence the longer the time delay becomes.

copper loss electric loss due to the re-sistance in conductors, windings, brush con-tacts or joints, in electric machinery or cir-cuits. Also referred to asI2R, the losses aremanifested as heat.

coprime 2-D polynomial matrices Seecoprimeness of 2-D polynomial matrices.

coprime 2-D polynomials Seecoprime-ness of 2-D polynomials.

coprimeness of 2-D polynomial matricesa mathematical relationship of interest in

control systems.A 2-D polynomial matrixC(z1, z2) (B(z1,

z2)) is called a right (left) divisor ofA(z1, z2)

if there exists a matrixB(z1, z2) (C(z1, z2))

such that

A (z1, z2) = B (z1, z2) C (z1, z2)

A 2-D polynomial matrixR(z1, z2) (L(z1,

z2)) is called a common right (left) divisor ofA(z1, z2), andB(z1, z2) if there exist two 2-Dpolynomial matricesA1(z1, z2), B1(z1, z2),(A2(z1, z2), B2(z1, z2)) such that

A (z1, z2) = A1 (z1, z2) R (z1, z2)

and

B (z1, z2) = B1 (z1, z2) R (z1, z2)

(A (z1, z2) = L (z1, z2) A2 (z1, z2)

and

B (z1, z2) = L (z1, z2) B2 (z1, z2))

A 2-D polynomial matrix P(z1, z2),

(Q(z1, z2)) is called greatest common right(left) divisor of A(z1, z2) and B(z1, z2) ifP(z1, z2) (Q(z1, z2)) is a greatest right(left) divisor ofA(z1, z2) andB(z1, z2) andany common right divisor ofA(z1, z2) andB(z1, z2) is a right divisor ofP(z1, z2).


2-D polynomial matricesA(z1, z2), B(z1,

z2) are called factor right (left) coprime iftheir greatest common right (left) divisoris a unimodular matrixU(z1, z2) (nonzerodetU(z1, z2) ∈ R).

2-D polynomial matricesA ∈ Rp×m[z1, z2],B ∈ Rp×m [z1, z2] (p+q ≥ m ≥ 1)are called zero right coprime if there exists apair (z1, z2) which is a zero of allm × m

minors of the matrix[A

B

]. The minor co-

primeness of two 2-D polynomial matricesimplies their factor coprimeness.

coprimeness of 2-D polynomials a math-ematical relationship of interest in controlsystems.

A 2-D polynomial

p (z1, z2) =n2∑i=0

ai (z1) zi2 =

n1∑j=0

aj (z2) zj

1

is called primitive ifai(z1), i = 1,2, . . . , n2and aj (z2), j = 1,2, . . . , n1 are coprime(have not a common factor). Two primitive2-D polynomials are called factor coprime iftheir greatest common divisor is a constant.The primitive 2-D polynomials

a (z1, z2) = an1 (z2) a (z1, z2)

= an2 (z1) a (z1, z2)

b (z1, z2) = bm1 (z2) b (z1, z2)

= bm2 (z1) b (z1, z2)

a (z1, z2) = zn11 + an1−1z

n1−11 + · · ·

+ a1z1 + a0 (ai = ai (z2))

a (z1, z2) = zn22 + an2−1z

n2−12 + · · ·

+ a1z2 + a0(ai = ai (z1)

)b (z1, z2) = z

m11 + bm1−1z

m1−11 + · · ·

+ b1z1 + b0(bi = bi (z2)

)b (z1, z2) = z

m22 + bm2−1z

m2−12 + · · ·

+ b1z2 + b0

(bi = bi (z1)

)are factor coprime if and only if

det[Bn1 + an1−1B

n1−1 + · · ·+ a1B + a0Im1

] 6= 0

and

det[Bn2 + an2−1B

n2−1 + · · ·+ a1B + a0Im2

]6= 0

or

det[Am1 + bm1−1A

m1−1 + · · ·+ b1A+ b0In1

] 6= 0

and

det[Am2 + bm2−1A

m2−1 + · · ·+ b1A+ b0In2

]6= 0

where

A =

0 1 0 . . . 00 0 1 . . . 0. . . . . . . . . . . . . . .

0 0 0 . . . 1−a0 −a1 −a2 . . . −an1−1

A =

0 1 0 . . . 00 0 1 . . . 0. . . . . . . . . . . . . . .

0 0 0 . . . 1−a0 −a1 −a2 . . . −an2−1

B =

0 1 0 . . . 00 0 1 . . . 0. . . . . . . . . . . . . . .

0 0 0 . . . 1−b0 −b1 −b2 . . . −bm1−1

B =

0 1 0 . . . 00 0 1 . . . 0. . . . . . . . . , , , . . .

0 0 0 . . . 1−b0 −b1 −b2 . . . −bm2−1

coprocessor a processor that is connectedto a main processor and operates concur-rently with the main processor, although un-der the control of the main processor. Copro-cessors are usually special-purpose process-ing units, such as floating point, array, DSP,or graphics data processors.

copy-back in cache systems an opera-tion that is the same as write-back—a write


operation to the cache that is not accompa-nied with a write operation to main memory.In copy-back, the data is written only to theblock in the cache. This block is written tomain memory only when it is replaced by an-other block.

CORDIC Seecoordinated rotation digitalcomputer.

core (1) the operating image of a pro-cess (sometimes used to refer to the part re-siding in physical memory), often written todisk if the program crashes (dumping core).Since magnetized ferrite rings (cores) wereonce used in main memory to store a singlebit each. The name remained and now corememory means the same as main memory,although currently, main memory is chip-based.See alsomagnetic core memory.

(2) the ferromagnetic portion of a trans-former or electric machine on which the coilsare mounted. Typically made of laminatedmagnetic material, encircled by the wind-ings, that provides a low reluctance path formagnetic flux.

(3) the central region of an optical fiber.The refractive index of the core must behigher than that of the cladding so that theoptical power is guided through the fiber bytotal internal reflection at the core-claddingboundary. The core refractive index may beconstant or may decrease with distance fromthe axis to the cladding.See alsogradedindexandstep index optical fiber.

(4) the section of a nuclear reactor inwhich the chain reaction is contained, com-prising fuel rods, control rods, moderator,and coolant.

core lamination Seelamination.

core loss loss in the ferromagnetic ma-terial comprising the core of an electric ma-chine or transformer, composed of the sumof hysteresis losses and eddy current losses.These magnetic losses are caused by timevarying fluxes in a ferromagnetic structure.Hysteresis losses are caused by friction in

molecules as the dipoles in a structure changedirection of alignment in response to an ap-plied alternating voltage, while eddy currentlosses are resistive losses(I2R), due to cir-culating currents in the core.

core memory See magnetic corememory.

core-type transformer a transformer inwhich the magnetic circuit upon which thewindings are wound takes the form of a singlering. When the coils are placed on the core,they encircle the core.See alsocore.

coriolis forces forces/torques that dependupon the product of joint velocities.

corner cube antenna a traveling waveantenna typically four wavelengths long andplaced 1.2 wavelengths from the apex of a90-degree corner reflector other used in thesubmillimeter wave and terahertz frequencyrange for mixers and detectors.

corner detection the detection of corners,often with a view to locating objects fromtheir corners, by a process of inference, indigital images.

corona a visible glow discharge whichemanates from high-voltage conductors in re-gions of extremely high electric field inten-sity.

corona effect flow of electrical energyfrom a high-voltage conductor to the sur-rounding ionized air. This effect only be-comes significant for potentials higher than1000 V. This effect is characterized by a faintglow, a crackling noise and conversion of at-mospheric oxygen to ozone.

corona loss the electric power lost in highvoltage lines due to the radiation of energyby corona discharge.Seecorona.

corona resistance capacity of a mate-rial to bear the action of the corona effect.


This capacity is particularly important forpolymeric materials, which have to withstandthe chemical degrading effect promoted byozone generated by the corona effect.Seealsocorona effect.

corona ring a toroidal metal ring con-nected to discontinuities on high-voltageconductors to reduce local field intensity andthus discourage the formation of corona dis-charge.

coronal projection a projection imageformed on a plane parallel to the chest andperpendicular to the transverse and sagittalplanes.

corporate feed also known as parallelfeed, a method of feeding a phased array an-tenna in which the signal from a transmitter issuccessively split until there are enough feed-lines to carry the signal to each array element.The name comes from the resemblance of thefeed structure to the organizational chart of acorporation.

correction layer control layer of a mul-tilayer controller, usually situated above thedirect control layer and below the optimizingcontrol layer, required to make such modifi-cations of the decisions supplied by the op-timizing layer — before these decisions arepassed to the direct layer — that some speci-fied objectives are met; for example, a correc-tion layer of the industrial process controllermay be responsible for such adjustment of aparticular set point value that an importantconstraint is satisfied by the controlled pro-cess variables — in case when the optimizinglayer is using inaccurate model of this con-trolled process.

correlation (1) the mathematical opera-tion of comparing the behavior of two signalsto determine how closely they are related. Itis usefully applied to system identificationwhen the input and output signals of a givensystem are compared by correlation to givethe system transfer function model. The rel-

evant equation is

g(jω) = y(jω)u∗(jω)u(jω)u∗(jω)

in terms of the FFTs of the input and outputsignals. The∗ superscript indicates the com-plex conjugate. As the system outputy(t)generally contains terms in addition to thosecaused by the inputu(t), its Laplace trans-form is

y(s) = g(s)× u(s)+ d(s)

Thus the correlation of input and output canenhance the accuracy of the estimate for thesystem frequency response functiong(jω),since the correlation function then yieldsthe following estimate of the actual processmodel:

g(jω) = g(jω)+ d(jω)u∗(jω)u(jω)u∗(jω)

Clearly, the effect of the disturbanced(t) isreadily reduced if the input is chosen eitherto be large in amplitude, compared to the dis-turbance, or to be pseudo-random so that itdoes not correlate with the disturbance andthe second numerator becomes zero, by def-inition.

(2) the temporal or spatial function orsequence resulting from integrating laggedproducts of two signals, i.e., the products ofa signal with the time-inverted version of asecond signal at various relative delays withrespect to each other. If the two signalsare identical, a maximum correlation outputamplitude results at full overlap.See alsoconvolution.

correlation bandwidth the frequencyfor which the autocorrelation of the transferfunction reaches a given threshold, for exam-plex% of the central value.

correlation coefficient a measure of theability to predict one random variablex as thelinear function of anothery. The correlationcoefficient

ρ = E[xy] − E[x]E[y]σxσy


satisfies−1 ≤ ρ ≤ 1, where|ρ| = 1 impliesa deterministic linear relationship betweenxandy, andρ = 0 implies lack of correlation.See alsocorrelation.

correlation detector the optimum struc-ture for detecting coherent signals in the pres-ence of additive white Gaussian noise.

correlation function a mathematical de-scription that describes the common relation-ship between two processes. For a singletime dependent quantity, the autocorrelationfunction describes the loss of information inthe process description of the quantity.Seeautocorrelation function.

correlation peak detector a photodetec-tor array that detects and outputs only thepeak in a spatial correlation function.

correlation receiver a receiver utilizingthe autocorrelation or cross-correlation prop-erties of the transmitted signal to detect thedesired information from the received signal.A correlation receiver is theoretically equiv-alent to a matched filter receiver. Correlationreceivers are often used in spread spectrumsystems and channel measurement systems.

correlation sidelobes the correlationfunction other than the peak; particularlywhere signals are designed to give a sharplypeaked correlation function.

correlation similarity the (unnormal-ized) correlation of two real vectorsx =(x0, . . . , xn) and (y0, . . . , yn) is defined astheir inner product:

C =n∑i=1

xiyi .

correlator a circuit that calculates thecorrelation function. See alsocorrelationfunction.

correspondence problem the problem ofmatching points in one image with their cor-responding points in a second image.

corresponding point a point, in a set ofimages representing a different view of thesame scene, onto which the same physicalpoint in the real world is projected.

corrugated horn a horn with grooves onits inner walls. The grooves create the sameboundary conditions on all four walls of thehorn. This results in the elimination of spuri-ous diffraction at the edges of the horn aper-ture.

coset leader each possible error patternof the code word representing the left-mostcolumn of the standard array.

cosine modulated filter bank a filter bankwith each of its analysis filters and synthe-sis filters being the modulation of a low-passprototype filter by a cosine function. In thefrequency domain, it is equivalent to shiftingthe low-pass prototype filter by different fre-quencies to form a bank of bandpass filterscovering the entire frequency band.

cosine roll-off filter a filter that has animpulse response which satisfies the NyquistI criterion for zero intersymbol interference.The filter has the following transfer function.

H(ω) =

1, |ω| ≤ π(1−α)

T

cos2 T4α(

|ω| − π(1−α)T

),π(1−α)T

< |ω| < π(1+α)T

0, |ω| ≥ π(1+α)T

whereT is the transmitted symbol period andα is a parameter that is known as the excessbandwidth (0≤ α ≤ 1). The caseα = 0yields the ideal low-pass filter, and the caseα = 1 yields a filter referred to as the raisedcosine filter.

cosine transform a transform that con-sists of a set of basis functions, which are co-sine functions. Usually referred to as discretecosine transform.See alsodiscrete cosinetransform.


cost function a nonnegative scalar func-tion that represents the cost incurred by aninaccurate value of the estimate. Also calledpenalty function.

Costas loop a carrier synchronizationloop in a digital communications receiver thatuses a quadrature phase detector in place ofa conventional square-law device.

cotree the complement of a tree in a net-work.

Cotton–Mouton effect second-orderanisotropic reciprocal magnetooptic effectthat causes a linearly polarized incident lightto transmit through as an elliptically polar-ized output light wave when the propagationdirection of the incident light is perpendicu-lar to the direction of the applied magnetiza-tion of the magnetooptic medium. It is alsoknown as magnetic linear birefringence.

Coulomb blockade the situation in whicha particle has insufficient thermal energy toallow the necessary energy exchange duringa tunneling process. Hence, the bias supplymust supply energy to the electron to accountfor the stored energy change in tunneling,which requiresV > e/2C, whereC is thecapacitance (eV << kBT ).

Coulomb force electric force exerted onan electrically charged body, which is pro-portional to the amount of the charge and theelectric field strength in which the chargedbody is placed.

Coulomb’s Law the force of repul-sion/attraction between two like/unlikecharges of electricity concentrated at twopoints in an isotropic medium is proportionalto the product of their magnitudes and in-versely proportional to the square of the dis-tance between them and to the dielectric con-stant of the medium.

Coulomb, Charles (1763–1806) Born:Angouleme, France

Coulomb is best known for his study ofelectric charge and magnetism resulting inCoulomb’s Law, as well as his studies in fric-tion. Coulomb also invented the torsion bal-ance in 1777. He used this device in manyexperiments. Coulomb began his career inthe military, but resigned when the FrenchRevolution began. His experience as a mil-itary engineer involved him in a wide vari-ety of different projects. It also gave himtime to continue his own experimental work.Coulomb’s law states that the force betweentwo charges is proportional to the productof the charges and inversely proportional tothe square of the distance between the twocharges. Coulomb is honored by having hisname used as the unit of electric charge, thecoulomb.

counter (1) a variable or hardware regis-ter that contains a value that is always incre-mented or decremented by a fixed amount,and always in the same direction (usually in-cremented by one, but not always).

(2) a simple Moore finite state machinethat counts input clock pulses. It can be wiredor enabled to count up and/or down, and invarious codes.

counter-EMF a voltage developed in anelectrical winding by Faraday’s Law that op-poses the source voltage, thus limiting thecurrent in the winding.

counter-EMF starter a type of DC-motorstarter that reduces the resistance in the start-ing circuit as the voltage across the armaturerises.

counter-propagation learning Seehi-erarchical feature map.

counter-rotating field theory mathemat-ical theory in which a magnetic field withstationary spatial direction but sinusoidallyvarying magnitude is decomposed into twoconstant-magnitude fluxes rotating in oppo-site directions. The angular velocity,ω, ofthe rotating fluxes equals the time-domain


angular frequency of the stationary flux, andthe magnitude of the rotating fluxes is halfthe magnitude of the stationary flux. Mathe-matically, this is described by the followingequation:

B cosθ sinωt = 0.5B sinωt − θ

+ 0.5B sinωt + θ

where the left-hand term represents the time-varying flux magnitude along the spatial di-rectionθ , and the right-hand terms representforward and backward rotating fields of con-stant magnitude rotating aboutθ . The fol-lowing diagram illustrates this relationship.Resolution of the stationary flux into two ro-tating fluxes allows induction machine per-formance to be analyzed using standard ro-tating field theories. The total machine per-formance is determined from the net resultof both rotating fluxes. Also called doublerevolving-field theory.

Counter-rotating field concept.

counter-torque torque developed in op-position to the rotation of a machine. It isproduced as load current flows in the pres-ence of and perpendicular to magnetic flux ina machine that is generating electric power.

counterpoise a ground wire buried be-neath an overhead line to lower footingimpedance.

counterpoise ground buried conductorrouted under transmission lines designed toachieve low earth electrode resistance.

coupled inductor two inductors within aswitching converter, which have similar volt-age waveforms, are wound on the same mag-netic core to steer the ripple from one wind-ing to another. A typical example is aCukconverter with an isolation transformer wherethe input and output inductors are coupledwith the transformer to provide zero rippleat the input and the output.See alsoCukconverter.

coupled line filter a type of microstripor stripline filter that is composed of paralleltransmission lines. Bandwidth is controlledby adjusting the transmission line spacing.Wider bandwidths are obtained by tightercoupling. A two-port circuit is formed byterminating two of the four ports in eitheropen or short circuits, which leaves ten pos-sible combinations. Different combinationsare used to synthesize low-pass, bandpass, allpass, and all stop frequency responses.

coupled lines the electromagnetic field oftwo unshielded transmission lines in prox-imity can interact with each other to formcoupled lines. Usually three conductors areneeded. Examples of coupled lines are cou-pled microstrip lines and coupled striplines.

coupled Riccati equations a set of dif-ferential or algebraic matrix equations ofthe Riccati type arising in the jump linearquadratic problem which should be solvedsimultaneously. The set of coupled differen-tial Riccati equations has a form

K(i, t)+ A(i)′K(i, t)+K(i, t)A(i)s∑j=1

qijK(j, t)+Q(i)

−K(i, t)B(i)R(i)−1B(i)′K(i, t) = 0

for i ∈ S = {1,2, . . . , s}, t ∈ [0, T ]and terminal conditionK(i, T ) = 0 where


A(i), B(i),Q(i), R(i) are, respectively, sys-tem and weighting, state and input matrices inthe modei, andqij is the transition rate frommodei to j . If the system is stochasticallystabilizable and the pairs(A(i),

√Q(i)) are

observable for alli where√Q(i) = C(i)

such thatC(i)′C(i) = Q(i), then the solu-tionK(i, t) is finite for eachi and

limT−t→∞K(i, t) = K(i)

whereK(i) is the set of solutions to the cou-pled algebraic Riccati equations:

A(i)′K(i)+ K(i)A(i)+s∑j=1

qij K(j)

+Q(i)− K(i)B(i)R(i)−1B(i)′K(i) = 0

Coupled Riccati equations are also used inN-person linear quadratic games. In thiscasei = 1,2, . . . , N is the number of theplayer and the set of coupled differential ma-trix equations has the following form:

K(i, t)+ A′K(i, t)+K(i, t)A+Q(i)

−K(i, t)

j=N∑j=1

BR(j)−1B ′K(j, t) = 0

for i ∈ {1,2, . . . , N}, t ∈ [0, T ] and termi-nal conditionK(i, T ) = 0 whereA,B are,respectively, state and input matrices in thestate equation,Q(i), R(i)are weighting stateand input matrices for thei-th player, and itssolution defines the gain in a Nash equilib-rium in the game.

coupled wave equations a special caseof Maxwell’s equations describing the prop-agation of two interacting electromagneticfields in a nonlinear material. For example,in four wave mixing or optical phase conju-gation, the probe and conjugate fields oftenobey coupled wave equations.

coupler (1) a passive, wavelength-insensitive, fiber optic component that com-bines all inputs and distributes them to theoutputs with a defined splitting ratio. The

most common types are termed fused taperand dichroic.

(2) a device inserted into an optical fibercommunications link that is used to insert ad-ditional optical energy or signals onto the linkand/or tap off some of the energy or signalon the link to another optical fiber. A cou-pler may also be used to combine power fromtwo or more input fibers into one output fiberor to distribute the combined input power to anumber of output fibers. Couplers are madein 1×N ,N×N , andN×M (input)×(output)configurations whereN andM are the num-ber of fibers. The power combining and/orsplitting ratios can vary from 50%/50% (a 3dB coupler) in a symmetric 2× 2 coupler to> 99%/ < 1% in an asymmetric coupler.

coupling capacitor voltage transformer(CCVT) a potential transformer that usesthe impedance of a small capacitance to re-duce the power line voltage to measureablelevels.

coupling factor in a coupler, the ratio ofpower in the coupled port to that applied atthe input port.

coupling efficiency the efficiency towhich a signal can be coupled from one trans-mission line or resonator to another.

coupling field a region of interactionwhere energy is transferred between systems.The energy may be transferred between sys-tems of a similar nature (i.e., electrical–electrical) or between systems of a differentnature (i.e., electrical–mechanical).

coupling loss measure of the power dis-sipated as a result of coupling a signal fromone transmission line or resonator to another.

covariance The expectation of the productof two mean-removed random quantities:

Cfg(t1, t2) = E[f (t1)g(t2)

T]

− E[f (t1)]E[g(t2)T ].


See alsoautocovariance,correlation, variance.

covert channel a mechanism by means ofwhich information can be communicated in-directly despite design features that preventdirect communication. Many covert chan-nels utilize side effects of operations, suchas effects on timing or scheduling that canbe seen from programs or processes that arenot supposed to be direct destinations of anycommunications from a sender that in factperforms operations that emit signals by af-fecting timing or scheduling within the sys-tem.

COVQ See channel optimized vectorquantization.

cow magnet a long, pill-shaped alnicomagnet that is placed in the second stomachof a cow to capture ferrous debris and preventit from passing through the digestive system.

CPA acronym for compensated pulsed al-ternator.Seecompulsator.

CPI Seecolor preference indexor cyclesper instruction.

CPM laser Seecolliding-pulse-modelockedlaser.

CPU Seecentral processing unit.

CPU time the time that is required to com-plete a sequence of instructions. It equals tothe (cycle time)× (number of instructions)× (cycles per instruction).

CR Seecavity ratio.

Cramer-Rao bound a lower bound onthe estimation error covariance for unbiasedestimators. In particular, the estimation errorcovariance3e(x), which is a function of theunknown quantityx to be estimated, mustsatisfy3e(x) ≥ I (x), whereI is the Fisherinformation matrix. If an estimator achieves

the Cramer-Rao bound with equality for allx,it is efficient; if an efficient estimator exists,it is the maximum likelihood estimator.Seealsobias, maximum likelihood estimation.

crash an incorrect operation of computersoftware or hardware leading to temporary orpermanent loss of part of the data.

CRC Seecyclic redundancy check.

CRC character a type of error detectioncode commonly used on disk and tape storagedevices. Data stored on a device using CRChas an additional character added to the endof the data that makes it possible to detectand correct some types of errors that occurwhen reading the data back.

CRC-code code that employs cyclic re-dundancy checking.Seecyclic redundancycheck.

CRCW Seeconcurrent read and concur-rent write.

credit assignment problem during neu-ral network learning, the problem of deter-mining how to apportion credit (blame) toindividual components for network behaviorthat is appropriate (inappropriate) to the taskbeing learned.

crest factor the ratio of the peak value ofa signal to its RMS value.

CREW Seeconcurrent read and exclusivewrite.

CRF Seecontrast rendition factor.

crisp set in fuzzy logic and approximatereasoning, this term applies to classical (non-fuzzy Boolean) sets that have distinct andsharply defined membership boundaries.Seealsofuzzy set.

critical refers to the state of a chain reac-tion which is self-sustaining but which pro-


duces just enough free neutrons to compen-sate for those lost to the moderator and leak-age.

critical angle the incidence angle, definedby Snell’s law, where the incident wave is to-tally reflected at the interface of two differentdielectric media.

critical band broadly used to refer to psy-choacoustic phenomena of limited frequencyresolution in the cochlea. More specifically,the concept of critical bands evolved in ex-periments on the audibility of a tone in noiseof varying bandwidth, centered around thefrequency of the tone. Increasing the noisebandwidth beyond a certain critical value haslittle effect on the audibility of the tone.

critical clearing angle (1) following abalanced three-phase fault at the stator ter-minals of a synchronous machine, the maxi-mum value of the angular position of the ro-tor prior to the removal (clearing) of the faultsuch that the rotor will obtain synchronousspeed without slipping poles following theremoval (clearing) of the fault. The corre-sponding time for the rotor to achieve thisangle is specified as the critical clearing time.

(2) the largest allowable angular deviationfrom synchronism that may be borne by apower system such that the system remainsstable: the edge of instability.

critical damping the least amount ofdamping such that the system does not freelyoscillate. For a characteristic equation of theform:

s2 + 2ζωns + ω2n,

the system is critically damped ifζ = 1.0;the roots of the characteristic equation arerepeated and real.

critical dimension (CD) the size (width)of a feature printed in resist, measured at aspecific height above the substrate.

critical frequency the rate of picture pre-sentation, as in a video system or motion pic-

ture display, above which the presented im-age ceases to give the appearance of flicker-ing. The critical frequency changes as a func-tion of luminance, being higher for higherluminance.

critical path a signal path from a primaryinput pin to a primary output pin with thelongest delay time in a logic block.

critical point Seeequilibrium point.

critical race a change in two input vari-ables that results in an unpredictable outputvalue for a bistable device.

critical region a set of instructions fora process that access data shared with otherprocesses. Only one process may execute thecritical region at a time.

critical section Seecritical region.

critically sampled sampling that at theNyquist frequency.

Crosby direct FM transmitter after itsinventor, Murray Crosby. Also known asthe “serrasoid modulator.” Direct frequencymodulation (FM) of an inductor/capacitor(LC) oscillator is essentially straightforward:One of the frequency determining elementsvalue is varied in accordance with the base-band information.

cross chrominance NTSC video arti-fact that causes luminance information to bepresent in the decoded chroma signal (lu-minance crosses into chrominance). Crosschrominance is a result of mixing highfrequency luminance information with thechrominance information in the compositevideo signal. An example of cross lumi-nance is the rainbow pattern observed whentweed or a herringbone pattern appears in aTV scene.

cross color Seecross chrominance.


cross luma Seecross luminance.

cross luminance NTSC video artifactthat causes chrominance information presentin the decoded luma signal (chrominancecrossed into luminance). Cross luminancehas the appearance similar to a zipper causedby the color subcarrier. Television receiversthat use a line comb filter to separate luma andchroma signals have cross-luminance com-ponents appearing on sharp horizontal edges.Television receivers that use a band-pass fil-ter for luma–chroma separation have cross-chrominance components appearing on sharpvertical edges.

cross modulation an undesired intermod-ulation of an electromagnetic carrier wave byanother electromagnetic carrier wave that iseither physically adjacent to it or near it interms of its radio frequency.

cross polar discrimination the ratioof co-polar to cross-polar received fieldstrengths in a depolarizing medium. Usuallyexpressed in decibels.

cross polar isolation the ratio between thefield received by an antenna with one polar-ization from co-polar and cross-polar trans-missions. This expresses the receiver’s abil-ity to detect one signal when dual-polarizedsignals are transmitted through a depolariz-ing medium.

cross power spectrum for two jointlywide sense stationary random processes theFourier transform of their cross-correlation.

cross spectra computation of the energyin the frequency distribution of two differentelectrical signals.

cross-assembler a computer program thattranslates assembly language into machinecode for a target machine different from theone on which the cross-assembler runs.

cross-compiler a computer program thattranslates a source code into machine or as-sembly code for a target machine differentfrom the one on which the cross-compilerruns.

cross-correlation a measure of the cor-relation or similarity of two signals. Forrandom processesx(t) andy(t), the cross-correlation is given by: Rxy(t1, t2) =Ex(t1)y(t2). See wide sense stationaryuncorrelated scattering channel.

cross-correlation function a function de-scribing the degree of similarity between twosignals, as a function of time-shift betweenthe signals.See alsocorrelation function.

cross-field theory a conceptual way toenvision the operation of a single-phase in-duction motor. The rotor current is assumedto produce a magnetic field electrically andspatially orthogonal to the field produced bythe main stator winding, thus contributing toa rotating magnetic field.

cross-polarization the field componentorthogonal to the desired radiated field com-ponent (co-polar component).

cross-quad the simplest form of the com-mon centroid concept, in which two matchedtransistors are formed by connecting diago-nally opposite devices in a two-by-two arrayof transistors.

crossarm a transverse, generally insu-lated member mounted horizontally on a util-ity pole. It carries insulators and allows widespacing of overhead conductors.

crossarm brace a brace, often insulated,which keeps a crossarm from rotating on itsattachment bolts.

crossbar switch a structure that allowsN units to communicate directly with eachother, point to point. Which pairs are con-nected depends on how the switch is config-


ured at that point in time. Crossbars are usu-ally implemented for small (8 or less) num-bers of nodes, but not always.

crossed field devices (CFD) radial or lin-ear forward traveling wave amplifier or back-ward wave oscillator where radial or trans-verse DC accelerating electric fields are per-pendicular to axial or longitudinal DC mag-netic fields, respectively.

crossing number the even number ob-tained by adding the number of changes inbinary value on going once around a particu-lar pixel location; crossing number is usefulfor finding skeletons of shapes and for help-ing with their analysis: it is a measure of thenumber of ‘spokes’ of an object emanatingfrom a specified location.

crossmodulation modulation of a desiredcarrier by an undesired interfering signal dueto interaction of the two signals in one ormore nonlinear elements.

crossover frequency the frequency wherethe magnitude of the open-loop gain is 1.

crossover point a crossover point of afuzzy setA is the point in the universe ofdiscourse whose membership inA is 0.5.

crosspoint a point at which two overlapneighboring fuzzy sets have the same mem-bership grade.

crosspoint level the membership grade ofa fuzzy set at a crosspoint.

crosstalk (1) undesired coupling (resis-tive, capacitive, or inductive) from one sig-nal wire or transmission line to a colocatedsignal wire or transmission line.

(2) capacitative interference between twoparallel transmission lines, in which a signalon one line affects a signal on the other line.

crowbar a triggered, shunt device that di-verts the stored energy of the beam powersupply.

CRR Seecochannel reuse ratio.

CRT Seecathode ray tube.

crystal filter reference amplifier ageneric phrase that refers to a cascade of am-plifying stages coupled together with the aidof crystal filter networks. Various topolo-gies are available for the connection of thesecrystals (lattice, bridged, hybrid, and ladder).The outstanding feature of such an arrange-ment is that the amplitude versus frequencyand phase versus frequency characteristicsare essentially determined by the crystal ar-rangements.

crystal oscillator an electronic circuitused for generating a sinusoidal waveformwhose frequency is determined by a piezo-electric crystal device.

crystalline phase in crystalline materialsthe constituent atoms are arranged in regu-lar geometric ways; for instance in the cubicphase the atoms occupy the corners of a cube(edge dimensions≈ 2–15A for typical ox-ides).

CSI Seecurrent-source inverter.

CSIM Seecapacitor-start induction motor.

CSMA Seecarrier-sense multiple access.

CSMA-CA CSMA with collision avoid-ance, a modification of the CSMA multipleaccess protocol to make it suitable for use in aradio environment. Prior to the transmissionof a message, a source and destination nodeundertake an exchange of short messages toensure that the destination node is idle andcapable of receiving a transmission from thesource node.See alsocarrier-sense multipleaccess.


CSMA/CD CSMA with collision detec-tion. A modification of the CSMA protocol,where a user continues to monitor (sense) thechannel after it initiates its transmission todetermine if other terminals have also initi-ated transmissions. The terminal aborts thetransmission if it detects a collision. Thisprotocol is used in the local area networkspecified in the standard IEEE 802.3, whichdefines a local area computer network com-monly referred to as the Ethernet.See alsocarrier-sense multiple access.

CSO Seecomposite second order.

CSR Seecontrolandstatus register.

CT Seecurrent transformer.

CT2 (cordless telephone, 2nd generation).A digital cordless communication system de-veloped in the United Kingdom, which oper-ates in the 864-868 MHz band and can pro-vide both private and public cordless tele-phony services. CT2-based public telephonyservices are often described as Telepoint ser-vices (q.v.). CT2 systems have 40 radio chan-nels, each 100 kHz wide. Each radio channelprovides a duplex traffic channel using sim-ple “ping-pong” TDD. The carrier bit rateis 72 Kb/s, and CT2 terminals have a peaktransmit power of 10 mW.

CTB Seecomposite triple beat.

CTE Seecoefficient of thermal expansionmismatch.

CU Seecoefficient of utilization.

cubic voxel Seevoxel.

cubical quad antenna a parasitic arrayantenna in which the elements are made ofsquare loops of wire with a perimeter of onewavelength. These antennas are typicallyused in the HF or VHF range and usuallyhave only two or three elements, resulting ina structure that fills a roughly cubical volume.

Cuk converter named after its inventor,SlobodanCuk, this device can be viewed as aboost converter followed by a buck converterwith topologic simplification. A capacitor isused to transfer energy from the input to theoutput. The output voltagevo is related tothe input voltagevi by vo = vid/(1−d) andit can be controlled by varying the duty ratiod. Note that the output voltage is oppositepolarity to the input.

Cuk converter.

cumulative distribution function (CDF)for a random variablex, the probability thatxis less than or equal to some valuex, denotedFx(x). Fx(−∞) is zero and increases mono-tonically toFx(+∞) = 1. For a continuousprobability density functionp(x), the CDFis

Fx(x) =∫ x

−∞p(t)dt.

The CDF is used in image processing tocarry out histogram equalization.See alsohistogram equalization, probability densityfunction.

cumulatively compounded a compound-wound DC machine in which the flux pro-duced by the MMF of the shunt field wind-ing and the flux produced by the MMF of theseries field winding are in the same direction.

Curie temperature the temperatureabove which a material ceases to be mag-netic; at this temperature a ferromagnetic ma-terial becomes paramagnetic.

current the flow of charge, measured inamperes (1 ampere= 1 coulomb/s).


current density vector field the field(commonly denotedJ ) that is related to theelectric field intensity vector field by the con-ductivity of the medium that the fields are lo-cated in. One of the quantities found on theright side of Ampere’s Law. The units are(amperes/square meter).

current distribution factor in economicdispatch studies, the proportion of a powerline’s total current which is contributed by aparticular generating plant.

current fed inverter an inverter in whichthe current is switched instead of the voltageto create AC current from a DC source. Alarge inductor is used to maintain a nearlyconstant current on the DC side which is thendirected to the load in an alternating fashionby the use of switching elements.

current feedback op-amp an op-amp inwhich the output voltage is controlled by thecurrent in the input stage, rather than thevoltage. The advantage of the current feed-back op-amp is that (to first order) its closed-loop performance is not subject to the gain-bandwidth tradeoff that affects voltage feed-back op-amps.

current gain short-circuit current gainthat helps describe the physical operation ofthe transistor. The current gain is the di-mensionless ratio of the peak RF output cur-rent (Iout.pk) to the peak incident RF current(Iincident.pk). Hence, current gain,GI is

GI = Iout.pk

Iincident.pk

current limiting the output current is lim-ited to a preset level even under a shortedoutput condition. This can be accomplishedby reducing the output voltage to prevent thecurrent limit from being exceeded.

current limiting fuse a fuse that limitsthe level of fault current from that which isavailable. It operates by developing a sub-

stantial voltage across the fuse following themelting of the fuse element.

current mirror a configuration of twomatched transistors in which the output is acurrent that is ideally equal to the input cur-rent. In the case of a BJT current mirror,the collector of the first transistor is forcedto carry the input current. This establishesa corresponding base-emitter voltage, whichis applied to the second transistor. If the twodevices are matched, then the collector cur-rent of the second transistor will equal that ofthe first transistor, thus “mirroring” the inputcurrent. This is a commonly used configura-tion in integrated circuits, which can take ad-vantage of the inherent matching available byfabricating the two transistors in close prox-imity to each other.

current regulator a device used to con-trol the magnitude and phase of the currentin DC, AC or other electrical variable speeddrives. May use different control strategieslike hysteresis current control or ramp com-parison current control.

current source amplifier the most com-mon group of amplifiers is made up of cur-rent sourcing amplifiers, in which the activedevice acts as a modulated current source.All class A, B, A-B, and C amplifiers fit intothis general group. Parameters such as devicecharacteristics, quiescent bias point, RF loadline, and amplitude and waveform of the ap-plied signal should be included with the classdefinition, thus defining the major contribu-tors to the physical actions taking place inone of these amplifiers.

current source inverter Seecurrent fedinverter.

current source region region of the I-V curve(s) of a device in which the outputcurrent is relatively constant (slope near zero)for changing output voltage.


current transducer a device used to mea-sure current in a variety of applications in-cluding variable speed drives. May give out acurrent proportional to the measured currentor a voltage proportional to the measured cur-rent. Electrical isolation may be obtained byusing a current transformer (cannot be usedat DC) or a hall effect transducer that can beused down to DC.

current transformer (CT) (1) a trans-former that is employed to provide a sec-ondary current proportional to primary cur-rent flowing. The primary “winding” is of-ten created by passing the system conductoror bus bar through an opening in the deviceand the secondary is typically rated at a stan-dard value to match common meters and dis-play units. Current transformers are used incurrent measurement, protective relays, andpower metering applications. The load (me-ter) on a CT should never be removed withoutfirst shorting the secondary of the CT, oth-erwise dangerous voltage levels may resultwhen the load is removed.

(2) a device which measures the instanta-neous current through a conductor of an elec-tric power line and transmits a signal propor-tional to this current to the power system’sinstrumentation.

current unit a protective relay that mon-itors the magnitude of a power transmissionline’s current flow.

current withstand rating the currentwithstand rating of a device is the maximumshort term current that can flow in the devicewithout causing damage.Seeampacity.

current-controlled multivibrator a mul-tivibrator with oscillation period controlledby the current that is charging and discharg-ing a timing capacitor. Most of the circuitsare based on two symmetric controlled cur-rent sources connected by a timing capaci-tor, and a steering circuit (connected betweenpower supply voltage and this capacitor),which includes two switches. The switches

take turns providing the conducting path forthe current sources, and the current in thetiming capacitor changes direction each halfof the period. If the current sources arecontrolled by an external voltage, the circuitbecomes a voltage-controlled multivibrator.Current- and voltage controlled multivibra-tors are a vital part of phase-locked loops andfrequency synthesizers.

current-limiting device when operatingwithin its current-limiting range, a current-limiting device is open in 0.25 cycle or lessand limits the maximum short-circuit cur-rent to a magnitude substantially less thanthe short-circuit current available at the faultpoint.

current-mode control the inductor cur-rent is sensed to control the pulse width.The switch is turned on by a constant fre-quency clock and turned off when the in-ductor current reaches the control reference.This method produces subharmonic oscilla-tions if the steady-state duty ratio is greaterthan or equal to 0.5. An artificial stabilizingramp is necessary to achieve stable controlover the full range of duty ratios. When theslope of the ramp is equal to the falling slopeof the inductor current, one cycle response isachieved. (i.e., the input perturbations are re-jected, and the current reference is followedwithin one cycle).

current-source inverter (CSI) an in-verter with a DC current input. Current-source inverters are most commonly used invery high power AC motor drives.

cursor (1) the symbol on a computerscreen that indicates the location on thescreen that subsequent input will affect.

(2) a movable, visible mark used to indi-cate a position of interest on a display surface.(ANSI).

curvature a geometric property that de-scribes the degree that a surface or a curve isbent. The curvature of a curve is the magni-


tude of the rate of change of the unit tangentvector with respect to arc length. The curva-ture of a surface is given in terms of a metrictensor which embodies two principal (pla-nar) curvatures,κ1 andκ2. The curvature ofa surface is sometimes characterized by theso-called Gaussian curvatureκ = κ1κ2.

curvature function a function that givescurvature values at different locations of acurve or surface.

cut-off frequency the minimum fre-quency at which a waveguide mode will prop-agate energy with little or no attenuation.

cut-off rate a measure of reliability of achannel. At information rates below the cut-off rate, the probability of error averaged overthe ensemble of randomly selected codes canbe forced to zero exponentially with the blocklength (for block codes), or constraint length(trellis codes). The cut-off rate is less than orequal to the channel capacity. Also knownas the computational cut-off rate, since fortrellis codes, it is at this rate that the ex-pected number of computations required toperform sequential decoding becomes un-bounded.

cut-off wavelength the wavelength abovewhich a fiber exhibits single mode operation(and below which exhibits multimode oper-ation).

cutoff frequency Seecut-off frequency.

cutoff rate Seecut-off rate.

cut-out a pole-mounted device whichcontains a high-voltage fuse integral witha disconnect switch. It is used to protectand disconnect primary circuits in distribu-tion work.

cutset a minimal subnetwork, the removalof which cuts the original network into twoconnected pieces.

CVD Seechemical vapor deposition.

CVQ Seeclassified vector quantization.

CW Seecontinuous wave.

cybernetics (1) the field of control andcommunication theory in general, withoutspecific restriction to any area of applicationor investigation.

(2) the behavior and design of mecha-nisms, organisms, and/or organizations thatreceive and generate information and re-spond to it in order to obtain a desired result.

cycle Seeclock cycle.

cycle stealing an arrangement in whicha DMA controller or I/O channel, in orderto use the I/O bus, causes the CPU to tem-porarily suspend its use of the bus. The CPUis said to hesitate.See alsodirect memoryaccess.

cycle time time required to complete oneclock cycle (usually measured in nanosec-onds).

cycles per instruction (CPI) a perfor-mance measurement used to judge the effi-ciency of a particular design.

cyclic access in devices such as magneticand optical disks (and older bubble memo-ries) that store data on rotating media, theproperty that any individual piece of data canbe accessed once during each cycle of the me-dia.

cyclic code a linear block code where ev-ery cyclic shift of a codeword is another code-word in the same code. This property is anoutcome of the significant algebraic structurethat underlies these codes.

cyclic redundancy check (CRC) (1)is used to detect transmission errors overcommunication channels in digital systems.When using the CRC technique, the modulo-


2 sum of the message bits is calculated aftergrouping them in a special way and then ap-pended to the transmitted message and usedby the receiver to determine if the messagewas received correctly. The number of bitsused in the CRC-sum is typically 8 or 16, de-pending on the length of the message and thedesired error-detection capability.

A value calculated from a block of data tobe stored or transmitted with the data blockas a check item. CRCs are generated using ashift register with feedback and are describedby the length of the register and the feedbackterms used.

(2) a coding scheme for detecting errorsin messages. Given a polynomialC(x) oforderc−1, and anm-bit messageM(x) rep-resented as a polynomial of orderm − 1,calculateR(x) = (xcM(x))/C(x), thenconstruct the transmitted messageT (x) =(xcM(x))⊕ R(x) which is exactly divisiblebyC(x). SupposingE(x) represents any er-rors introduced during transmission, the re-ceiver calculates(T (x) + E(x))/C(x). Ifthis is not zero,E(x) is non-zero and the re-ceived message is erroneous. If this is zero,either no errors were introduced orE(x) issuch that the CRC is not strong enough to de-tect the error. A suitable choice ofC(x) willmake this unlikely.

cycloconverter a frequency changingconverter that synthesizes lower frequencysine waves from portions of a polyphase set ofhigher frequency sine waves. For example,a three-phase, 400 Hz input could be used tocreate a 60 Hz supply of any desired phaseorder.

cyclotron frequency a frequency of elec-tron oscillation in a cyclotron. A frequencyof circular motion of an electron under mag-netic fields applied perpendicularly to theplane of the circular motion.

cylinder a stack of tracks where thesetracks are at a constant radial position (thesame track number) on a disk or disk pack.

cylindrical extension of a fuzzy set letAbe a fuzzy set in a Cartesian product spaceXi ,andXn be another Cartesian product spaceincludingXi , then the cylindrical extensionof A in Xn is a fuzzy set inXn, denoted ascext (A;Xn), with membership function de-fined equal to membership inXi . See alsofuzzy set, membership function, projectionof a fuzzy set.

cylindrical lens a lens that has curvature,or optical focusing power, in one directiononly.

cylindrical microstrip line structure thatconsists of an angular strip on a cylindricalground plane separated by a dielectric sub-strate. The cylindrical conductor acts as aground plane, the angular strip guides theelectromagnetic wave along the longitudinaldirection.

cylindrical stripline the angular strip inthe homogeneous media between the twocylindrical conductors. The inner and outerconductors act as ground planes, the angularstrip guides the electromagnetic wave alongthe longitudinal direction.

cylindrical wave an electromagneticwave in which each wavefront (surface ofconstant phase) forms a cylinder of infinitelength and propagates in toward or away fromthe axis of the cylinder. A uniform cylin-drical wave has the same amplitude over anentire wavefront; a nonuniform cylindricalwave has varying amplitude.

cylndrical-rotor machine a synchronousmachine with a cylindrical rotor containinga distributed field winding and an essentiallyuniform air-gap. This design is limited to twoand four pole machines (3600 and 1800 rpmat 60 Hz) and is usually used in large gener-ators.See alsosalient-pole rotor machine.


Dd axis Seedirect axis.

D flip-flop a basic sequential logic cir-cuit, also known as bistable, whose outputassumes the value (0 or 1) at its D input whenthe device is clocked. Hence it can be used asa single bit memory device, or a unit delay.

D’Arsenval meter a permanent-magnetmoving-coil instrument with a horseshoeform magnet. It measures direct current only.

D-TDMA See dynamic time divisionmultiple access.

D/A Seedigital-to-analog converter.

DAC Seedigital-to-analog converter.

Dahlin’s method a synthesis procedurefor finding a digital control law that will pro-duce a defined closed loop behavior for agiven process, when it is controlled in a feed-back configuration. The synthesis equationis

k(z) =(1 − e−T/λ

)z−N−1

1 − e−T/λz−1 − (1 − e−T/λ

)z−N−1

× 1

g(z)

wherek(z) is the pulse transfer function forthe synthesized controller,T is the samplingtime of the digital control loop,λ is the timeconstant specified for the closed loop system,NT is the process deadtime, andg(z) is thepulse transfer function for the process.

daisy chain (1) a type of connection whendevices are connected in series.

(2) a hardware configuration where a sig-nal passes through several devices. A signal

will be passed through a device if that de-vice is not requesting service, or not passedthrough if the device is requesting service.

(3) an interrupt-prioritizing scheme inwhich the interrupt acknowledge signal fromthe CPU is connected in series through alldevices. A shared interrupt-request line con-nects all devices to the CPU with a singlecommon line. When one (or several) devicesactivates its request line, the CPU will (af-ter some delay) respond with an acknowl-edge to the first device. If this device did re-quest an interrupt, it will be serviced by theCPU. However, if the device did not requestan interrupt, the acknowledge is just passedthrough to the next device in the daisy chain.This process is repeated until the acknowl-edge signal has passed through all the con-nected devices on this chain. The schemeimplements prioritized service of interruptsby the way the devices are electrically con-nected in the daisy chain: the closer a deviceis to the CPU, the higher its priority.

A more general case exists where severaldaisy chains are used to form priority groups,where each chain has a unique priority. TheCPU will service interrupts starting with thedaisy chain having the highest priority. Inthis scheme, any device may be connected tomore than one priority group (chain), usingthe interrupt priority level appropriate for theparticular service needed at this moment.

DAMA protocol See demand assignmultiple access protocol.

damage the failure pattern of an electronicor mechanical product.

damped sinusoid a sinusoidal signal mul-tiplied by a decaying. An example isx(t) =10e−3t cos(10t).

damper winding an uninsulated wind-ing, embedded in the pole shoes of a syn-chronous machine, that includes several cop-per bars short-circuited by conducting ringsat the ends, used to reduce speed fluctuation


in the machine by developing an induction-type torque that opposes any change in speed.

damping a characteristic built into electri-cal circuits and mechanical systems that pre-vents rapid or excessive corrections that maylead to instability or oscillatory conditions.

damping coefficient electrical torquecomponent in phase with the rotor speed.

damping factor a measure of the abilityof the PLL to track an input signal step. Usu-ally used to indicate the amount of overshootpresent in the output to a step perturbation inthe input.

damping ratio the ratio of the real part ofthe resonant frequency to the undamped res-onant frequency, in a second-order system.

dark current a noise source in photode-tectors, corresponding to undesired outputsignals in the absence of light. The mainsource of dark currents are ohmic leakagedue to imperfect insulation and thermionicemission.

darkfield performance the amount oflight that is blocked in a light valve projectorwhen a totally dark scene is being displayed.This parameter is critical to good contrast ra-tio.

Darlington bipolar junction transistora combination of two bipolar junction tran-sistors (BJT) where the emitter current ofone transistor drives the base of the secondtransistor. The arrangement reduces the cur-rent required from the base driver circuit, andthe effective current gain of the combinationis approximately the product of individualgains. The configuration can be made fromtwo discrete transistors or can be obtained asa single integrated device.

dart leader in lightning, a continuouslymoving leader lowering charge preceding areturn stroke subsequent to the first. A dart

leader typically propagates down the residualchannel of the previous stroke.

dashpot timer a fluid time-on timer thatcan be used in definite time DC motor accel-eration starters and controllers. The dashpottimer functions where a magnetic field forcesa piston to move within a cylinder when thecoil is energized. The movement of the pistonis limited by fluid passing through an orificeon the piston. The amount of fluid passingthrough the orifice is controlled by a throttlevalue, which in turn determines the amountof time delay. After the fluid pressure equal-izes across the piston, movement stops andcontacts change state. The fluid can eitherbe air (pneumatic dashpot) or oil (hydraulicdashpot). When the timer is deenergized, acheck valve allows the pressure to equalizeacross the piston rapidly, thereby causing thecontacts to change state “immediately.”

data any information, represented in bi-nary, that a computer receives, processes, oroutputs.

data access fault a fault, signaled in theprocessor, related to an abnormal conditiondetected during data operand fetch or store.

data acquisition (1) method used for cap-turing information and converting it to a formsuitable for computer use.

(2) process of measuring real-world quan-tities and bringing the information into acomputer for storage or processing.

data bottleneck a computer calculationin which the speed of calculation is limitedby the rate at which data is presented to theprocessor rather than by the intrinsic speed ofthe processor itself. Ultra high speed parallelprocessors are very frequently limited in thisway.

data buffer Seebuffer.


data bus set of wires or tracks on a printedcircuit or integrated circuit that carry binarydata, normally one byte at a time.

data cache a small, fast memory thatholds data operands (not instructions) thatmay be reused by the processor. Typical datacache sizes currently range from 8 kilobytesto 8 megabytes.Seecache.

data communications equipment (DCE)a device (such as a modem) that establishes,maintains, and terminates a session on a net-work.

data compression theorem Claude Shan-non’s theorem, presenting a bound to the op-timally achievable compression in (lossless)source coding.See alsoShannon’s sourcecoding theorem.

data dependency the normal situation inwhich the data that an instruction uses or pro-duces depends upon the data used or pro-duced by other instructions such that the in-structions must be executed in a specific orderto obtain the desired results.

data detection in communications, amethod to extract the transmitted bits fromthe received signal.

data flow architecture a computer ar-chitecture that operates by having sourceoperands trigger the issue and execution ofeach operation, without relying on the tra-ditional, sequential von Neumann style offetching and issuing instructions.

data fusion analysis of data from mul-tiple sources — a process for which neuralnetworks are particularly suited.

data logger a special-purpose processorthat gathers and stores information for latertransfer to another machine for further pro-cessing.

data path the internal bus via which theprocessor ships data, for example, from thefunctional units to the register file, and viceversa.

data pipeline a mechanism for feeding astream of data to a processing unit. Data ispipelined so that the unit processing the datadoes not have to wait for the individual dataelements.

data preprocessing the processing of databefore it is employed in network training.The usual aim is to reduce the dimensionalityof the data by feature extraction.

data processing inequality informationtheoretic inequality, a consequence of whichis that no amount of signal processing on asignal can increase the amount of informationobtained from that signal. Formally stated,for a Markov chainX → Y → Z,

I (X;Z) ≤ I (X;Y )The condition for equality is thatI (X;Y |Z) =0, i.e.,X → Z → Y is a Markov chain.

data reduction coding system any al-gorithm or process that reduces the amountof digital information required to represent adigital signal.

data register a CPU register that may beused as an accumulator or a buffer register oras index registers in some processors. In pro-cessors of the Motorola M68000 family, dataregisters are separate from address registersin the CPU.

data segment the portion of a process’virtual address space allocated to storing andaccessing the program data (BSS and heap,and may include the stack, depending on thedefinition).

data stripe storage methodology wheredata is spread over several disks in a diskarray. This is done in order to increase thethroughput in disk accesses. However, la-


tency is not necessarily improved.See alsodisk array.

data structure a particular way of orga-nizing a group of data, usually optimized forefficient storage, fast search, fast retrieval,and/or fast modification.

data tablet a device consisting of a sur-face, usually flat, and incorporating meansfor selecting a specific location on the sur-face of the device and transmitting the coor-dinates of this location to a computer or otherdata processing unit that can use this infor-mation for moving a cursor on the screen ofthe display unit.

data terminal equipment (DTE) a de-vice, such as a subscriber’s computer, Ex-change workstation, or Exchange central sys-tem, that controls data flow to and from acomputing system. It serves as a data source,data sink, or both, and provides for the datacommunication control function according toprotocols. Each DTE has an address that isa 12-digit number uniquely identifying thesubscriber’s connection to the network. DTEterm is usually used when referring to the RS-232C serial communications standard, whichdefines the two end devices of the communi-cations channel: the DTE and the DCE (datacommunications equipment). The DCE isusually a modem and the DTE is a UARTchip of the computer.

data-oriented methodology an applica-tion development methodology that consid-ers data the focus of activities because dataare more stable than processes.

database computer a special hardwareand software configuration aimed primarilyat handling large databases and answeringcomplex queries.

dataflow computer a form of computerin which instructions are executed when theoperands that the instructions require becomeavailable rather than being selected in se-

quence by a program counter as in a tradi-tional von Neumann computer. More thanone processor is present to execute the in-structions simultaneously when possible.

Daubechies wavelets a class of com-pactly supported orthogonal and biorthog-onal wavelets, first proposed by IngridDaubechies, that can be obtained by impos-ing sufficient conditions on wavelet filters.

daughter board a computer board thatprovides auxiliary functions, but is not themain board (motherboard) of a computer sys-tem (and is usually attached to the mother-board).

dB Seedecibel.

dBc ratio of the signal power (p) to a ref-erence signal power (pref), usually the mod-ulation carrier signal, expressed in decibelsreferenced to a carrier (dBc). Thus a har-monic signal that is 1/100th of the power ina desired fundamental signal is at−20 dBc.

PdBd = 10 log10

(p

pref

)

dBm Power ratio in decibels referencedto 1 milliwatt.

DBS Seedirect broadcast satellite.

DBS receiver electronic assembly thataccepts as input a microwave signal from asatellite, containing transmitted TV signals“modulated” onto the signal. The receiverfirst amplifies the low-level signal, then pro-cesses the signal by first converting it to alower “IF” frequency and then demodulatingthe signal to separate the TV signals fromthe microwave carrier signal. A basic way oflooking at the relationship of the microwavecarrier to the TV signal is to think of the car-rier signal as an envelope with a messageletter inside. The message letter is the TVsignal. Demodulation is the process of care-fully removing the message from the enve-


lope (carrier). The noise figure of receiveris a measure of the amount of noise that willbe added to the signal (carrier and TV signal)by the receiver. If the receiver adds too muchnoise, the result will be a snowy picture onthe TV screen.

dBW ratio of the signal power in watts(pwatts) to a 1 Wreference power, expressedin decibels referenced to 1 W (dBW). Thus1 watt signal power is equal to 0 dBW, and−30 dBW is equal to 0 dBm.

PdBW = 10 log10

( pwatts

1 ·Watt)

DC direct current. SeeDC current, DCvoltage.

DC block A circuit simulation componentthat behaves like a capacitor of infinite value.

DC chopper a DC to DC converter thatreduces the voltage level by delivering pulsesof constant voltage to the load. The averageoutput is equal to the input times the dutycycle of the switching element.

DC circuit electrical networks in whichthe voltage polarity and directions of currentflow remain fixed. Thus such networks con-tain direct currents as opposed to alternatingcurrents, thereby giving rise to the term.

DC current constant current with no vari-ation over time. This can be considered ingeneral terms as an alternating current (AC)with a frequency of variation of zero, or azero frequency signal. For microwave sys-tems, DC currents are provided by batteriesor AC/DC converters required to “bias” tran-sistors to a region of operation where theywill either amplify, mix or frequency trans-late, or generate (oscillators) microwave en-ergy.

DC drain conductance for an FET deviceunder DC bias, the slope of the output drainto source current (IDS) versus output drain

to source voltage (VDS) for a fixed gate tosource voltage (VGS), expressed in siemens.

gd = ∂IDS

∂VDS

∣∣∣∣VGS = constant

DC generator commutator exciter asource of energy for the field winding of asynchronous machine derived from a directcurrent generator. The direct current gen-erator may be driven by an external motor,a prime mover, or by the shaft of the syn-chronous machine.

DC input power the total DC or biaspower dissipated in a circuit, which is usuallydependent on signal amplitudes, expressedin watts. This may include input bias, biasfiltering, regulators, control circuits, switch-ing power supplies and any other circuitryrequired by the actual circuit. These con-siderations should be explicitly specified, asthey will affect how efficiency calculationsare performed.

DC link the coupling between the rectifierand inverter in a variable speed AC drive.

DC link capacitor a device used on theoutput of a rectifier to create an approxi-mately constant DC voltage for the input tothe inverter of a variable speed AC drive.

DC link inductor an inductor used on theoutput of a controlled rectifier in AC currentsource drives to provide filtering of the inputcurrent to the current source inverter. If usedin conjunction with a capacitor, then it is usedas a filter in voltage source drives.

DC load flow a fast method of estimatingpower flows in an electric power system inwhich the problem is reduced to a DC circuit,with line impedances modeled as resistancesand all generator bus voltages presumed toremain at their nominal values.

DC machine an electromechanical (rotat-ing) machine in which the field winding is on


the stator and carries DC current, and the ar-mature winding is on the rotor. The currentand voltage in the armature winding is actu-ally AC, but it is rectified by the commutatorand brushes.

DC motor a motor that operates from aDC power supply. Most DC motors have afield winding on the stator of the machinethat creates a DC magnetic field in the airgap.The armature winding is located on the rotorof the machine and the DC supply is invertedby the commutator and brushes to provide analternating current in the armature windings.

DC motor drive a converter designed tocontrol the speed of DC motors. Controlledrectifiers are generally used and provide avariable DC voltage from a fixed AC volt-age. Alternatively, a chopper, or DC–DCconverter, can be employed to provide a vari-able DC voltage from a fixed DC voltage.

DC offset current the exponentially de-caying current component that flows imme-diately following a fault inception. DC off-set is the result of circuit inductance, and isa function of the point in the voltage wavewhere the fault begins. The offset for a givenfault can range from no offset to fully offset(where the instantaneous current peak equalsthe full peak–peak value of the AC current).

DC restoration reinsertion of lost DClevel information into a signal after using ACsignal coupling; in television applications,the DC component of a composite video sig-nal represents the average brightness of thepicture. After AC coupling of the compos-ite video, the DC level includes the averageluminance signal plus the fixed average ofthe sync and lanking signals, causing picturelevel racking errors. For a positive video sig-nal, the average value of mostly white sceneswill be slightly lower than it should be; formostly dark scenes, the DC average could be-come negative (due to the sync and blankingsignals) when it should be slightly positive.Clamping circuits restore the DC average of

the video by establishing the DC level of thesync tips.

DC servo drive a feedback, speed controldrive system used for position control. Ser-vos are used for applications such as roboticactuators, disk drives, and machine tools.

DC test tests that measure a static param-eter, for example, leakage current.

DC to RF conversion efficiency dimen-sionless ratio of RF power delivered to theload (pout) versus total DC input power dis-sipated in the amplifier (pDC). With the DCto RF conversion efficiency given myηDCwe have

ηDC = pout

pDC

DC transconductance for an FET de-vice under DC bias, the slope of the out-put drain to source current(IDS) versus in-put gate to source voltage (VGS) for a fixeddrain to source voltage (VDS), expressed insiemens. Given bygm we have

gm = ∂IDS

∂VGS

∣∣∣∣VDS = constant

DC voltage constant voltage with no vari-ation over time. This can be considered ingeneral terms as an alternating current (AC)with a frequency of variation of zero, or azero frequency signal. For microwave sys-tems, DC voltages are provided by batteriesor AC/DC converters required to “bias” tran-sistors to a region of operation where theywill either amplify, mix or frequency trans-late, or generate (oscillators) microwave en-ergy.

DC–AC inverter Seeinverter.

DC–DC converter a switching circuitthat converts direct current (DC) of one volt-age level to direct current (DC) of anothervoltage level. A typical DC–DC converter


includes switches, a low pass filter (to at-tenuate the switching frequency ripple), anda load. The size of magnetic componentsand capacitors can be reduced and bandwidthcan be increased when operating at high fre-quency. Most DC–DC converters are pulse-width modulated (PWM), while resonant orquasi-resonant types are found in some appli-cations. Commonly used topologies includethe buck converter, boost converter, buck-boost converter, andCuk converter. Isola-tion can be achieved by insertion of a highfrequency transformer.

DC-free code Seebalanced code.

DCA Seedynamic channel assignment.

DCE Seedata communications equipmentor distributed computing environment.

DCS-1800 digital communication system–1800 MHz. A micro-cell version of GSMthat operates at a lower transmitter power,higher frequency band, and has a larger spec-trum allocation than GSM, but in most otherrespects is identical to GSM. The DCS-1800spectrum allocation is 1710–1785 MHz forthe up-link and 1805–1880 MHz for thedown-link (i.e., 2× 75 MHz). The peaktransmit power for portable DCS-1800 ter-minals is 1 W.

DCT Seediscrete cosine transform.

de Haas–Shubnikov oscillation SeeShubnikov–de Haas oscillation.

de-assert to return an enabling signal toits inactive state.

de-emphasis refers to the receiving pro-cess of correcting the amplitude of cer-tain signal components that have been “pre-emphasized” prior to transmission in order toimprove signal-to-noise (S/N) ratio. In com-mercial FM broadcast receivers, de-emphasisis accomplished with a simple resistor ca-pacitor lowpass filter that represents the in-

verse transfer function characteristic of thepre-emphasis network situated at the trans-mitter. See the figure for a de-emphasis net-work. Seepre-emphasis.

de-regulation the removal of some gov-ernment controls on public utilities, generallyincluding the unbundling of certain services,the dismantling of vertically-integrated util-ities, and the introduction of competitionamong various utility companies for cus-tomer services.

dead band (1) the portion of the operatingrange of a control device or transducer overwhich there is no change in output.

(2) referring to an automatic controller be-havior, a range of values of the controlledvariable in which no corrective action occurs.This type of controller behavior is responsi-ble for the time lag, called dead zone lag,which can cause instability of the controlledsystem if other conditions are present.

Also known as dead zone.

dead end an installation in which an elec-tric power line terminates at a pole or tower,typically for purposes of structural stabilty.

dead man (1) a stand on which to rest autility pole when setting the pole by hand.

(2) a buried log used as a guy anchor.

dead-end shoe a fixture for securing awire or strain insulator to a utility pole.

dead tank breaker a power circuitbreaker where the tank holding the interrupt-ing chamber is at ground potential. Oil cir-cuit breakers, for example, are typically deadtank breakers.

dead zone Seedead band.

deadbeat 2-D observer a system de-scribed by the equations

zi+1,j+1 = F1zi+1,j + F2zi,j+1 +G1ui+1,j

+G2ui,j+1 +H1yi+1,j +H2yi,j+1


xi,j = Lzi,j +Kyi,j

i, j ∈ Z+ (the set of nonnegative integers) iscalled a full-order deadbeat observer of thesecond generalized Fornasini–Marchesini 2-D model

Exi+1,j+1 = A1xi+1,j + A2xi,j+1

+ B1ui+1,j + B2ui,j+1

yi,j = Cxi,j +Dui,j

i, j ∈ Z+ if there exists finite integersM,Nsuch thatxi,j = xi,j for i > M, j > N , andanyuij , yij and boundary conditionsxi0 fori ∈ Z+ andx0j for j ∈ Z+ wherezij ∈ Rnis the local state vector of the observer at thepoint (i, j), uij ∈ Rm is the input,yi,j ∈Rp is the output, andxi,j ∈ Rn is the localsemistate vector of the model,F1, F2, G1,G2,H1,H2, L,K, E, A1, A2, B1, B2, C,Dare real matrices of appropriate dimensionswith E possibly singular or rectangular,Z+is the set of nonnegative integers. In a similarway, a full-order asymptotic observer can bedefined for other types of the generalized 2-Dmodels.

deadbeat control of 2-D linear systemsgiven the 2-D Roesser model[xhi+1,jxvi,j+1

]=[A1 A2A3 A4

][xhijxvij

]+[B1B2

]uij

i, j ∈ Z+ (the set of nonnegative integers)

yij = [C1 C2

] [xhijxvij

]

with boundary conditionsxh0j = 0 for j ≥N2 andxvi0 = 0 for i ≥ N1, find an inputvector sequence

uij ={ 6= 0 for 0 ≤ i ≤ N1,0 ≤ j ≤ N2

= 0 for i > N1 andj > N2

such that the output vectoryij = 0 for alli > N1 and j > N2 wherexhij ∈ Rn1 andxvij ∈ Rn2 are the horizontal and vertical statevectors, respectively, andA1, A2, A3, A4,C1, C2 are real matrices.

deadlock a condition when a set of pro-cesses using shared resources or commu-nicating with each other are permanentlyblocked.

deadtime the time that elapses betweenthe instant that a system input is perturbedand the time that its output starts to respondto that input.

debug to remove errors from hardware orsoftware.See alsobug.

debug port the facility to switch the pro-cessor from run mode into probe mode toaccess its debug and general registers.

debugger (1) a program that allows in-teractive analysis of a running program, byallowing the user to pause execution of therunning program and examine its variablesand path of execution at any point.

(2) program that aids in debugging.

debugging (1) locating and correcting er-rors in a circuit or a computer program.

(2) determining the exact nature and loca-tion of a program error, and fixing the error.

debuncher a radio frequency cavityphased so that particles at the leading edgeof a bunch of beam particles (higher momen-tum particles) are decelerated while the trail-ing particles are accelerated, thereby reduc-ing the range of momenta in the beam.

Debye material a dispersive dielec-tric medium characterized by a complex-valued frequency domain susceptibility func-tion with one or more real poles. Water is anexample of such a material.

Debye media SeeDebye material.

decade synonymous with power of ten.In context, a tenfold change in frequency.

decade bandwidth 10:1 bandwidth ratio(the high-end frequency is ten times the low-


end frequency). The band edges are usuallydefined as the highest and lowest frequencieswithin a contiguous band of interest at whichthe loss equalsLAmax, the maximum attenu-ation loss across the band.

decay a transformation in which an atom,nucleus, or subatomic particle changes intotwo or more objects whose total rest energyis less than the rest energy of the originalobject.

decay heat the fraction of the total en-ergy obtained from a nuclear fission reactionwhich is produced by delayed neutrons andby the secondary decay of fission daughters.

decay length the average distance aspecies of a particle at a given energy travelsbefore decaying.

decay time in the absence of any pump orother excitation mechanisms, the time afterwhich the number of atoms in a particularstate falls to 1/E of its initial value.

decentralized control a structure of large-scale control systems based on system de-composition onto interconnected subsystemsin order to simplify control design. Decen-tralized control systems are usually designedin the form of local feedback controllers andare chosen to fit information structure con-straints imposed by the decomposition.Toensure robustness with respect to intercon-nections between subsystems, the local con-trollers should be robust and/or coordinationshould ensure robustness of the overall sys-tem. See alsodecomposition.

decibel (dB) a unit of measure that de-scribes the ratio between two quantities interms of a base 10 logarithm. For example,the ratio between the power level at the in-put and output of an amplifier is called thepower gain and may be expressed in decibelsas follows:

G(dB) = 10 log10 (Pout/Pin)

Terms such as dBm, dBuV, dBW indicatethat the decibel measurement was made rel-ative to an established standard. A commonpower measure reference is 0 dBm, which isdefined to be 1 mW (milliwatt, 0.001W). Acommon voltage reference is 1µV (1 micro-volt).

decimal from the number system that hasbase 10 and employs 10 digits.

decimation an operation that removessamples with certain indexes from a discrete-time signal and then re-indexes the remainingsamples. Most frequently, decimation refersto keeping everynth sample of a signal. Alsoknow as down-sampling.

decision boundary a boundary in featurespace which separates regions with differentinterpretations or classes, e.g., the bound-ary separating two adjacent regions charac-terizing the handwritten characters ‘E’ and‘F’. In practice, the regions associated withneighboring classes overlap; consequentlymost decision boundaries lead to some erro-neous classifications, so an error criterion isused to select the “best” boundary.See alsoclassifier, Bayesian classifier.

decision directed the use of previouslydetected information bits in an estimator, de-tector, or adaption algorithm in an adaptivefilter. Usually improves performance com-pared to non-decision directed counterparts,but introduces potential problems with errorpropagation (erroneous bit feed back).

decision level the boundary betweenranges in a scalar quantizer. On one side ofthe decision level, input values are quantizedto one representative level; on the other side,input values are quantized to a different rep-resentative level.

decision mechanism rules and principlesby which the information available to a givendecision unit (control unit) is processed andtransformed into a decision; typical decision


mechanisms within control systems are fixeddecision rule and optimization-based mech-anism. Decision mechanisms can assume ahierarchical form — one may then talk abouta hierarchical decision structure.

decision support system a system whosepurpose is to seek to identity and solve prob-lems.

decision tree analysis decomposing aproblem into alternatives represented bybranches where nodes (branch intersections)represent a decision point or chance eventhaving probabilistic outcome. Analysis con-sists of calculating expected values associ-ated with the chain of events leading to thevarious outcomes.

decision-directed adaptation a methodfor adapting a digital equalizer in which de-cisions at the output of the equalizer are usedto guide the adaptive algorithm.See alsoadaptive algorithm, LMS.

declaration phase or statement in which anew variable is requested; the declaration ofa variable is based on the definition of its typeor class. This phase leads to a instantiationof the variable.

decode cycle the period of time duringwhich the processor examines an instructionto determine how the instruction should beprocessed. This is the part of the fetch-decode-execute cycle for all machine instruc-tion.

decode history table a form of branchhistory table accessed after the instructionis decoded and when the branch address isavailable so that the table need only store aBoolean bit for each branch instruction to in-dicate whether this address should be used.

decoder (1) a logic circuit withN inputsand 2N outputs, one and only one of whichis asserted to indicate the numerical value oftheN input lines read as a binary number.

(2) a device for converting coded infor-mation to a form that can be understood by aperson or machine.

decoder source the coded signal input tothe decoder. In information theory the de-coder source is modeled as a random process.

decoding (1) the act of retrieving the orig-inal information from an encrypted (coded)data transmission.

(2) the process of producing a single out-put signal from each input of a group of inputsignals.

(3) the operation of the decoder. The in-verse mapping from coded symbols to recon-structed data samples. Decoding is the in-verse of encoding, insofar as this is possible.

decomposition (1) an operation per-formed on a complex system whose purposeis to separate its constituent parts or subsys-tems in order to simplify the analysis or de-sign procedures.

1.For large-scale systems, decompositionis performed by neglecting links intercon-necting subsystems. It is followed by designof local control systems on the base of localobjectives and coordination, which enablesreaching global goals.See alsodecentralizedcontrol.

2.For optimization algorithms, decompo-sition is reached by resolving the objectivefunction or constraints into smaller parts, forexample, by partitioning the matrix of con-strains in linear programs followed by the so-lution of a number of low dimensional linearprograms and coordination by Lagrange mul-tipliers.

3.For uncertainties, decomposition is per-formed to make their model trackable, forexample, by dividing them into matchedand mismatched parts.See alsomatchingconditions.

4. For linear time-invariant systems instate form, Kalman’s decomposition is atransformation of state matrix in a way thatindicates its controllable and observable,controllable and unobservable, uncontrol-


lable and observable, and uncontrollable andunobservable parts.See alsocontrollability.

5. For matrix transfer functions, decom-position transforms it to the form composedof specific blocks used in chosen design pro-cedure.See alsoH infinity design.

(2) in fuzzy systems, if the fuzzy relationR is a composition of the fuzzy relationsAandB, then the determination ofA givenBandR is referred to as a decomposition ofRwith respect toB.

decorrelation the act of removing orreducing correlation between random vari-ables. For random vectors a discrete lineartransform is often used to reduce the correla-tion between the vector components.

decorrelator a linear code-division multi-ple-access receiver, which takes the data vec-tor output from a bank of filters matched tothe desired users’ spreading sequences, andmultiplies by a matrix which is the inverse ofthe spreading-sequence correlation matrix.

decoupled load flow a load-flow study inwhich certain simplifying assumptions per-mit an accelerated solution.

decoupling problem for 2-D linear systemsgiven the 2-D Roesser model[xhi+1,jxvi,j+1

]=[A1 A2A3 A4

][xhijxvij

]+[B1B2

]uij

yij = [C1 C2

] [xhijxvij

]i, j ∈ Z+ (the set of nonnegative integers),find matricesK = [K1,K2] andH(detH 6=0) of the control law

uij = K

[xhijxvij

]+Hvij i, j ∈ Z+

such that the transfer function matrix of theclosed-loop system

Tc(z1, z2) = [C1 C2

]×[In1z1 − A1 − B1 −K1

−A3 − B2 −K1

−A2 − B1K2In2z2 − A4 − B2K2

]×[B1B2

]H

is diagonal and nonsingular, wherexhij ∈ Rn1

and xvij ∈ Rn2 are horizontal and verticalstate vectors, respectively, anduij ∈ Rm isthe input vector,vij ∈ Rm is a new inputvector, yij ∈ Rp is the output vector, andA1, A2, A3, A4, B1, B2, C1, C2, are givenreal matrices.

decrement to reduce the value of a vari-able or content of a register by a fixed amount,normally one.

decryption a process, implemented inhardware or software, for reconstructing datapreviously coded by using a cryptography al-gorithm, that is encrypted data. These al-gorithms are typically based on keywords orcodes.

DECT See digital European cordlesstelephone.

deep-bar rotor a squirrel-cage inductionmotor rotor in which the rotor bars are deepand narrow, to make the effective resistance,and therefore the torque, higher at starting.

deep-UV lithography lithography usinglight of a wavelength in the range of about150 nm to 300 nm, with about 250 nm beingthe most common.

default the value or status that is assumedunless otherwise specified.

deferred addressing Seeindirect address-ing.

deferred evaluation a scheduling policyunder which a task is not scheduled until ithas been determined that the results of thattask are required by another task that is inexecution and cannot proceed without theseresults.


definite time DC motor accelerationwhen DC motors accelerate during their start-ing sequence, starting resistors are removedfrom the armature circuit in steps. In definitetime DC motor acceleration (also referredto as open loop DC motor acceleration), thestarting resistors are removed in definite timeincrements, whether the motor is actually ac-celerating or not.

definite-purpose motor any motor de-sign, listed and offered in standard ratingswith standard operating characteristics, withspecial mechanical features for use under ser-vice conditions other than usual or for use ona particular type of application.

definition phase or statement in which anew type, or a class, or a frame for variablesis defined. The definition of typed constantis also typically allowed.

deflector any of a number of optical de-vices that change the direction of an opticalbeam, using mechanisms such as diffraction,mechanical mirror motion, and refraction.

defocus the distance, measured along theoptical axis (i.e., perpendicular to the planeof the best focus) between the position ofa resist-coated wafer and the position if thewafer were at best focus.

defocusing quadrupole magnet a quadru-pole magnet that focuses beam in the verticalplane and defocuses in the horizontal plane.

DeForest, Lee (1873–1961) Born: Coun-cil Bluffs, Iowa, U.S.A.

DeForest is best known for his contribu-tions to the development of radio communi-cations. DeForest’s greatest invention wascalled the audion triode. This vacuum tubewas based on an earlier patented tube devel-oped by John A. Fleming. This tube, whichwas both an amplifier and a rectifier, allowedthe development of radios, radar, television,and some early computers. DeForest’s life

was noted for controversial, and often poor,business decisions.

deformable mirror device a type of de-vice for light modulation, especially spatiallight modulation, employing micromechan-ical structures, such as cantilevered mirrorsor mirrors with torsional motions, to deflectincident light rays.

defuzzification the process of transform-ing a fuzzy set into a crisp set or a real-valuednumber.

defuzzifier a fuzzy system that producesa crisp (non fuzzy) output from the results ofthe fuzzy inference engine. The most useddefuzzifiers are

1. maximum defuzzifier that selects as itsoutput the value ofy for which the mem-bership of the output membership functionµB(y) is maximum;

2.centroid defuzzifier determines the cen-ter of gravity (centroid),y ofB, and uses thisvalue as the output of the fuzzy inference sys-tem.

See alsofuzzy inference engine, fuzzyinference system.

degenerate common emitter a combi-nation of the common-emitter and emitter-follower stages with a very well-defined gain.

degenerate four-wave mixing a four-wave mixing process in which all of the in-teracting waves have the same frequency. Incertain geometrical arrangements, this pro-cess leads to optical phase conjugation andin addition can be for certain types of opticalinformation processing.

degenerate modes two modes with dif-ferent field structures having the same cutofffrequency in a waveguide or the same reso-nant frequency in a cavity.

degenerate two-wave mixing a specialcase of two wave mixing in which the twobeams are of exactly the same frequency. In


two-wave mixing, if the two laser beams areof the same frequency, a stationary interfer-ence intensity pattern is formed. This leads toa stationary volume refractive index grating.Such a kind of two-wave mixing is referredto as degenerate two-wave mixing.

degradation situation in which a signalhas been corrupted by noise, blurred by somepoint-spread function or distorted in someother fashion.

degree of membership the degree towhich a variable’s value belongs in a fuzzyset. The degree of membership varies from0 (no membership) to 1 (complete member-ship).

degree of mobility each prismatic or rev-olute joint has one degree of freedom andprovides the mechanical structure with a sin-gle degree of mobility.

degrees of freedom the number of in-dependent position variables that have to bespecified in order to locate all parts of themechanism is defined as a number of de-grees of freedom. Therefore, the degrees offreedom is defined as the minimal number ofposition variables necessary for completelyspecifying the configuration of the mecha-nism.

degree of visual angle the angle sub-tended by an object of a given width a givendistance away from the viewer.

delay (1) the time required for a signal topropagate along a wire.

(2) the difference in the absolute anglesbetween a point on a wavefront at the deviceoutput and the corresponding point on the in-cident input wavefront, expressed in secondsor degrees. Delay can exceed 360 degrees.Given bytd , we have

td = θout − θin

delay angle Seefiring angle.

delay locked loop Seedelay-locked loop.

delay power spectrum a function charac-terizing the spread of average received poweras a function of delay. Can be obtained fromthe scattering function by integrating over theDoppler shift variable. See alsoscatteringfunction, multipath propagation.

delay profile Seedelay power spectrum.

delay range the difference in arrival timesbetween the first and last significant compo-nent of the impulse response of a widebandcommunication channel. Also known as thetotal excess delay.

delay resolution the capability, measuredin units of delay (seconds), of a signal usedfor channel measurement to resolve receivedsignal components which arrive with differ-ent delays. If two signal components arrive atthe receiver with a delay separation less thanthe delay resolution, they will be observedas one signal, superimposed on each other.The actual value of the delay resolution de-pends on the criterion by which two signalcomponents are defined to be resolved. Anapproximate measure is given by the inverseof the channel (or signal) bandwidth.Seealsomultipath propagation.

delay slot in a pipelined processor, a timeslot following a branch instruction. An in-struction issued within this slot is executedregardless of whether the branch condition ismet, so it may appear that the program is ex-ecuting instructions out of order. Delay slotscan be filled (by compilers) by rearrangingthe program steps, but when this is not pos-sible, they are filled with “no-operation” in-structions.

delay spread a measure of the timethrough which the duration of a transmittedsignal is extended by dispersion in a wide-band communication channel. Usually mea-sured as the RMS delay spread, i.e., the sec-


ond moment of the time-averaged channelimpulse response.

delay-line a transmission line of the ap-propriate length to result in a specific timedelay. As an example, a line at 100 MHz thatis 90 degrees long (one-quarter wavelength)will exhibit a time delay of 2 ns.

delay-locked loop (DLL) (1) a pseudo-noise (PN) sequence tracking loop typicallyused in receivers for spread spectrum signals.The tracking loop has two branches that cor-relate the received signal with two shifts of alocally generated PN sequence: an advancedand a retarded time shift of the phase of thesignal being tracked.

(2) a technique for symbol synchroniza-tion based on time-shifted and reversed corre-lation functions of the desired symbol wave-form, which results in a control function withan s-shape (termed an s-curve). The controlfunction is used in a feedback loop similar toa PLL to adjust the timing of receiver clockused in sampling the received signal. DLLsare used, e.g., in spread-spectrum receiversto maintain chip synchronization.

delayed AGC Seedelayed automatic gaincontrol.

delayed automatic gain control auto-matic gain control in which the control mech-anism becomes effective only when the inputsignal is larger than a predetermined level.

delayed branch instruction a form ofconditional branch instruction in which one(or executed irrespective of the outcome ofthe branch). Then the branch takes effect.Used to reduce the branch penalty.

delayed neutrons neutrons emitted by fis-sion daughters after some time delay.

delectric resonator an unmetallized di-electric object of high dielectric constant andhigh quality factor that can function as an en-ergy storage device.

delta connection a three-phase powersource or load in which the elements are con-nected in series and are thus represented ona schematic diagram as a triangular configu-ration.

delta function in discrete-time, the func-tion given by

δ[n] ={

0 n 6= 01 n = 0.

In continuous-time, the (Dirac) delta “func-tion” is not a function at all and, althoughan abuse of mathematical rigor, is a tremen-dously important concept throughout signaland system theory. The Dirac delta “func-tion” is defined as

δ(t) ={

0 t 6= 0∞ t = 0

such that∫f (t)δ(t − to)dt = f (to)

delta gun cathode ray tube (CRT) elec-tron gun structure has the red, green, and blueelectron gun components configured in theshape of an equilateral triangle; the structureprovides the smallest CRT neck size and hasthe smallest deflection yoke diameter, but re-quires color registration (color convergence)correction in both the horizontal and the ver-tical CRT face.Seecathode ray tube.

delta modulation a special case of differ-ential pulse code modulation (DPCM) wherethe digital code-out represents the change, orslope, of the analog input signal, rather thanthe absolute value of the analog input signal.A “1” indicates a rising slope of the inputsignal. A “0” indicates a falling slope of theinput signal. The sampling rate is dependenton the derivative of the signal, since a rapidlychanging signal would require a rapid sam-pling rate for acceptable performance.

delta modulation control a pulse-timemodulation method transplanted from the


delta modulation in signal processing. Thedifference between the control reference andthe switched signal is integrated and fed toa Schmitt trigger. When the integrated valuereaches a predefined upper bound the switchis turned off. When the integrated valuereaches a predefined lower bound the switchis turned on.

delta rule a supervised learning algo-rithm, based upon gradient descent, that wasdeveloped for application to single-layer net-works of linear threshold units. For eachinput patternx, the weightwij connectinginput xj to unit i is adjusted according toDwij = h(ti − wi · x)xj , wherewi is thevector of weightswij, ti is the target outputfor unit i, andh is a positive constant. Thisrule is also known as the Widrow–Hoff ruleand the LMS algorithm.

delta–delta transformer a three-phasetransformer connection formed by connect-ing three single-phase transformers in whichthe windings on both the primary and the sec-ondary sides are connected in series to forma closed path.

delta-wye transform a transformationbetween delta and wye connections.

delta–wye transformer a three-phasetransformer connection formed by connect-ing three single-phase transformers in whichthe primary windings are connected in seriesto form a closed path while one end of eachof the secondary windings is connected to acommon point (the neutral).

demagnetization the act of removing adevice from being in a magnetic state, i.e.,rearranging the atomic magnetic domains ina disoriented fashion.

demagnetizing field the magnetic fieldproduced by divergences in the magnetiza-tion of a magnetic sample.

demand assign multiple access protocoltelephone signalling mechanism in which theaccess is established for the duration of a call.

demand fetch in a cache memory, thename given to fetching a line from the mem-ory into the cache on a cache miss when it isrequested and not before.

demand meter an electric meter whichshows both the energy used and the peakpower demand in a given period.

demand paging condition where eachpage in virtual memory is loaded into mainmemory, after first being referenced by theprocessor (i.e., not in advanced). The firstreference to each page will thus always causea “page fault” (page not in main memory).After these initial page faults, most of thepages needed by the program are in mainmemory.

demodulation the process by which amodulated signal is recovered back to itsoriginal form. It is the general processof extracting the information-bearing signalfrom another signal. Modulation is the gen-eral process of embedding an information-bearing signal into a second carrier sig-nal. An important objective in modulation isto produce a signal whose frequency rangeis suitable for transmission over the com-munication channel to be used.See alsomodulation.

DeMorgan’s theorem a formula for find-ing the complement of a Boolean expression.It has two forms:

1.A ∨ B = A ∧ B2.A ∧ B = A ∨ B

whereA andB are Boolean variables and∧represents logical AND and∨ represents log-ical OR and the overbar represents the logicalcomplement.

demultiplexer a logic circuit withK in-puts andI controls which steers theK inputs


to one set of 2I sets of output lines.Comparewith multiplexer.

demultiplexing the inverse operation ofmultiplexing that enables the transmission oftwo or more signals on the same circuit orcommunication channel.

Denavit–Hartenberg notation a systemthat describes the translational and rotationalrelationships between adjacent links. TheD-H representation results in 4× 4 homo-geneous transformation matrix representingeach link’s coordinate system at the joint withrespect to the previous link’s coordinate sys-tem. The D-H representation of a rigid linkdepends on four geometric parameters as-sociated with each link. Every coordinateframe is assigned according to the three rules:

1.Thezi−1 axis lies along the axis of mo-tion of theith joint.

2. Thezi axis is normal to thezi−1 axis,and pointing away from it.

3. Theyi axis completes the right-handedcoordinate system as required.

Referring to the figure, the four D-H param-eters are defined as follows:• qi is the joint angle from thexi−1 axis to thexi axis about thezi−1 axis (using the right-hand rule),• di is the distance from the origin of the(i − 1)-th coordinate frame to the intersec-tion of thezi−1 axis with thexi axis alongzi−1 axis,•ai is the offset distance from the intersectionof thezi−1 axis with thexi axis to the originof theith frame along thexi axis (in anotherwords it is the shortest distance between thezi−1 andzi axes),• αi is the offset angle from thezi−1 to theziaxis about thexi axis (using the right-handrule).

For a revolute jointdi ,ai , andαi are called thelink parametersor joint parametersand re-main constant.qi is called thejoint variable.For a prismatic joint,qi ,ai , andαi are the linkparameters and remain constant, whiledi isthe joint variable. The D-H transformation

matrix for adjacent coordinate frames has thefollowing form:

i−1Ai =

cosqi − cosαi sinqi sinαi sinqi ai cosqisinqi cosαi cosqi − sinαi cosqi ai sinqi

0 sinαi cosαi di0 0 0 1

denormalized number nonzero numberwhose leading significand bit is zero andwhose exponent has a fixed value. Thesenumbers lie in the range between the smallestnormalized number and zero.

density estimation statistical methodsfor estimating the probability density froma given set of examples.

density function (DF) an alternativename for probability density function (PDF).

density matrix representation for thewave functions of quantum mechanics interms of binary products of eigenfunction ex-pansion amplitudes; with ensemble averag-ing the density matrix representation is con-venient for phenomenological inclusion ofrelaxation processes.

density matrix formalism of quantum me-chanics a mathematical formulation of thetheory of quantum mechanics more generalthan those based on a description in termsof a wavefunction or a state vector, becauseit can treat situations in which the state ofthe system is not precisely known. The den-sity matrix formalism is often used in laserphysics and in nonlinear optics, for example,under situations in which collisional dephas-ing effects are important.

dependability system feature that com-bines such concepts as reliability, safety,maintainability, performance, and testability.

dependency a logical constraint betweentwo operations based on information flow-ing among their source and/or destinationoperands; the constraint imposes an ordering


The Denavit–Hartenberg parameters.

A PUMA 560 robot arm illustrating joints, links, and

the D-H parameters.

on the order of execution of (at least) portionsof the operations. For example, if the firstoperation in a sequential program produces aresult that is an operand of the second oper-ation of the program, that second operationcannot be performed until the first operationhas been completed, since its operand valuewill not be available earlier.

dependent source Seecontrolled source.

depleted uranium uranium in which theproportion of fissile U-235 has been reducedbelow useful levels.

depletion layer space charge region orlayer adjacent to a semiconductor junctionwhere the majority carrier concentration hasbeen depleted or reduced below the thermalequilibrium value.

depletion mode an FET that is on whenzero volts bias is applied from gate to source.

depolarization (1) the change of the po-larization state of a wave propagating throughan anisotropic medium.

(2) phenomenon that occurs when a wavebecomes partially or completely unpolarized.

(3) changing the original polarization of apropagating wave into a different type.

depolarizing scattering change in the po-larization of the light due to strong scattering.

deposited multi-chip module (MCM-D)a multi-chip module built using the deposi-tion and thin-film lithography techniques thatare similar to those used in integrated circuitmanufacturing.

depth in computer vision, the distance to asurface, as perceived subjectively by the ob-server. Also, the number of bits with whicheach pixel is represented in a digital image.

depth map a map of depth in a scene cor-responding to each pair of coordinates in animage of the scene.


depth of field the range of depths overwhich objects in the field of vision are in ac-ceptable focus.

depth of focus (DOF) the total range offocus that can be tolerated; that is, the rangeof focus that keeps the resulting printed fea-ture within a variety of specifications (such ascritical dimension, photoresist profile shape,and exposure latitude).

depth of penetration distance inside amaterial interface that an impinging wave hasattenuated by 1/ewheree = 2.7183 (Euler’sconstant).

depth-first search a search strategy fortree or trellis search where processing is per-formed depth first, i.e., a particular path isprocessed through the depth of the tree/trellisas long as it fulfills a certain threshold crite-rion (e.g., based on the Fano metric). If a pathfails the threshold test, a new path is consid-ered. Also known as sequential search.

derating factor the fraction (or percent)of nominal rating to which a specified quan-tity must be reduced due to unusual operat-ing conditions. Examples of conditions thatmay require application of a derating factorare high altitude, high ambient temperature,frequent motor starting, and “plugging” op-eration of a motor.

derivative control control scheme where-by the actuator drive signal is proportional tothe time derivative of the difference betweenthe input (desired output) and the measuredactual output.

descrambling the process of restoring atransmitted signal to its original form com-monly used in CATV systems. The positive-trap, negative-trap, and baseband are com-mon methods of scrambling requiring sucha means to restore the signal to its originalform.

descriptor an object describing an area ofspace within memory. A descriptor containsinformation about the origin and length of thearea.

desensitization a reduction in a deviceoutput signal power due to one or more addi-tional signals that compress the device out-put, expressed as a negative ratio of the de-sensed signal output power to the outputpower without the additional signals, in deci-bels.

D = −10 log10

(pdesensed

pundesensed

)

design for testability designing a semi-conductor component such that it is easierto feed it a set of test vectors that guaran-tees, or provides sufficient reassurance, thatthe component was manufactured (and de-signed) correctly.

design model a mathematical model thatis used to design a controller. The designmodel may be obtained by simplifying thetruth model of the process. The truth modelis also called the simulation model. The truthmodel is usually too complicated for con-troller design purposes. The controller per-formance is tested using the truth model.Seealsotruth model.

designed distance the guaranteed mini-mum distance of a BCH forward error con-trol code that is designed to correct up toterrors.

destination operand where the results ofan instruction are stored, e.g., the instruc-tion MOV AL,7 uses AL as the destinationoperand (7 is the source operand).

destructive physical analysis (DPA) de-vices are opened and analyzed for processintegrity and workmanship.


destructive read reading process in whichthe information is lost from memory after be-ing read out.

detectability the property of a system con-cerning the existence of a stabilizing outputinjection. For linear time-invariant systems,it is characterized by the observability (seethe definition) of the unstable modes.

detective quantum efficiency (DQE) ofa photodetector, the ratio of its quantum ef-ficiency (Qe) to its power noise factor (k),i.e., DQE = 3DQe/k. This is a generalrelationship that applies to electrical, photo-graphic, and biological (e.g., the eye) typesof photodetectors.

detector (1) a device that converts RF in-put signals to a corresponding DC output sig-nal.

(2) in optics, a circuit, usually containinga diode that converts the intensity of light intoan electrical signal. Used in a variety of ap-plications including power sensing, leveling,and modulation envelope reproduction.Seealsophotodetector.

deterministic control of uncertain systemsan approach to the control of uncertain sys-tems that is effective over a specified rangeof the system parameter variations. Thesemethods do not require on-line identificationof the values of the system parameters toyield the desired robust performance. No sta-tistical information about uncertain parame-ters is required. The two main approachesare variable structure sliding mode controland Lyapunov theory approach.

deuteron the nucleus of the deuteriumatom (an isotope of the hydrogen) consistingof a proton and a neutron.

developed power the power convertedfrom electrical to mechanical form in an elec-tric motor. The developed power of a motorcan be calculated from the developed torqueand motor speed.

developed torque the torque created byan electric motor including torque requiredto overcome the friction and windage lossesof the motor. This value will be higher thanthe shaft torque, which is actually deliveredto the load.

development the process by which a liq-uid, called the developer, selectively dis-solves a resist as a function of the exposureenergy that the resist has received.

deviation a measure of the dispersionamong the elements in a set of data. Some-times, a deviation is defined as failure to meetthe specified critical limits.

deviation ratio the allowable frequencydeviation for an FM signal divided by thehighest modulating frequency.

device a hardware entity that exists out-side of the motherboard, and is accessedthrough device drivers. Devices often relateto I/O (floppy drives, keyboards, etc.).

device controller (1) a device used to con-nect a peripheral device to the main com-puter; sometimes called a peripheral con-troller.

(2) software subroutine used to communi-cate with an I/O device.

device driver program that controls an in-put/output device, usually providing a stan-dard form of interface to the programs thatutilize the device. Device drivers performthe basic functions of device operation.

device register register in an I/O devicethat may be read or written by the processorto determine status, effect control, or transferdata.

device scaling to increase device speedand circuit density as well as reduce powerconsumption, all three dimensions of transis-tors are reduced and the impurity concentra-tions increased by interrelated scale factors.


dewar a vacuum insulated, cryogenic ra-diation insulated, low conduction heat lossvessel for cryogenic fluids. A dewar is usedfor keeping material, chiefly liquids, cold orhot. The common thermos bottle is an ex-ample of a dewar. In the accelerator businessdewars are often used to store large quantitiesof cryogenic liquids.

DF Seedissipation factor.

DFB laser Seedistributed feedback laser.

DFD Seedisplaced frame difference.

DFT Seediscrete Fourier transform.

Dhrystone SeeDhrystone benchmark.

Dhrystone benchmark synthetic bench-mark program consisting of a representativeinstruction mix used to test the performanceof a computer. Does not compute anythingin particular. Another synthetic benchmarkis the Whetstone benchmark.

DIAC a two-terminal AC device that, oncegated on by sufficient forward voltage, per-mits the flow of current until reverse biased.It is often used as a trigger device to drive thegate of other power electronic devices.

Diac’s delta function Seedelta function.

diagnostic (1) one of a set of tests to runthrough a system that determines whether thesystem is functioning correctly.

(2) pertaining to the detection and isola-tion of faults or failures. For example, a di-agnostic message, a diagnostic manual.

diagonal clipping distortion that occursin an AM demodulator (usually associatedwith diode detection), where the capacitordischarge time constant is set too long for thedetector to accurately follow fast changes inthe AM signal envelope. Sometimes referredto as “failure to follow distortion,” diagonalclipping can also occur in AM modulators

when the intelligence bandwidth exceeds thatof the modulator.

diagonal dominance a measure ofthe amount of interaction that exists be-tween variables in a multi-input–multi-output(MIMO) system. It is quantified by Gersh-gorin circles or bands that are often plottedon an inverse Nyquist array (or INA) diagramthat shows the frequency response matrix ofthe system in a graphical form. Its practicalsignificance relates to the fact that a diago-nally dominant system can be controlled bymultiple single variable controllers, whereasa nondominant process might require so-phisticated and costly multivariable tech-niques for effective control.See alsoinverseNyquist array.

diamagnetic materials with magnetiza-tion directed opposite to the magnetizingfield, so that the permeability is less than one;metallic bismuth is an example.

die an individual MMIC circuit or subsys-tem that is one of several identical chips thatare produced after dicing up an MMIC wafer.

dielectric (1) a medium that exhibits neg-ligible or no electrical conductivity and thusacts as a good electrical insulator.

(2) a medium characterized by zero con-ductivity, unity relative permeability, and arelative permittivity greater than one. Alsoknown as an insulator.

Dielectries are usually used to separatetwo conducting bodies such as to form a ca-pacitor.

dielectric constant (1) a quantity that de-scribes how a material stores and dissipateselectrical energy.

(2) ratio of the electrical capacity of a con-denser, which has a given material as the di-electric, to the capacity of an identical con-denser, but with air as the dielectric.

(3) permittivity of a medium normalizedto the permittivity of free space; a measure


of the response of a dielectric to an appliedelectric field.

(4) an electric property of an insulator orsemi-conducting material, which describeshow differently electric fields will behave in-side of the material as compared to air. Asan example,er = 12.9 for GaAs as com-pared toer = 1 for air. In integrated circuits,an effective dielectric constant (eeff ) is used,since the electric fields supported by the sig-nals traveling through the conductors on thecircuit flow through both air and the insulatoror semiconductor simultaneously.

dielectric discontinuity interface be-tween two media with different dielectric per-mittivity properties.

dielectric medium medium that is polar-izable but relatively nonconducting.

dielectric resonator an unmetallized di-electric object of high dielectric constant andhigh quality factor that can function as an en-ergy storage device.

dielectric resonator [stabled] oscillator(DRO) a dielectric resonator is a cylindri-cally shaped piece of material, or “puck,” thathas the properties of having low-loss resonantfrequencies that are determined primarily bythe size of the cylinder. Placing a dielectricresonator near a microstrip line can form aresonant circuit that will frequency stabilizea voltage-controlled oscillator.

dielectric resonator antenna (DRA) anantenna where a dielectric resonator is usedas the radiation element.

dielectric slug tuner system of two mov-able dielectric pieces of material placed on atransmission line for the purpose of matchinga wide range of load impedances by meansof placing the dielectrics in proper positions.

dielectric step discontinuity the junctionbetween different dielectric waveguides.

dielectric waveguide a waveguide that re-lies on differences in permittivity among twoor more materials to guide electromagneticenergy without the need for ground planes ormetallic strips. Such guides of rectangular,circular, elliptical, and other cross sectionsare made of dielectric materials and used fortransmitting signals. Transmission is accom-plished by the total internal reflection mech-anism inside the waveguide.

difference amplifier See differentialamplifier.

difference engine a mechanical calculatordeveloped by Babbage in 1823.

difference equation the mathematicalmodel of a LTIL discrete time system.Seealsodiscrete time system, LTIL system.

difference of Gaussian filter a bandpassfilter whose point spread function is the dif-ference of two isotropic Gaussians with dif-ferent variances. The result is a “Mexicanhat” shape similar to the Laplacian of a Gaus-sian (SeeMarr-Hildreth operator). Variousphysiological sensors, including some filtersin early vision, appear to have DOG pointspread functions.

difference-frequency generation a sec-ond-order nonlinear optical process in whichtwo input beams are applied to a nonlinearoptical material and an output is produced atthe difference of the frequencies of the twoinput beams.

difference-mode signal if two arbitrarysignalsv1 and v2 are applied to the inputsof a differential amplifier, then the common-mode signal is the arithmetic average of thetwo signals. That is,

(v1 + v2)/2

differential amplifier an amplifier in-tended to respond only to the difference be-


tween its input voltages, while rejecting anysignal common to both inputs.

The differential amplifier is designed suchthat the difference between the two inputs isamplified (high differential gain), while thesignals appearing at either individual input(referenced to ground potential) sees a verylow gain (low common-mode gain, usuallyloss). The differential amplifier is usuallyused as the first component at the receivingend of a communications link using twistedpair cable (either shielded or unshielded) asthe transmission medium. This provides amethod to reject any common-mode noiseinduced onto the twisted pair transmissionline, including common-mode noise fallingwithin the useful bandwidth of the commu-nications link. The figure of merit for thedifferential amplifier is its common mode re-jection ratio (CMRR), computed by dividingthe differential-mode gain by the common-mode gain.

differential coding a coding scheme thatcodes the differences between samples.Seepredictive coding.

differential entropy the entropy of a con-tinuous random variable. For a random vari-able X, with probability density functionf (x) on the support setS, the differentialentropyh(X) is defined as

h(X) = −∫Sf (x) logf (x)

provided the integral exists. Also writ-tenh(f ), emphasizing the sole dependenceupon the density.See alsoentropy, relativeentropy, mutual information.

differential gain the amplification factorof a circuit that is proportional to the differ-ence of two input signals. The differentialgain may be expressed in percentage formby multiplying the above amplification fac-tor by 100, or in decibels by multiplying thecommon logarithm of the differential gain by20.

differential inclusion a multivalued dif-ferential equation,

x ∈ F(t, x) ,

whereF(t, x) is a nonempty set of velocityvectors atx ∈ R

n for each timet on sometime interval. The setF(t, x) can be viewedas the set of all possible “velocities”x(t) of adynamical system modeled by the multival-ued, or multifunction, differential equation.A solution x(t) is an absolutely continuousfunction on some time interval whose veloc-ity vector x lies in the setF(t, x) for almostall t . See alsoFilippov method.

differential kinematics equationv =J (q)q can be interpreted as the differentialkinematics mapping relating then compo-nents of the joint velocity vector to ther ≤ m

components of the velocity vectorv of con-cern for the specific task. Heren denotesnumber of degrees of mobility of the struc-ture,m is the number of operational spacevariables, andr is the number of operationalspace variables necessary to specify a giventask.See alsogeometric Jacobian.

differential length vector the vector sumof the differential length changes in each ofthe three coordinate directions along a givencurve.

differential mode gain for a differentialamplifier, the ratio of the output signal ampli-tude to the amplitude of the difference signalbetween the amplifier input terminals.

differential pair a two-transistor BJT(FET) amplifier in which a differential in-put signal is applied to the base (gate) ter-minals of the two transistors, the output istaken differentially from the collector (drain)terminals, and the emitter (source) terminalsare connected together to a constant currentsource. Also known as an emitter-coupledpair (BJT) or source-coupled pair (FET). Thisconfiguration is often used as the basis of thedifferential input stage in voltage feedbackop-amps.


differential pair oscillator a device usedinstead of a transistor in any LC-oscillator.Two distinct advantages result from employ-ing the differential pair as the active element.The first is that the output signal may be takenat the collector of transistor that is externalto the oscillator feedback loop, and second isthat, if a tuned circuit is used as a load, the dis-tortion of the output signal is much less thanit would be for a single transistor oscillator.The second advantage follows from the factthat the differential pair collector currents donot include even harmonic components and,in addition, the amplitudes of existing high-order harmonics are smaller than they are fora single transistor.

differential peak detector a circuit com-monly used for the demodulation of FM sig-nals; it utilizes two peak detectors, a differ-ential amplifier and a frequency selective cir-cuit. Also known as a balanced peak detector.

differential protection a protective relay-ing scheme in which the currents entering andleaving the protected line or device are com-pared.

differential protection unit a protectiveunit based on the difference of currents flow-ing in and out of a protected zone.

differential pulse code modulation(DPCM) (1) a class of methods forpulse code modulation (or scalar quantiza-tion) where (linear) prediction is used in or-der to utilize the temporal redundancy in thesource signal to enhance performance. Alsoreferred to as predictive PCM or predictiveSQ.See alsopulse-code modulation, scalarquantization, adaptive differential pulse codemodulation.

(2) in image processing, a lossy predictivecoding scheme. In this schemem pixels ina causal neighborhood of the current pixel isused to estimate (predict) the current pixel’svalue. The basic components of the predic-tive coder comprises predictor, quantizer, andcode assigner.

differential relay a differential relay isa protective relay that measures current go-ing into a device from all sources by meansof a network of paralleled current transform-ers. Ideally, the operational current is zero fornormal conditions, and rises to a high value(proportional to fault current) when a faultcomes on inside the differential zone. Dif-ferential relays are commonly applied in busprotection, transformer protection, generatorprotection, and large motor protection.

differential volume element in a givencoordinate system, the product of the differ-ential length changes in each of the three co-ordinate directions.

differential-mode coupling pick-up froman electromagnetic field that induces achange in potential on both signal leads ofequal magnitude but opposite phase relativeto the ground reference potential.

differentially compounded a compoundmachine in which the flux produced by theMMF of the shunt field winding and theflux produced by the MMF of the seriesfield winding oppose each other. Most of-ten obtained by incorrectly connecting themachine, the differentially compounded ma-chine may demonstrate very erratic behavior.

differentiator a filter that performs a dif-ferentiation of the signal. Since convolutionand differentiation are both linear operations,they can be performed in either order.

(f ∗ g)′(x) = f ′(x) ∗ g(x) = f (x) ∗ g′(x).

Thus, instead of filtering a signal and thendifferentiating the result, differentiating thefilter and applying it to the signal has the sameeffect. This filter is called a differentiator.A low-pass filter is commonly differentiatedand used as a differentiator.

diffracted beam diffraction that takesplace when the wavelength of an incidentbeam is short compared to the interaction


distance. Particles exhibit wave-like charac-teristics in their passage through matter. Instriking a target the incident beam scatters offnucleons. The scattered waves then combineaccording to the superposition principle, andthe peak of this scattered wave is called thediffracted beam.

diffraction (1) distortion of an electro-magnetic wave due to the proximity of aboundary or aperture.

(2) a bending or scattering of electromag-netic waves. Basically a redistribution withina wavefront when it passes near the edge ofan opaque object.

(3) the propagation of light in the presenceof boundaries. It is the property of light thatcauses the wavefront to bend as it passes anedge.

diffraction angle angle corresponding ap-proximately to the rate of spreading of anelectromagnetic wave that has been trans-mitted through an aperture; with Gaussianbeams the far field half angle for a radiusequal to the spot size.

diffraction coefficient in the GeometricTheory of Diffraction, the coefficient that isproportional to the contribution to the scat-tered field due to the fringe currents near anedge or corner of a scattering target.

diffraction efficiency of Bragg cell ra-tio of the intensity of the principal diffractedbeam to the intensity of the undiffractedbeam.

diffraction grating an array of reflectingor transmitting lines that mutually enhancethe effects of diffraction.

diffraction loss loss from an electromag-netic beam due to finite aperture effects.

diffraction tomography generalizationof computerized tomography incorporatingscattering effects.

diffuse density signal that has uniformenergy density, meaning that the energy fluxis equal in all parts of a given region.

diffuse intensity the energy scattered inall directions out of the forward or speculardirections. Sometimes also called incoherentcomponent of the intensity.

diffuse multipath the result of multipathpropagation observed as overlapping signalcomponents, due to delay differences of mul-tipath components being less than the delayresolution of the signal. Observable in the de-lay power spectrum as a continuous distribu-tion of power over delay.See alsomultipathpropagation, delay power spectrum.

diffuse scattering the component of thescattering from a rough surface that is not inthe specular direction. It is caused by reflec-tions from local surfaces oriented in planesdifferent from that of the mean surface.Seealsospecular scattering.

diffuse transmittance a transmitted sig-nal that has uniform energy density.

diffusion a region of a semiconductor intowhich a very high concentration of impurityhas been diffused in order to substantially in-crease the majority carrier concentration inthat region.

diffusion pump second stage of the vac-uum system. Hot oil showers the particles ina vacuum and creates a better vacuum. Af-ter a mechanical (roughing) pump is used toremove about 99.99% of the air in the beamtube, the remaining air can then be removedby a diffusion pump, down to about 1E−9

torr.

diffusion under field (DUF) a local thinlayer of semiconductor with a very high car-rier concentration located under and in con-tact with the collector of a vertical bipolartransistor to provide a low-resistivity connec-tion to it.


diffusive scattering when the photonmean free path is much smaller than the scat-terer dimensions and then the energy is scat-tered uniformly in all directions.

DigiCipher HDTV system a high-defini-tion television (HDTV) digital transmissiontelevision system proposed to the FCC bythe American Television Alliance composedof General Instruments and MassachusettsInstitute of Technology. The DigiCipherHDTV proposal submitted to the FCC in Au-gust of 1991 was the first system to providean all digital television system that promisedspectrum compatibility with the existing tele-vision channel allocation. The system usedquadrature amplitude modulation (16-QAM)digital transmission at a 4.88 MHz symbolrate.

digital circuits or systems that employ twovalued (binary) signals denoted by the digits0 and 1. Normally binary 1 is used to indicatehigh/true and binary 0 to indicate low/false(Positive Logic).

digital cellular radio cellular radio prod-uct designed to transmit its signals digitally.

digital communications communicationtechniques that employs binary bits to encodeinformation.

digital European cordless telephone(DECT) a digital microcell system op-erating in the 1.88–1.90 GHz band designedto provide high-capacity wireless voice anddata services indoors and outdoors in smalllocal networks. DECT systems use aTDMA/FDMA multiple access scheme, with10 radio frequency carriers, each 1728 kHzwide, divided into 2× 12 time slots. Thisprovides a total of 120 duplex traffic chan-nels, with duplexing via TDD. The carrier bitrate is 1152 kb/s. DECT portable terminalshave a peak transmit power of 250 mW.

digital filter the computational processor algorithm by which a sampled signal or

sequence of numbers (acting as an input) istransformed into a second sequence of num-bers termed the output signal. The compu-tational process may be that of low-pass fil-tering (smoothing), bandpass filtering, high-pass filtering, interpolation, the generation ofderivatives, etc.

digital halftone halftone technique basedon patterns of same size dots designed to sim-ulate a shade of gray between white paper andfull colorant coverage.

digital image (1) an array of numbers rep-resenting the spatial distribution of energy ina scene obtained by a process of samplingand quantization.

(2) a discrete-value functionI (k, l) =I (k1x, l1y) obtained by sampling, at equi-spaced positions, the continuous functionI (x, y) which measures image intensity atpositionx, y of the image plane.I can besingle-valued for monochrome images orm-valued (usuallym = 3) for color images.

digital modulation signal generator anRF signal generator capable of providing sig-nals with digital modulation formats suchas Gaussian minimum shift keying (GMSK),Pi/4 differential quadrature phase shift key-ing (DQPSK), and code division multiple ac-cess (CDMA).

digital optical computing optical com-puting that deals with binary number opera-tions, logic gates, and other efforts to even-tually build a general-purpose digital opticalcomputer. In digital optical computing, newoptical devices are sought to replace elementsin an electronic computer. The digital opti-cal computer may be primarily based on al-ready known computer architectures and al-gorithms.

digital optics optical systems that handledigital data.


digital relay a relay in which decisionsare made by a digital computer, typically amicroprocessor system.

digital serial processing processing ofmore than one but not all bits in one clockcycle. If the digit size isW1 and the wordlength isW , then the word is processed inW/W1 clock cycles. IfW1 = 1, then thesystem is referred to as a bit-serial, and ifW1 = W , then the system is referred to as abit-parallel system. In general, the digit sizeW1 need not be a divisor of the word lengthW , since the least and most significant bitsof consecutive words can be overlapped andprocessed in the same clock cycle.

digital signal processor (DSP) micro-processor specifically designed for process-ing digital signals. DSPs are typically wellsuited to perform multiplications and addi-tions in chain, even in floating point. Theyare less suitable for managing interrupts andlarge amounts of memory. For reaching highperformance, a neat division between mem-ory for data and memory for programs isadopted, with the constraint of having a highnumber of pins.

digital simulator a simulator that allowsthe user to check the function and perform atiming analysis of a digital system.

digital subscriber line (DSL) in tele-phony, a digital connection between a cus-tomer premise and a central switching office(CO) using twisted-pair (copper) as the trans-mission medium. Although DSLs were orig-inally introduced for narrowband ISDN ap-plications (144 kbps), recent enhancementsof DSLs (definitions follow) now support abroader range of higher-rate services.

digital sum variation a measure of themaximum possible imbalance in a line codedsequence. Definitions vary; a common defi-nition is the total number of running digitalsum values that can arise in the encoded se-quence.

digital tachometer a device with a sensorthat senses pulses from a rotating axis andconverts them to digital output calibrated inrotations per minute (rpm).

digital voltmeter (DVM) a modern solid-state device capable of measuring voltage anddisplaying the value in digitized form. Theterm is also used loosely for the digital mul-timeter, which can also measure current andresistance.

digital–optical computing that branch ofoptical computing that involves the develop-ment of optical techniques to perform digitalcomputations.

digital-to-analog converter (DAC, D/A)a device that changes a digital signal to ananalog signal of corresponding magnitude.

digital-video effects application of digi-tal technology to manipulate video informa-tion for production, to compress video data,to transmit video signals, and to process ortransform video signals for various displaysystems.

The original analog video is digitized forapplication to computer-type circuits that canproduce effects such as video mixing or over-lay; editing of video signals; compressionof video data; synchronizing video systems;signal transformation and timebase correc-tion for various display formats and signalconversion back to analog form. The primaryvideo artifacts observed from the digital-video effects result from either digitizing toan insufficient number of bits per picture el-ement (pixel) or from too few samples forthe video block processing algorithms. Us-ing fewer than 8 bits/pixel for each colorcomponent will cause poor signal-to-noiseratios and produce scenes that have the ap-pearance of a poster or cartoon. The posterappearance results from contours that are tooabrupt and from colors that are not smoothlyblended. Similarly, algorithms that processlarge blocks of pixels tend to reduce resolu-tion and produce blocks with color shifts.


digitization a process applied to a con-tinuous quantity that samples the quantityfirst, say, in time or spatial domain, and thenquantizes the sampled value. For instance, acontinuous-time signal can be first sampledand then quantized to form a digital signal,which has been discretized in both time andmagnitude.

digitize the action of converting informa-tion from analog to digital form.

digitizer Seedata tablet.

dilation a fundamental operation in math-ematical morphology. Given a structuringelementB, the dilation byB is the opera-tor transformingX into the Minkowski sumX ⊕ B, which is defined as follows:1. If bothX andB are subsets of a spaceE,

X ⊕ B = {x + b | x ∈ X, b ∈ B}

2. If X is a gray-level image on a spaceE andB is a subset ofE, for everyp ∈ E we have

(X ⊕ B)(p) = supb∈B

X(p − b)

3. If bothX andB are gray-level images ona spaceE, for everyp ∈ E we have

(X ⊕ B)(p) = suph∈E

[X(p − h)+ B(h)]

with the convention∞ − ∞ = −∞ whenX(p − h), B(h) = ±∞. (In the two itemsabove,X(q) designates the gray-level of thepointq ∈ E in the gray-level imageX.) Seeerosion, structuring element.

dilation equation the equation

π(t) =inf ty∑

n=−inf tya(n)π(2t − n)

withπ(t) being scaling function anda(n) be-ing the coefficients. It states the fact that, inmultiresolution analysis, a scaling space iscontained in a scaling space with finer scale.

Dill parameters three parameters, namedA, B, and C, that are used in the Dill exposuremodel for photoresists. A and B representthe bleachable and nonbleachable absorptioncoefficients of the resist, respectively, and Crepresents the first-order kinetic rate constantof the exposure reaction. (Named for Freder-ick Dill, the first to publish this model.) Alsocalled the ABC parameters.

dilution transformer one of severalhedges, (e.g., somewhat, quite, rather, andsort of) that dilute the characteristics of afuzzy set.

diminished radix complement form ofthe complement representation of negativenumbers. In the binary system, the radixcomplement is called the 2s complement andthe diminished radix complement is calledthe 1s complement.

diode a two-terminal device that permitsthe flow of electric current in only one direc-tion.

Diodes are most often constructed byabutting n-type and p-type regions of a semi-conductor, that has significantly higher elec-trical conductivity in one direction (forward-biased) than the other (reverse-biased).

Diode devices may be specially designedfor low-power, high switching speed appli-cations (signal diodes) or higher-power ap-plications (rectifier diodes).

diode detector a device that by use of rec-tification and the use of inherent nonlinearityseparates a modulating signal from its carrier.

diode gun Plumbicon a Plumbicon tubewith an electron gun that operates with posi-tive voltage applied toG1 with respect to thecathode. The diode gun principle provides afiner beam spot size and lower beam temper-ature. This results in higher resolution andimproved lag performance compared to tri-ode gun tubes. The diode gun also providesa much higher current reserve for highlight


handling when used in conjunction with a dy-namic beam control circuit.

diode laser laser in which the amplifi-cation takes place in an electrically pumpedsemiconducting medium. Also known as asemiconductor laser or a heterojunction laser.

diode rectifier a circuit in which the out-put voltage is fixed by the circuit parametersand the load. The direction of power flow isnot reversible. An example of a single-phasediode-bridge rectifier with a capacitor filteris shown. Note that the diodes are on onlyfor a short duration, while the rectified linevoltage is greater than the capacitor voltage.

dioptric an optical system made up ofonly refractive elements (lenses).

dip Seesag.

DIP Seedual in-line package.

DIP switch set of micro-switches (on/offor deviators) that are compliant with the DIPfor the position of their pin (connections);thus, they can be installed in standard socketsfor integrated circuits.

dipole Seemagnetic dipole.

dipole antenna a straight wire, with twoarms, of oppositely pulsating charges, typi-cally λ/2 orλ/4.

Dirac’s delta function not actually afunction, Dirac’s delta function is definedfunctionally by its property of “choosing” asingle value of the integrand when integrated:

f (r0) =∫f (r)δ(r − r0)dr

when the volume of integration includes thepoint r0. See delta function.

direct addressing address of the operand(data upon which the instruction operates) ofthe instruction is included as part of the in-struction.

direct axis (d axis) the magnetic axis ofthe rotor field winding of a synchronous ma-chine. The axis between the poles of a DCmachine.

direct axis magnetizing (armature) reac-tance a reactance that represents all theinductive effects of the d-axis stator currentof a synchronous machine, except for thatdue to the stator winding leakage reactance.In Park’s d-axis equivalent circuit of the syn-chronous machine, this reactance is the onlyelement through which both the stator androtor currents flow. Its value may be deter-mined by subtracting the stator winding leak-age reactance from the steady-state value ofthe d-axis operational impedance or from thegeometric and material data of the machine.

direct axis synchronous reactance thesum of the stator winding leakage reactanceand the direct-axis magnetizing (armature)reactance of a synchronous machine. Thisrepresents the balanced steady-state value ofthe direct-axis operational impedance of thesynchronous machine, and thus characterizesthe equivalent reactance of the machine dur-ing steady-state operation.

direct axis transient reactance a valuethat characterizes the equivalent reactanceof the d-axis windings of the synchronousmachine between the initial time followinga system disturbance (subtransient interval)and the steady state. This reactance can-not be directly mathematically related to thed-axis operational impedance. However, inmodels in which the rotor windings are rep-resented as lumped parameter circuits, the d-axis transient reactance is expressed in closedform as the sum of the stator winding leak-age reactance, and the parallel combinationof the d-axis magnetizing reactance and thefield winding leakage reactance.

direct broadcast satellite (DBS) refersto TV signal transmission and distributionfrom a base station up to a satellite, and thendown to consumers who have suitable satel-


lite receiving antennas and down-converterreceivers.

F(dB) = 10 log10(SNRin/SNRout)

direct burial (1) the practice of burying aspecially-armored power or communicationscable in a ditch without the use of a surround-ing conduit.

(2) a term applied to any cable which ismeant for direct burial.

direct control bottom, first, control layerof a multilayer controller, directly responsi-ble for adjusting the manipulated inputs to thecontrolled process; typical example of directcontrol is the regulation layer of an industrialcontrol system, where the manipulated inputsare used to make the controlled variables fol-low the desired set point values.

direct control layer Seedirect control.

direct converter a frequency converterthat converts an RF signal to a baseband sig-nal directly in receivers. It converts a base-band signal to an RF signal directly in trans-mitters.

direct current machine a DC machineis an electromechanical dynamo that eitherconverts direct current electrical power intomechanical power (DC motor), or convertsmechanical power into direct current electri-cal power (DC generator). Some DC ma-chines are designed to perform either of thesefunctions, depending on the energy source tothe dynamo.

direct current motor a rotation machineenergized by DC electrical energy and usedto convert electrical energy to mechanical en-ergy.

direct digital synthesizer an oscillatorthat generates sinusoidal wave by digital cal-culation and digital to analog conversion. Itcan generate an arbitrary frequency signal ina fine channel step.

direct drive a drive in which no gear re-ducer is used.

direct drive robot a mechanical armwhere all or part of the active arm joints areactuated with the direct drive. Due to the factthat many actuators are best suited to rela-tively high speeds and low torques, a speedreduction system is required. Gears are themost common elements used for reduction.A robot with a gear mechanism is called ageared robot. Gears are located at differentjoints; therefore, usually geared robots havea transmission system, which is needed totransfer the motion from the actuator to thejoint.

direct dynamics the direct dynamicsproblem consists of determining the jointgeneralized accelerationsq(t) assuming thatjoint generalized forcesτ(k), joint positionsq(t), and joint velocitiesq(t) are known. So-lution of the direct dynamics leads to the dy-namic simulation which is defined as follows:for a given set of joint generalized forces andinitial values of the joint positions and veloci-ties integrate the equations of motion in orderto find a set of joint accelerations, velocities,and positions. The dynamics simulation isvery useful for manipulator simulation.

direct fuzzy control the use of fuzzy con-trol directly in the inner control loop of afeedback control system.

direct kinematics direct kinematics (orforward kinematics) for an arbitrary manip-ulator and given the joint variables vectorq(t) = [q1(t), q2(t), . . . , qn(t)]T and the ge-ometric link parameters finds the position andorientation of the end-effector of the manip-ulator with respect to a reference coordinateframe. The direct kinematics problem canbe solved by successive multiplication of theD-H transformation matrices from the baseof the manipulator towards its end-effector.More precisely the kinematics can be repre-sented mathematically as a continuous mapassigning to every joint position (configu-


ration of the manipulator) a correspondingposition and orientation of the end-effector.k : Q → E whereQ andE are called, re-spectively, the inertial (joint) space and theexternal space of the manipulator. Supposethat a manipulator hasr-revolute and(n− r)prismatic joints. Then the internal space isdefined asQ = T r×Rn−r , whereT r denotesther dimensional torus, andRn−r (n−r) di-mensional space of real numbers.

direct mapped Seedirect mapped cache.

direct mapped cache a cache where eachmain memory (MM) block is mapped di-rectly to a specific cache block. Since thecache is much smaller than the MM, severalMM blocks map to the same cache block.

If, for example, the cache can hold 128blocks, MM block k will map onto cacheblock k modulo 128. Because several MMblocks map onto the same cache block, con-tention may arise for that position. This isresolved by allowing the new block to over-write the old one, making the replacementalgorithm very trivial in this case.

In its implementation, a high-speed ran-dom access memory is used in which eachcache line and the most significant bits of itsmain memory address (the tag) are held to-gether in the cache at a location given by theleast significant bits of the memory address(the index). After the cache line is selected byits index, the tag is compared with the mostsignificant bits of the required memory ad-dress to find whether the line is the requiredline and to access the line.

direct memory access (DMA) used ina computer system when transferring blocksof information between I/O devices (e.g.,disk memory) to/from the main memorywith minimal intervention from the CPU. A“DMA controller” is used and can, after initi-ation by the CPU, take control of the address,control, and data busses. The CPU initiatesthe DMA controller with parameters such asthe start address of the block in main mem-ory, number of bytes to be transferred, and

the type of transfer requested (read or write).The transfer is then completely handled bythe DMA controller, and the CPU is typicallynotified by an interrupt when the transfer ser-vice is completed. While the DMA transferis in progress, the CPU can continue execut-ing the program doing other things. How-ever, as this may cause access conflicts of thebusses between the CPU and the DMA con-troller, a “memory bus controller” handlesprioritized bus requests from these units. Thehighest priority is given to the DMA trans-fer, since this normally involves synchronousdata transfer that cannot wait (e.g., a disk ortape drive). Since the CPU normally origi-nates the majority of memory access cycles,the DMA control is considered as “stealing”bus cycles from the CPU. For this reason, thistechnique is normally referred to as “cyclestealing.”

direct method Lyapunov’s second methodof investigating the stability of dynamicalsystems. The method is called a directmethod because no knowledge of the solu-tion of the differential equations modeling adynamical system is required when investi-gating the stability of an equilibrium solution.See alsoequilibrium solutionandstability.

direct method coordination coordina-tion by the direct method amounts to iteratingthe coordination variables (direct coordina-tion instruments) defined as the interactioninputs and outputs and, if needed, any othervariables that, when fixed, provide for inde-pendence of the local decision problems. Theresults of these problems are used by the co-ordinator to check whether its objectives aresatisfied — if not, then the direct coordinat-ing instruments are changed (iterated), etc.

direct modulation modulation of the op-tical intensity output from a semiconductordiode laser by direct modulation of the biascurrent.

direct scattering theory predicts the dis-tribution of scattered intensity from knowl-


edge of the structure of the inhomogeneousmedium.

direct semiconductor a semiconductorwhose band gaps are between electronicstates with the same momentum, thus allow-ing optical transitions to occur between them.

direct sequence refers to a technique fordigital spread spectrum modulation. Thedata symbol is modulated by a higher-ratepseudo-random sequence. The resulting sig-nal is thus at a higher rate, or equivalently,occupies a larger bandwidth. This is denoteddirect sequence spreading.

direct stroke (1) a lightning strike whichterminates on a piece of equipment

(2) an approach to the design of over-head electric transmission lines which as-sumes that only direct strokes to the line willbe disruptive to a power system.

direct write lithography a lithographymethod whereby the pattern is written di-rectly on the wafer without the use of a mask.

direct-access storage storage in whichan item can be accessed without having tofirst access all other items that precede it inthe storage; however, sequential access maybe required to a few preceding systems. Anexample is a disk, in which blocks may beaccessed independently, but access to a lo-cation within a block is preceded by accessto earlier locations in the block.See alsosequential-access storage.

direct-axis subtransient open-circuit timeconstant a constant that characterizes theinitial decay of transients in the d-axis vari-ables of the synchronous machine with thestator windings open-circuited. The intervalcharacterized is that immediately followinga disturbance, during which the effects of allamortisseur windings are considered. A de-tailed (derived) closed-form expression forthe subtransient open-circuit time constant ofa machine with a single d-axis amortisseur

windings is obtained by taking the reciprocalof the smallest root of the denominator of thed-axis operational impedance. An approxi-mate (standard) value is often used, in whichit is assumed the field winding resistance isvery small and the detailed expression sim-plified.

direct-axis subtransient reactance thehigh-frequency asymptote of the d-axis op-erational impedance of a synchronous ma-chine. This value characterizes the equiva-lent reactance of the d axis of the machineduring the initial time following a systemdisturbance. In models in which the rotorwindings are represented as lumped parame-ter circuits, the d-axis subtransient reactanceis expressed in closed form as the sum of thestator winding leakage reactance, and the par-allel combination of the d-axis magnetizingreactance and the d-axis rotor leakage reac-tances.

direct-axis subtransient short-circuit timeconstant a constant that characterizes theinitial decay of transients in the d-axis vari-ables of the synchronous machine with thestator windings short-circuited. The inter-val characterized is that immediately follow-ing a disturbance, during which the effectsof amortisseur windings are considered. Adetailed (derived) closed-form expression forthe subtransient short-circuit time constant ofa machine with a single d-axis amortisseurwinding is obtained by taking the reciprocalof the largest root of the numerator of the d-axis operational impedance. An approximate(standard) value is often used, in which it isassumed the field winding resistance is smalland the detailed expression simplified.

direct-axis transient open-circuit timeconstant a constant that characterizes thedecay of transients in the d-axis variablesof the synchronous machine with the statorwindings open-circuited. The interval char-acterized is that following the subtransient in-terval, but prior to steady-state, during whichthe effects of the amortisseur windings are


small (possibly negligible). A detailed (de-rived) closed-form expression for the tran-sient open-circuit time constant of a machinewith a single d-axis amortisseur winding isobtained by taking the reciprocal of the small-est root of the denominator of the d-axis op-erational impedance. An approximate (stan-dard) value is often used, in which it is as-sumed the amortisseur winding resistance isinfinite and the detailed expression simpli-fied.

direct-axis transient short-circuit timeconstant a constant that characterizes thedecay of transients in the d-axis variablesof the synchronous machine with the sta-tor windings short-circuited. The intervalcharacterized is that following the subtran-sient interval, but prior to steady-state, inwhich the effects of the amortisseur wind-ings are small (possibly negligible). A de-tailed (derived) closed-form expression forthe short-circuit transient time constant is ob-tained by taking the reciprocal of the small-est root of the numerator of the d-axis op-erational impedance. An approximate (stan-dard) value is often used, in which it is as-sumed the amortisseur winding resistance isinfinite and the detailed expression simpli-fied.

direction cosine similarity between twovariablesx = (x0, ..., xn) andy = (y0, ...,

yn), it is defined as

cosθ = (x, y)

‖ x ‖ ‖ y ‖where(x, y) is the inner product ofx andyand‖ x ‖ is the Euclidean norm ofx.

direction line a curve to which the givenfield is tangential at every point on the curve.Also called stream line or flux line.

direction of arrival Seeangle of arrival.

directional coupler a passive, 3 or 4 portdevice used to sample a portion of the forward(incident) signal or the reverse (reflected) sig-

nal, or both (dual directional coupler) in anRF, microwave, or mmW circuit. Directionalcouplers are usually described in terms ofcoupling factor and directivity. The couplingfactor describes what fraction of the incident(or reflected) power appears at the desiredcoupled port. Directivity describes the frac-tion of power coupled to the same port due toreverse power in the main arm of the coupler.

directional overcurrent relay an over-current relay that operates only for overcur-rents flowing in the tripping direction. Di-rection sensing is typically done with respectto a voltage or current signal, which is notaffected by fault location.

directional power relay a protective re-lay that operates for power flow in a givendirection. Applications are in cases wherenormal power flow is in one direction, includ-ing anti-motoring protection on a turbine-generator and fault backfeed protection onparallel step-down transformers.

directivity the maximum ratio of an an-tenna’s ability to focus or receive power ina given direction relative to a standard; thestandard is usually an isotropic radiator or adipole. Only depends on the radiation patternshape and does not include the efficiency ofthe antenna.

directory a table used in the directorymethod to maintain cache coherence in mul-tiprocessors. Contains entries identifying thecaches that hold copies of memory locations.

directory look-aside table (DLT) Seetranslation lookaside buffer.

Dirichlet conditions a set of conditionsguaranteeing that a signalx(t) will be equalto theN -term Fourier approximation ofx(t)asN → ∞ except at isolated values of t forwhichx(t) is discontinuous. The conditionsare that

(1) the signalx(t) must be absolutely in-tegrable,


(2)x(t)must have a finite number of max-ima and minima during a period, and (3)x(t)must have a finite number of (finite) discon-tinuities in any finite interval of time.

dirty bit a status bit used to indicate if ablock (e.g., cache block, page, etc.) at somelevel of the memory hierarchy has been mod-ified (written) since it was first loaded in.When the block is to be replaced with an-other block, the dirty bit is first checked tosee whether the block has been modified. Ifit has, the block is written back to the nextlower level. Otherwise, the block is not writ-ten back.

dirty page a page in memory that has beenaltered since last loaded into main memory.See alsodirty bit.

disable action that renders a device inca-pable of performing its function; the oppositeof enable.

disagreement of interests situation inwhich there are several decision units withconflicting goals.Compare withconsistencyof interests.

disassembler a computer program thatcan take an executable image and convert itback into assembly code.

disc Seemagnetic disk. Also spelled“disk.”

DISC Cerenkov counter stands for dif-ferential isochronous self-collimating Cerenkovcounter. This device is used to identify parti-cles over a wide range of masses and can alsobe used to give an independent calibration ofthe average momentum of a beam line.

disco contraction of "distribution com-pany," a firm which owns the electric distri-bution network in a service area but neithergenerates nor transmits electric power.

disconnect switch a manually operatedswitching device used to disconnect circuitconductors and their associated load(s) theirsource of electrical power.

discontinuity effect an appropriate equiv-alent circuit model for discontinuities thathave a tendency to disturb the electric andmagnetic fields in their vicinity.

discontinuity manifold See slidingsurface.

discontinuous control the control lawthat is allowed to vary discontinuously to ac-count for sudden switching. An example ofa discontinuous controller is

u = −Usign(s(e)) ,

wheree is a control error ands is a functionof e, where

sign(s) ={−1 if s < 0

1 if s > 0.

discrete cosine transform (DCT) a popu-lar format for video compression. The spatialsignal is expanded in a cosine series, wherethe higher frequencies represent increasedvideo resolution.

The forward 2-D DCT of ann × n blockis defined as

F(u, v) = 4C(u)C(v)

n2

n−1∑j=0

n−1∑k=0

f (j, k)

cos

((2j + 1)uπ

2n

)cos

((2k + 1)vπ

2n

)and the inverse is defined as

f (j, k) =n−1∑j=0

n−1∑k=0

C(u)C(v)F (u, v)

cos

((2j + 1)uπ

2n

)cos

((2k + 1)vπ

2n

)whereC(w) = 1√

2for w = 0 and 1 for

w = 1,2, . . . , n− 1.


For images that exhibit high pixel to pixelcorrelation, the DCT is indistinguishablefrom the Karhunen–Loeve transform (KLT).

There is a family of DCTs, of which theDCT-II described above is the one commonlyused. These other types of DCT, specificallythe DCT-IV, are sometimes used in calculat-ing fast transforms. TheN = 8 elementDCT is particularly important for image datacompression and is central to the JPEG andMPEG standards. As a matrix, the 8-elementDCT isasin thefigurefor DCT .

discrete data channel the concatenationof all communication system elements be-tween and including the modulator and de-modulator.

discrete Fourier transform (DFT) thesum of complex exponentials representinga sampled sequence. This transform is ob-tained to represent a reasonable approxima-tion of a signal for which only a finite sampleexists. Defined as

X(x) =N−1∑n=0

x(n)e−j (2π/N)nk

wherex(n) represents a sequence of finitesamples of a signal;N is the number of sam-ples in the sequence.

discrete fuzzy set a fuzzy set that includesonly those sample points of a continuous vari-able.

discrete Hadamard transform SeeHada-mard transform.

discrete Hopfield network a single layer,fully connected network that stores (usuallybipolar) patterns by setting its weight valueswij equal to the(i, j) entry in the sum of theouter products of the patterns. The networkcan be used as an associative memory so longas the number of stored patterns is less thanabout 14% of the number of neural elements.Compare withcontinuous Hopfield network.

discrete multipath the result of multi-path propagation observed as clearly separa-ble, discrete signal components, seen in thedelay power spectrum as a set of discretepeaks at various delays.See alsomultipathpropagation, delay power spectrum, specularreflection.

discrete network an electronic networkcomposed of separate, i.e., individual, com-ponents.

discrete sine transform (DST) a unitarytransform mappingN samplesg(n) toN co-efficientsG(k) according to:

G(k) =√

2

N + 1

N−1∑n=0

g(n) sinnkπ

N + 1

with inverse

g(n) =√

2

N + 1

N−1∑k=0

G(k) sinnkπ

N + 1

As with the discrete cosine transform thereis a family of DSTs, the other members ofwhich are rarely used. While the DST isclosely related to the DCT, the latter is theform which has attained supremacy for im-age data compression.

discrete spectrum the eigenvalues of adifferential equation with real coefficientsand finite boundary conditions form a dis-crete spectrum. By extension, also the modesof closed waveguides originate a discretespectrum.

discrete time Fourier series representa-tion of a periodic sequencexn with periodNby the sum of a series of harmonically relatedcomplex exponential sequences:

xn = 1

N

N−1∑k=0

Xkej2πknN .

TheXk are the Fourier series coefficients, ob-tained by

Xk =N−1∑n=0

xne− j2πkn

N .


0.3536 0.3536 0.3536 0.3536 0.3536 0.3536 0.3536 0.35360.4904 0.4157 0.2778 0.0975−0.0975−0.2778−0.4157−0.49040.4619 0.1913−0.1913−0.4619−0.4619−0.1913 0.1913 0.46190.4517−0.0975−0.4904−0.2778 0.2778 0.4904 0.0975−0.45170.3536−0.3536−0.3536 0.3536 0.3536−0.3536−0.3536 0.35360.2778−0.4904 0.0975 0.4157−0.4157−0.0975 0.4904−0.27780.1913−0.4619 0.4619−0.1913−0.1913 0.4619−0.4619 0.19130.0975−0.2778 0.4157−0.4904 0.4904−0.4157 0.2778−0.0975

Discrete cosine transform (DCT).

discrete time signal a signal representedby samples at discrete moments of time (usu-ally regularly spaced). The samples may takevalues from a continuous range, so the termis usually used to differentiate a sampled ana-log signal from a digital signal which is quan-tized. Seesignal.

discrete time system a process that trans-forms discrete time input signals to discretetime output signals.

discrete time white Gaussian noise noisesamples modeled as independent and identi-cally distributed Gaussian random variables.

discrete Walsh-Hadamard transformSeeHadamard transform.

discrete wavelet transform (DWT) acomputation procedure that calculates the co-efficients of the wavelet series expansion fora given finite discrete signal.

discriminator a circuit whose output volt-age varies in magnitude and polarity in directproportion to the difference between the in-put voltage and a standard signal. A discrim-inator that converts phase deviations at theinput to a linearly proportionate variation inoutput voltage is called a phase discriminatorand is used in FM detection.

disk Seemagnetic disk. Also spelled“disc.”

disk actuator a mechanical device thatmoves the disk arms over the disk surface(s)

in order to position the read/write head(s)over the correct disk track.

disk arm a mechanical assembly that po-sitions the head over the correct track forreading or writing a disk device. The armis not movable on a fixed-head disk, but is ona moving-head disk.

disk array a number of disks groupedtogether, acting together as a single logicaldisk. By this, multiple I/O requests can beserviced in parallel, or that the bandwidth ofseveral disks can be harnessed together to ser-vice a single logical I/O request.

disk cache a buffer memory area in mainmemory used to hold blocks of data from diskstorage. The cache can hide much of the ro-tational and seek latencies in disk accessesbecause a complete data block (disk sector)is read or written together. Disk caches arenormally managed by the machine’s operat-ing system, unlike a processor cache, whichis managed by hardware.

disk capacitor a small single-layer ce-ramic capacitor with a dielectric insulatorconsisting of conductively silvered opposingsurfaces.

disk controller unit that carries out theactions required for the proper operations ofa disk unit.

disk drive assembly consisting of elec-tronics and mechanical components, to con-trol disk and disk-head movement and to ex-


change data, control, and status signals withan input/output module, as required for theproper reading or writing of data to or froma disk. The head disk assembly plus all theassociated electronics.

disk format the (system-dependent) man-ner in which a track of a disk is partitionedso as to indicate for each sector: the identity,the start and end, synchronization informa-tion, error-checking information, etc. A diskmust be formatted before any initial writingcan take place.

disk head read/write head used in a diskdrive. Such a head may be fixed-gap, inwhich the head is positioned at a fixed dis-tance from the disk surface; contact-head, inwhich the head is always in contact with thesurface; and aerodynamic, in which the headrests lightly on the surface when the disk ismotionless but floats a small distance abovewhen the tape is rotating. Typically, contactheads are used in floppy disks, and aerody-namic head are used in Winchester disks.

In earlier systems, “fixed-head disks” hav-ing one read/write head per track were usedin some disk systems, so the seek time waseliminated. However, since modern diskshave hundreds of tracks per surface, placinga head at every track is no longer consideredan economical solution.

disk latency time between positioning aread/write head over a track of data and whenthe beginning of the track of data passes un-der the head.

disk operating system (DOS) a set ofprocedures, services, and commands that canbe used by the computer user for managingits resources with a special attention to disksmanaging. The most famous DOS for per-sonal computers is the MS-DOS (Microsoft-DOS). It is a mono-task and mono-user op-erating system.

disk pack a stack of disk platters that canbe removed for off-line storage.

disk platter metal disks covered with amagnetic material for recording information.

disk scheduling algorithms used to re-duce the total mechanical delays in disk ac-cesses, as seen by a queue of simultane-ous I/O requests. E.g., if a “shortest-seek-time-first” scheduling algorithm is used, seektimes can be reduced. That is, among thequeue of pending I/O requests, the one nextserviced is the one requiring the shortestseek time from the current location of theread/write head. The disk scheduling algo-rithm is run by the computer operating sys-tem.

disk sector the smallest unit that can beread or written on a track; adjacent sectorsare separated by a gap. A typical track has10–100 sectors.

disk spindle a stack of disk platters.

disk striping the notion of interleavingdata across multiple disks at a fine grain, sothat needed data can be accessed from all thedisks simultaneously, thus providing muchhigher effective bandwidth.

disk track connectric circle over whicha read/write head moves; adjacent tracks areseparated by a gap. A typical disk has hun-dreds to thousands of tracks.

diskette a floppy disk is a flexible plasticdiskette coated with magnetic material. It isa smaller, simpler, and cheaper form of diskstorage media than the hard disk, and alsoeasily removed for transportation.

disocclusion the uncovering of an object.Seeocclusion.

disparity in binocular vision, the relativedifference of position of an object with re-spect to a background between the left andright images. It is usually measured in min-utes of arc.


dispatch the determination of the poweroutput of each plant in an electric power sys-tem.

dispersed generation problem in eco-nomic dispatch calculations, the task of ac-counting for the production capacity of co-generation or dispersed sources such as pho-tovoltaic or wind generation plants.

dispersion (1) a characteristic of amedium in which the propagation velocityof a wave varies as a function of signal fre-quency.

As a pulse propagates through an opticalfiber, its chromatic components will spreadout or “disperse” in time. This phenomenonlimits the distance between optical regenera-tors in fiber communication systems. Thereare four sources of dispersion: modal disper-sion, material dispersion, waveguide disper-sion, and nonlinear dispersion.

(2) the variation of the index of refractionof a material as a function of wavelength.

dispersion compensator a device thatcompensates for the accumulated chromaticdispersion in a fiber optic transmission sys-tem. Three main schemes exist: fiber gratingdispersion compensators, dispersion com-pensating fiber, and phase conjugation ormid-span spectral inversion.

dispersion diagram a plot of propagationconstant versus frequency.

dispersion of authority a situation inwhich the decisions, for example, decisionsconcerned with the manipulated inputs to thecontrolled process, are distributed betweenseveral (control) decision units; dispersionof authority may result either from natural,legislative, or other reasons, or may appeardue to design of the controller; decentralizedcontrol, in particular, is based on the disper-sion of authority between the local decisionunits. Also known as dispersion of control.

dispersion of control Seedispersion ofauthority.

dispersion shifted fiber single mode opti-cal fibers with zero dispersion in the 1550 nmtelecommunications window. Prior to 1985,single mode optical fibers were designed tohave zero dispersion in the 1310 nm telecom-munications window.

dispersive medium (1) a medium forwhich the permittivity or the permeability (orboth) are frequency dependent.

(2) in optics, medium in which the in-dex of refraction varies significantly with fre-quency.

displaced frame difference (DFD) thedifference between a given digital imageframe and its estimate obtained by using themotion compensation technique. It is usefulin image (sequence) data compression andmotion estimation.

displacement current a field quantity thatdescribes the completion of a circuit when aconducting path is not present.

displacement parameter complex pa-rameter representing the displacement of theamplitude and phase centers of a Gaussianbeam from the axis of an optical system.

display a device that provides a visualnon-permanent display of system input andor output. Common display technology in-cludes CRT (cathode ray tube), LED (lightemitting diode), PDP (plasma display panel),EL (electroluminescense) and LCD (liquidcrystal display).

dissipated power the power dissipated asa heat, which is defined by subtracting an RFoutput power from a DC input power.

dissipation the phenomenon associatedwith the attenuation of a propagating wave ina medium with material losses.


dissipation factor (DF) the ratio of theeffective series resistance of a capacitor to itsreactance at a specified frequency measuredin percentage. Also known as loss tangent.

dissipation power (1) ratio of real power(in phase power) to reaction power (shifted90◦ out of phase).

(2) the ratio of the imaginary to real partsof the complex permittivity, expressed as adimensionless ratio.

dissipative half-space used for analy-sis of complex systems, it is constructed bydividing the infinite space in two by someconvenient fashion. The resulting two half-spaces can both be infinite, or one finite andone infinite. The resulting half-space is thenfilled with a dissipative material (other thana perfect insulator or conductor).

dissipator a form of air terminal which ismeant to prevent lightning strikes to a struc-ture by reducing the surface charge on theearths’s surface in the immediate region ofthe structure.

distance Seechamfer distance,chessboarddistance, Euclidean distance, Hammingdistance, Hausdorff distance, inter-featuredistance, Mahalanobis distance, Manhattandistance.

distance between symbol strings a mea-sure of the difference between two symbolstrings. The most frequently used distancemeasures between two strings include Ham-ming distance, edit or Levenshtein distance,and maximum posterior probability distance.See alsoHamming distance, edit distance,maximum posterior probability distance.

distance measure a functiond(x, y) de-fined on a metric space, such thatd(x, y) ≥ 0, where = holds iffx = y,d(x, y) = d(y, x), andd(x, y) ≤ d(s, z)+ d(z, y).

See alsosimilarity measure.

distance profile for convolutional codes,the minimum Hamming weight of all se-quences of a specific length emerging fromthe zero state. A distance profile for one codeis superior to that of another if

1. all values of the distance profiles of thetwo codes up to a certain depthp (lower thanthe constraint length) are equal; and

2. the superior distance profile code hashigher values of the distance profile for alldepths above the given depthp.

distance protection relaying principlebased upon estimating fault location (dis-tance) and providing a response based uponthe distance to the fault.

distance relay Seeimpedance relay.

distance resolution delay resolutionmapped into the spatial domain, measured inunits of distance (meters). Each unit of timedelay corresponds to a unit of distanced trav-eled by the radio wave through the equation

d = cτ,

whereτ is the delay andc is the speed oflight. Distance resolution gives the smallestdifference in path length resolved by a signal.Is also a measure of the spatial resolution ca-pability of a signal (or measurement system).

distance transform a map of all the pixelsin a shape showing the closest distance ofeach point in the shape from the background;also, an image in which the distance maps ofall the shapes in the image are indicated.Seealsodistance.

distortion (1) addition of an unwantedcomponent to an electronic signal.

(2) undesired change in a signal’s ampli-tude and/or phase as the result of it passingthrough an active nonlinear circuit. Numer-ous figures of merit have been adopted to de-scribe various aspects of signal distortion.

distortion cross modulation nonlineardistortion of a system or carrier signal char-


acterized by the appearance in the outputof frequencies equal to the sum and differ-ence of the desired carrier’s frequency andthe unwanted cross modulating carrier’s fre-quency. Although harmonic components arealso caused to be present, they are not usuallya part of measurements of this effect.

distortion factor See total harmonic dis-tortion.

distortion-rate theory Seerate-distortiontheory.

distributed amplifier (1) composed oftwo main artificial transmission-line sectionsconsisting of series inductors and shunt ca-pacitors, which are usually supplied by theFET transistor. Excellent bandwidth per-formers are obtainable and the amplifier canbe designed as wideband low-noise ampli-fiers and are relatively easy to simulate andfabricate.

(2) the input and output capacitance ofthe active devices can be absorbed into dis-tributed circuits (i.e., transmission lines) toobtain a very broad bandwidth.

distributed antenna typically consists ofa set of discrete radiators fed by a commoncable from a single signal source.

distributed arbitration a scheme usedfor bus arbitration where multiple bus mas-ters can access the bus. Arbitration is notdone centrally (by a bus arbiter), but insteaddone in a distributed fashion. A mechanismto detect when more than one master tries totransmit on the bus is included. When thishappens, one (or all) stops transmitting andwill reattempt the transmission after a short(e.g., random) time delay.

Compare withcentralized arbitration.

distributed computing an environmentin which multiple computers are networkedtogether and the resources from more thanone computer are available to a user.Seealsodistributed computing environment.

distributed computing environment (DCE)an industry-standard, comprehensive, and

integrated set of services that supports thedevelopment, use, and maintenance of dis-tributed computing technologies. DCE is in-dependent of the operating systems and net-work types. It provides interoperability andportability across heterogeneous platforms,and provides security services to protect andcontrol access to data. DCE also providesservices that make it easy to find distributedresources; for instance, directory service, aDCE component, is a central repository forinformation about resources in a distributedsystem. Distributed time service (DTS), an-other DCE component, provides a way tosynchronize the times on different hosts ina distributed system. DCE gives a model fororganizing widely scattered users, services,and data. It runs on all major computing plat-forms and is designed to support distributedapplications in heterogeneous hardware andsoftware environments. Particularly impor-tant for the World Wide Web and security ofdistributed objects.

distributed computing system a systemwhose different parts can run on differentprocessors.

distributed control a control techniquewhereby portions of a single control processare located in two or more places.

distributed element a circuit element inwhich dimension is not negligible relative tothe wavelength. The characteristics of a dis-tributed element depend upon a dimensionsuch as a length.

distributed feedback laser (DFB) a lasersource where the optical feedback is dis-tributed throughout the length of the gainmedium. Feedback then occurs throughBragg diffraction, and the laser operates inonly one optical mode.


distributed generation small powerplants at or near loads and scattered through-out the service area.

distributed memory denotes a multipro-cessor system where main memory is dis-tributed across the processors, as opposedto being equally accessible to all. Eachprocessor has its own local main memory(positioned physically “close”), and accessto other processors’ memory takes placethrough passing of messages over a bus. Theterm “loosely coupled” could also be used todescribe this type of multiprocessor architec-ture to contrast it from shared-memory archi-tectures, which could be denoted as “stronglycoupled.”

distributed memory architecture a mul-tiprocessor architecture in which physicalmemory is distributed among the processingnodes, as opposed to being in a central loca-tion, equidistant from all processors.

distributed refresh in a DRAM, carryingout refresh operations one at a time, at regularintervals. Requires that all rows be refreshedin a time less that the time before which anygiven row needs to be refreshed.See alsoburst refresh.

distributed sample scrambling a modi-fication of the reset scrambling technique inwhich information regarding the state of thescrambling sequence generator is embeddedinto the encoded sequence in a distributedfashion for purposes of synchronizing the de-scrambling sequence generator.

distribution (1) the possibility to executedifferent parts of a system on different pro-cessors.

(2) that class of electric power systemwork which is concerned with the distribu-tion of electric power within a load area suchas a residential or commercial area, or withinan industrial installation. The distributioncircuit extends from the local substation andterminates at the customer’s meter.

distribution function See cumulativedistribution function.

distribution management system (DMS)a system that helps manage the status of thedistribution network, crews and their workflow, system safety, and abnormal conditions.

distribution switchboard a switchboardused in the distribution system, typicallywithin a building.

distribution transformer a transformerdesigned for use on a power distribution sys-tem (typically 2.4 kV to 34.5 kV) to supplyelectrical power to a load at the proper uti-lization voltage.

disturbance a sudden change or a se-quence of changes in the components or theformation of a power system. Also calledfault.

disturbance decoupling of generalized 2-Dlinear systems given the second general-ized 2-D Fornasini–Marchesini model withdisturbances

Exi+1,j+1 = A1xi+1,j + A2xi,j+1

+ B1ui+1,j + B2ui,j+1

+H1zi+1,j +H2zi,j+1

yij = Cxij

i, j ∈ Z+ (the set of nonnegative integers),find a state-feedbackuij = Kxij such thatthe outputyij is not affected by the distur-bances fori, j ∈ Z+ wherexij ∈ Rn, uij ∈Rm, yij ∈ Rp, zij ∈ Rq are semistate vec-tor, input vector, output vector, and distur-bance vector, respectively, andE, Ak, Bk,Hk (k = 1,2, ) are real matrices withE pos-sibly singular or rectangular.

dithering computer technique allowingthe display and printing of gray-level imageson devices having a small number of avail-able colors (generally two, as in black-and-white CRTs and printers). The two main ap-proaches are matrix dithering, where a matrix


of black and white dots is associated witheach gray-level, and error-diffusion dither-ing, where each gray-level pixel in turn isassigned an available color, and the error isspread to its unprocessed neighbors (as in theFloyd-Steinberg) method.Seehalftone.

divergence the angle that the trajectory ofeach particle makes with the beam axis. Ac-celerator systems always try to reduce beamdivergence.

divergence theorem let us consider a vol-umeV bounded by a surfaceS and an outernormaln. Let alsoF be a vector functionwhich, together with its partial derivatives,is continuous at all points ofV andS. Thedivergence theorem states that∫

V

divFdx =∫S

F · ndS .

The above equation may also be used to de-fine the divergence.

diversity combining a communicationtechnique that combines the signals re-ceived through different, possibly indepen-dent channels. The most common methodsof combining are maximum-ratio combining,equal gain combining, selection combining,and switched combining.

diversity frequency a method for in-creasing the reliability of digital communi-cations in which multiple copies of the sig-nal, or other types of redundant informa-tion, are transmitted. Frequency diversityimplies that the received signal occupies amuch wider bandwidth than the minimumbandwidth needed to carry the information.

diversity path a form of diversity in whichmultiple copies of the signal are created viadifferent paths from the transmitter to re-ceiver.

diversity selection a form of diversityreception in which the receiver selects the

strongest signal among the copies received.The weaker signals are simply ignored.

divide by zero error occurring when adivision per zero is operated. This case ismathematically undefined. In many cases,this problem is detected directly at level ofmicroprocessor that activates an exceptionand leaving true a status flag. The excep-tion can be managed for recovering the errorand avoiding the interruption of the programexecution. Also known as divide per zero.

divide per zero Seedivide by zero.

divider functional unit consisting ofcircuits that implement either integer orfloating-point division.

DLL Seedelay-locked loop.

DLT directory look-aside table.See alsotranslation lookaside buffer.

DMA Seedirect memory access.

DMS Seedistribution management system.

DOA direction of arrival.See alsoangleof arrival.

DOF Seedepth of focus.

domain module or area of execution thatis to be kept isolated from other domains; adomain may have special properties that de-fine the nature of the isolation and the limitson communication to and from the domain.In the context of secure system design, a do-main may be an execution of a process thathas specific security attributes.

dominant mode the mode of a waveguidehaving the lowest cutoff frequency or of acavity having the lowest resonant frequency.

don’t care a function that can be takeneither as a minterm or a maxterm at the con-venience of the user.


donor an impurity in a semiconductorthat donates a free electron to the conductionband.

door Seelid.

dopant an impurity substance (such asphosphorus or boron) added in very smallcontrolled quantities to a semiconductor basematerial (such as silicon or gallium arsenide),thereby changing the material conductioncharacteristics by modifying electron and/orhole concentration. A donor dopant is onethat gives rise to electrons and an acceptordopant is one that gives rise to holes.

doping the process of introducing impu-rity atoms into pure silicon to change its elec-trical properties. The impurities may be ei-ther n-type (introducing an additional con-ducting electron) or p-type (introducing theabsence of a conducting electron, also calleda “hole”).

Doppler broadened lineshape functionspectral function that results from Dopplershifts caused by the velocity distribution ofatoms or molecules in a gas; a Gaussian func-tion for a Maxwellian velocity distribution.

Doppler broadening broadening of aspectral line due to Doppler shifts caused bythe random motion of atoms or molecules ina vapor.

Doppler effect SeeDoppler shift.

Doppler filter a filter used to resolve tar-gets from each other and from extraneous re-turns from other objects (called clutter) byfiltering in the velocity or Doppler domain.So-called because the Doppler effect causesfrequency shifts proportional to velocity vari-ations in tracked objects.

Doppler frequency a shift in frequency ofthe returned power from a target as a result ofthe target’s motion relative to the illuminatingsource.

Doppler linewidth characteristic widthof a Doppler-broadened spectral line; usuallythe full width at half maximum when the lineis Gaussian.

Doppler power spectrum a functioncharacterizing the spread of average receivedpower as a function of Doppler shift. Can beobtained from the scattering function by in-tegrating over the delay variable.See alsoscattering function, multipath propagation.

Doppler radar radar-based techniqueused in measuring the velocity of a movingtarget or wind by measuring the Doppler shift(Doppler effect).

Doppler shift a frequency shift in a re-ceived signal caused by time-variant trans-mission delay, or equivalently time-variantpropagation path length. This in turn iscaused by movement of terminals with re-spect to each other, or by movement of re-flecting objects. In optics, frequency shiftsimposed on laser beams such as when usedin laser radar or when diffracted by an acous-tic wave.

The Doppler shift depends on the fre-quency of the signal and the angle of arrival ofthe signal relative to the direction of move-ment of the receiver. For a signal consist-ing of a range of frequencies, each frequencycomponent will experience a different shift.Hence the received signal will have a dif-ferent bandwidth than the transmitted signal(Doppler dispersion).

For a tone (continuous-wave) signal of fre-quencyf , the Doppler shiftfD observed ona single propagation path of changing lengthis given by

fD = f v

c,

wherev is the rate (in m/s) of path lengthchange, andc is the speed of light.

Also known as Doppler effect.

Doppler spread the increase of band-width of a signal due to doppler shifting of


multipath energy arriving at a mobile receiverfrom differing directions.

Doppler-free spectroscopy a spectro-scopic technique in which two or more lightfields are used to produce a spectral line thatis not Doppler shifted or Doppler broadened.For example, in the case of lambda systemsor vee systems, this requires that the elec-tromagnetic fields co-propagate, whereas forcascade systems, the fields must counter-propagate.

DOS Seedisk operating system.

dose to clear the amount of exposureenergy required to just clear the resist in alarge clear area for a given process.

dosimeter an instrument used for mea-suring or evaluating the absorbed dose of ra-diation. It may depend on the measurementof ionization for its operation or may simplyinvolve the darkening of a piece of photo-graphic film ("film badge").

dot pitch the center-to-center distance be-tween adjacent green phosphor dots of thered, green, blue triad in a color display.

dot-matrix printer a printer that pro-duces readable characters by imprinting alarge number of very small dots.

dots per inch (DPI) a measure of the den-sity of line-printer plots in dots per inch.

double bridge a Wheatstone bridge mod-ification designed to increase the precisionof measurements for low-value resistors. Toavoid the error due to resistance of the con-nection (called yoke) between the unknownresistor and the standard resistor, the ends ofthe unknown resistor and the standard resis-tor are connected to the balance detector (usu-ally a galvanometer for this type of bridge)via two small resistors, the ratio of which isthe same as the ratio of resistors in the ratioarms. Then the yoke resistance is eliminated

from the balance condition and the unknownresistor can be found using the same formulaas in an ordinary Wheatstone bridge.

double buffering (1) generally, anybuffering scheme where two identical mem-ory images are used. One memory imageis used to hold received data, while data issimultaneously read from the other memoryimage.

(2) in terminal-to-computer communica-tion scheme where a number of remote ter-minals are connected to a single computerthrough a “multiplexer.” This unit connectsn low-speed bit-serial transmission lines ontoa single high-speed bit-serial line (runningntimes faster) using STDM (synchronous timedivision multiplex) of the connection. In themultiplexer, each low-speed line is connectedto a “one-character buffer,” converting the re-ceived low-speed bitstream to a high-speedbitstream using two shift-registers (buffers).The first one receives the low-speed line char-acter bits (8 bits), clocked by the low-speed“receive clock.” When a complete charac-ter has been received, it is moved to thesecond shift-register, where it is stored un-til it can be shifted out on the high-speedline on the appropriate time slot. Mean-while, the next character is assembled in thefirst shift-register. For full-duplex operation(simultaneous two-way communication), asimilar structure is needed for the opposite(computer-to-terminal) connection.

double conversion the process where anincoming RF signal is mixed with a local os-cillator (LO) signal to produce the first inter-mediate frequency (IF). This IF is then mixedwith a second fixed LO to produce a secondIF signal.

double conversion receiver a receiverthat uses two heterodyne operations be-fore detection generating two intermedi-ate frequencies, first intermediate frequency(IF) and second IF, respectively.See alsointermediate frequency.


double heterostructure dye laser laser inwhich the amplification takes place in a dye;a broadly tunable laser.

double lambda system a quantum me-chanical system composed of two interact-ing lambda systems that share common lowerstates. A double lambda system is also a dou-ble vee system. Double lambda systems canbe used to construct resonant closed loop in-teractions in which some linear combinationof the phases of the four fields is constant intime.

double line to ground fault Seedoublephase ground fault.

double phase ground fault a fault withtwo transmission lines being connected to theground.

double revolving-field theory Seecounter-rotating field theory.

double sideband modulation (DSB) amodulation scheme resulting in a spectrumconsisting of the carrier frequency, one signalthat is the sum of the carrier and the modu-lating signal, and one signal that is the differ-ence between the carrier and the modulatingsignal.

double tuned a circuit, amplifier, or otherdevice having a response that is the same astwo single-tuned circuits.

double word data block that containstwice the number of bits as the machine wordsize in a microprocessor.

double-cage rotor Seedual-cage rotor.

double-frequency recording Seemagneticrecording code.

double-line contingency a malfunctionof a power system which involves the simul-taneous failure of two transmission lines.

double-sided assembly a packaging andinterconnecting structure with componentsmounted on both the primary and secondarysides.

double-sided disk a disk in which bothsides of a platter are covered with magneticmaterial and used for storing information.

doublet a system of two quadrupole mag-nets in close proximity and with oppositepolarity used to simultaneously constrainthe beam size in two dimensions at somepoint downstream. Doublet (Quadrupole)is a beam optical system consisting of twoquadrupoles of opposite sign, which providesnet particle focusing in all planes.

doubly-fed induction motor an induc-tion motor with a wound-rotor. The rotorand stator windings are connected to sepa-rate sources of electric energy. The machinecan be used as a generator to provide fre-quency conversion or precise speed controlin the motor-mode.

down-sampling Seedecimation.

downconductor the cable which connectsan air terminal or lightning rod to ground.Seeair terminal.

download to bring data from a remotesource to local storage.

downstream in power distribution work,the direction in which power flows, i.e., to-wards the load.

DPA Seedestructive physical analysis.

DPCM See differential pulse codemodulation.

DPCM encoding Seedifferential pulsecode modulation.

DPI Seedots per inch.

DPLL Seedual phase-locked loop.


DQE Seedetective quantum efficiency.

DRA Seedielectric resonator antenna.

drain terminal of an FET (usually identi-cal in structure to the source) to which elec-trons flow. Electrons in the FET channel flowdown the drain, and current flow is defined asthe negative direction of electron movement,since electrons are negative. In p-channelFETs, current flows from source to drain. Inn-channel FETs, current flows from drain tosource. The drain is usually considered to bethe metal contact at the surface of the die.

drain conductance the increase in draincurrent when the magnitude of the appliedFET drain-to-source voltage is increased.Mathematically, the derivative of drain cur-rent with respect to drain voltage.

drain saturation current the drain-to-source current flow through the JFET underthe conditions thatVGS = 0 and| VDS |>|VP | such that the JFET is operating in theactive or saturated region.

drain-source leakage the current flowingin the channel of a MOSFET when its gateand source are shorted together. The mag-nitude of the leakage current is strongly in-fluenced by the applied drain-source voltage,the gate length and substrate doping concen-tration.

drain-to-source voltage (VDS) poten-tial difference between the FET drain andsource terminals, this voltage determines thedevice operational region and limits the out-put power. For an n-channel device this volt-age is normally positive, and negative for ap-channel device. Magnitudes usually rangeup to as high as 10 V for a low noise deviceand much higher for power devices.

DRAM See dynamic random accessmemory.

drift (1) movement of free carriers in asemiconductor due to the electric field.

(2) the relatively uninpaired fluctuationof adaptive filter coefficients in the direc-tion of least sensitivity along the eigenvectorcorresponding to the minimum eigenvalue isknown as coefficient drift.

Coefficient drift can be a problem if thiseigenvalue gets very small, as coefficientscan drift out of the allowed region. Alsoimportant as random, often temperature-induced, fluctuations in the output levels ofDC amplifiers.

drift chamber a series of chambers usedto detect particle trajectories. They are simi-lar to multi-wire proportional chambers, ex-cept the wire spacing is increased. The cor-relation between the position of an ionizedtrack produced by a charged particle and thetime of appearance of an electric pulse at thewire is used to measure the distance of thetrajectory from the wires.

drift space space where electrons movedue only to their inertia.

drip-proof machine a machine with ven-tilating openings constructed in such a waythat drops of liquid or solid particles fallingon it, at an angle less than 15 degrees fromthe vertical, can enter the machine neitherdirectly nor by striking on it, run along a hor-izontal or inwardly inclined surface.

drive circuit a circuit that produces gatetrigger pulses, of desired level and timing, toturn on and off active switches (just turn onfor natural-commutated switches) in switch-ing circuits.

driving-point admittance the admittancemeasured at the antenna terminals when theantenna is in free space (not loaded).

DRO Seedielectric resonator (stabled)oscillator.


dropout equipment misoperation due toan interruption, noise, or sag.

dropout compensation signal compen-sation provided when a loss of informationfrom magnetic tape defects occurs duringVCR tape playback. Signal dropouts resultfrom oxide or dust interfering with the tape-head to tape contact when the VCR is inthe record or the playback operational mode.A horizontal delay line stores the luminancesignal of the previous line that is used to re-place the lost video information. The sig-nal dropout is detected as a loss of the tapeplayback. FM signal activates an electronicswitch. The switch substitutes the stored de-lay line contents into the video path to com-pensate for the lost video from the tape play-back.

dropout current the current at which amagnetically-operated device will revert toits de-energized position.

dropout voltage the voltage level whereproper equipment operation is hindered.

Drude material a frequency-dependentdielectric whose complex permittivity is de-scribed by an equation with two poles, one ofthem atw = 0. A collisional plasma is an ex-ample of such a material duration–bandwidthreciprocity relation.

Drude media SeeDrude material.

drum memory an old form of backingmemory. Similar to magnetic disks in opera-tion, but here the magnetic film is depositedon the surface of a drum, instead of a disk.

dry-type transformer a transformer thatis cooled by circulating air or gas through oraround the transformer housing.

DS-CDMA direct sequence code divisionmultiple access.See alsodirect sequenceandcode division multiple access.

DSB Seedouble sideband modulation.

DSI Seediscrete sine transform.

DSL Seedigital subscriber line.

DSP Seedigital signal processor.

DTE Seedata terminal equipment.

DTW Seedynamic time warping.

dual control an adaptive control that per-forms two different functions at the sametime: one that applies probing signals tolearn more about the system dynamics andthe other that tries to keep the output at a de-sired value.

dual functions mathematical relationshipbetween two sets used to represent a signal. Ifa signals(t) can be represented by a completeset {πn(t)}, i.e., s(t) =

∑n

anπ(t)n, there

must exist a dual set{πn(t)} such that theexpansion coefficientsan =< s(t), πn(t) >.πn(t) andπn(t) are called dual functions. Asa special case, if the set{πn(t)} is orthonor-mal, then it is its own dual.

dual in-line package (DIP) a standardcase for packaging integrated circuits. Thepackage terminates in two straight, paral-lel rows of pins or lead wires. This stan-dard has been more recently substituted bysurface-mount standards.Compare withsingle in-line packaging.

dual mode filters filters realized by us-ing two resonances inside the same cavity,hence allowing saving of space, volume, andweight.

dual phase-locked loop (DPLL) pro-grammable, low-jitter, low-power, and high-performance devices. DPPLs are capableof synthesizing two low-jitter clocks withuser-selected, industry-standard frequencies,phased-locked to the system reference tim-


ing. They accept a wide range of popular tele-com and networking input frequencies andcan be programmed to generate a range ofoutput frequencies. Used for wide area net-work (WAN) and ISDN applications.

dual port memory memory system thathas two access paths; one path is usually usedby the CPU and the other by I/O devices.

dual-cage rotor a three-phase inductionmotor rotor with two separate squirrel cagewindings, that give the effect of varying ro-tor resistance. The outer cage has high re-sistance to obtain high starting torque, whilethe inner cage has low resistance to reducelosses at full load.

dual-element fuse a fuse constructed withtwo different types of fusible elements in se-ries. One element is designed to melt veryquickly in the presence of fault current. Theother is designed to melt after a time delaywhen exposed to overload conditions. Thefusible elements are somewhat similar in op-eration to the thermal and magnetic elementsof an inverse-time circuit breaker. Dual-element and time-delay are often used inter-changeably.

duality property of Fourier transformthe property that results from the symmetrybetween the Fourier transform synthesis andanalysis equations. To illustrate, letf (w)be the Fourier transform of the time func-tion g(t) (f (w) = F {g(t)}). Substitutingt for w in f gives the time functionf (t).Applying a Fourier transform tof (t) givesthe function 2πg(−w), a frequency domainfunction similar to the original time functiong(t) (g(−w) = 1

2π F {f (t)}).

DUF Seediffusion under field.

dump an area of steel and dense concreteinto which unwanted particle beam can besteered so that its energy can be dissipatedin a safe and controlled manner. The dumpresistor is switched into the magnet/capacitor

circuit in order to dissipate the stored energyin the magnets/capacitors.

duoplasmatron a type of ion-producingsource that develops protons by extractingpositive ions from an arc struck in hydrogengas.

duplex a method of winding the arma-ture of a commutated electric machine suchthat the number of parallel electrical pathsbetween brushes is double that provided by asimplex winding. Duplex windings are con-structed by placing consecutive coils in al-ternate coil slots and continuing the windingtwice around the rotor, filling the empty slotson the second pass. The result is two com-plete, identical windings between brush posi-tions rather than the one winding that is pro-duced when coils are placed in adjacent slots.See alsosimplex, multiplex, reentrancy.

duplex channel two-way simultaneous(and independent) data communication, e.g.,between a computer and remote terminals.

duplex ultrasound simultaneous displayof speed versus time for a chosen region andthe two-dimensional B-mode image.

duplication with complementary logicfault detection based on circuit duplicationand comparison. One module is designed us-ing positive logic and the other module usesnegative logic. This assures detecting com-mon mode faults.

dust cover a cover to protect the terminalsof a pad-mount transformer.

dust-ignition-proof machine a machinedesigned with a casing or specialized en-closure to safely contain any internal ig-nition or flammable substances or compo-nents, and prevent them from igniting exter-nal flammables such as explosive gases, va-pors, and dust particles.

DUT device under test.


duty cycle (1) the ratio of the turn-on timeof a semiconductor switch to the sum of theturn-on and turn-off times;

Turn-on / (Turn-on + Turn-off)

(2) the mode of operation that an elec-tric machine is classified, in consideration ofthermal limits, e.g., continuous operation, in-termittent operation.

Also known as duty ratio.

duty ratio Seeduty cycle.

DVM Seedigital voltmeter.

DWT Seediscrete wavelet transform.

DX-center stands for defect unknown(X)/complex. Name originally given to de-fects that caused persistent photoconductiv-ity (PPC) in GaAs. The term is now com-monly applied to similar defects in bothIII-V and II-VI compound semiconductors.The presently accepted interpretation of themicroscopic nature of the DX-center is alarge deformation associated with a nega-tively charged substitutional impurity.

dyadic decomposition wavelet decompo-sition in which the dilation is a dyadic num-ber, i.e., 2j with j being the decompositionlevel; this corresponds to a octave frequencyband division in the frequency domain.

dyadic Green’s functions Green’s func-tions that relate together vector field quanti-ties.

dyed resist a photoresist with an addednonphotosensitive chemical that absorbslight at the exposing wavelength.

dynamic allocation allocation of memoryspace that is determined during program ex-ecution. Dynamic allocation can be used todesignate space on the stack to store objectswhose lifetimes match the execution interval

of a subroutine; this allocation is performedupon entry to the subroutine.

dynamic beam control a circuit in acamera designed to instantaneously increasebeam current to handle highlights in a scene.

dynamic brake the braking operation ofa machine by extracting electrical energy andthen dissipating it in a resistor.

dynamic braking since AC motors do nothave separate field excitation, dynamic brak-ing is accomplished by continuing to excitethe motor from the drive. This causes a gen-erative current to the drives DC intermediatebus circuit. The dynamic brake resistors arethen placed across the DC bus to dissipate thepower returned. The brake resistor is usuallyswitched by a transistor or other power switchcontrolled by the drive.

dynamic channel assignment (DCA) atechnique of assigning or reassigning radiochannels in a communications system in or-der to respond to current interference or prop-agation conditions. Such assignment or reas-signment may be made before or during radiotransmission. DCA is typically implementedin a distributed fashion, with mobile termi-nals and/or fixed cell sites taking measure-ments to determine local interference levels,and then applying an algorithm to choose thebest radio channel on which to initiate or re-assign communication.

dynamic convergence the process ofmaking the electron beam fall on a specifiedsurface during scanning, as in the case of aCRT. This process is necessary for all three-gun CRTs that do not have in-line video dis-plays or CRTs.

dynamic linking deferring the determi-nation of the association between a symbolused in a program module and the object towhich it refers until that object must be ac-cessed (i.e., during program execution).


dynamic load line graphical plot showingthe instantaneous relationship between volt-age across and current through a transistorwhen driven by an input AC signal.

dynamic matching See variationalsimilarity.

dynamic memory allocation the run-time assignment of small units of memory toan active program. Used typically to supportgrowing structures such as lists.

dynamic path reconnect used in IBM’shigh-end computer systems to allow a “sub-channel” to change its channel path each timeit cycles through a disconnect/reconnect witha given device. This enables it to be assignedto another available path, rather than just waitfor the currently allocated path to becomefree.

dynamic programming introduced byBellman, one of the best known methods forsolving the optimal control problems. A re-cursive method to compute the optimal con-trol as a function of the state, dynamic pro-gramming is used in multistage systems byworking backward from the final stage.

Dynamic programming is based on TheOptimality Principle. The Principle says thatoptimal control strategy has the feature thatregardless of initial state and initial decision,decision in the next step must form an opti-mal control strategy with respect to the finalstate of the previous decisions. This principleallows us to find an optimal strategy in a nu-merical way. The principle serves to limit thenumber of potentially optimal control strate-gies that must be investigated. It also impliesthat optimal control strategies must be deter-mined by working backward from the finalstage.

dynamic random access memory (DRAM)a semiconductor memory using one capac-

itor and one access transistor per cell (bit).The information is stored dynamically on asmall charge on the cell capacitance, and can

be read or written through the “access tran-sistor” in the cell. Since the charge willslowly leak away (through semiconductorjunctions), the cells need to be “refreshed”once every few milliseconds. This is typi-cally done using on-chip circuitry. DRAMshave very high storage density, but are slowerthan SRAMs (static RAMs).See alsoburstrefresh, distributed refresh.

dynamic range refers to the range of inputsignal amplitudes over which an electronicdevice will operate within a set of specifiedparameters. Usually expressed in decibels.In a communications receiver, the upper endof the dynamic range is determined by thelargest tolerable input signal, while the lowerend is set by the receiver’s sensitivity. Thesensitivity is the minimum discernible signalfor a specific signal-to-noise ratio (SNR).Seealsosignal-to-noise ratio.

dynamic reconfiguration changes of op-tical paths from sources to detectors whichare instantly controllable. Paths of opticalsignals are controlled and changed by an op-tical crossbar switch that is usually a spatiallight modulator.

dynamic scattering procedure to studythe change of state of atoms and molecules byanalyzing the frequency shift and fluctuationsof scattered light.

dynamic scheduling (1) creating the ex-ecution schedule of instructions at run-timeby the hardware, which provides a differentschedule than strict program order (i.e., a pro-cessor issues instructions to functional unitsout of program order). The processor can dy-namically issue an instruction as soon as allits operands are available and the required ex-ecution unit is not busy. Thus, an instructionis not delayed by a stalled previous instruc-tion unless it needs the results of that previousinstruction.

(2) changing the software program sched-ule dynamically depending on data or oper-ating conditions.


(3) automatic adjustment of the multipro-cessing program at run time that reflects theactual number of CPUs available presently.For instance, a DO loop with 100 itera-tions is automatically scheduled as 2 blockswith 50 iterations on a two-processor sys-tem, as 10 blocks with 10 iterations on aten-processor system, and as one block on asingle-processor machine. This enables oneto run multiprocessor programs on single-processor computers.

dynamic simulation Seedirect dynamics.

dynamic stability a measure of a powersystem to return to a pre-disturbance steady-state condition following a disturbance.

dynamic system Seestatic system.

dynamic time division multiple access(D-TDMA) time division multiple accessscheme in which the channels are assigneddynamically.See alsotime division multipleaccess.

dynamic time warping in problems oftemporal pattern recognition, each exemplarcan be regarded as a sequence of vectors.The process of pattern matching requires tocarry out an optimal alignment of the vectorscomposing the sequences so as to minimizea proper distance. For example, in automaticspeech recognition, the problem of isolatedword recognition requires producing an op-timal alignment between the incoming wordto be classified and a reference template. Let

A = [a1, . . . ,aM ]

B = [b1, . . . ,bN]

be two sequences of vectors(ai,bi ∈ Rp)

that must be aligned optimally. Formally, de-termining the optimal alignment consists offinding a warping function

C = c1, . . . , cK

wherec(k) = [ik, jk], such that the distance

d(A,B) = .=K∑k=1

(ai(k) − bj(k) )2

is minimum. The optimization must takeplace under the following conditions:1. monotonic conditioni(k) ≥ i(k − 1) andj (k) ≥ j (k − 1)2. boundary conditionsi(1) = k(1) = 1i(K) = M

j(K) = N

3.non-skip conditioni(k)− i(k−1) ≤ 1 andj (k)− j (k − 1) ≤ 14. efficiency condition|i(k)− j (k)| < Q

The solution of this problem can be ob-tained by Belmann’s dynamic programming.The algorithm that produces the optimal tem-plate alignment is referred to as dynamicaltime warping (DTW).

dynamic time warping (DTW) a recog-nition technique based on nonlinear timealignment of unknown utterances with ref-erence templates.

dynamical linear nonstationarycontinuous-time finite-dimensional sys-tem a system described by the linearordinary differential state-equation

x′(t) = A(t)x(t)+ B(t)u(t)

and the linear algebraic output equation

y(t) = C(t)x(t)+D(t)u(t)

wherex(t) ∈ Rn

is the state vector,

u(t) ∈ Rm

is the input vector,

y(t) ∈ Rq


is the output vector.

u ∈ L2loc([t0,∞), Rm)

is an admissible control,A(t) isn×n dimen-sional matrix, with piecewise-continuous el-ements,B(t) is n × m dimensional matrix,with piecewise-continuous elements,C(t) isq × n dimensional matrix with piecewise-continuous elements,D(t) is q × m dimen-sional matrix with piecewise-continuous ele-ments. The solution of the state equation hasthe form

x(t, x(t0), u) = F(t, t0)x(t0)

+∫ t

t0

F(t, s)B(s)u(s)ds

whereF(t, s) isn×n dimensional transitionmatrix for a dynamical system.

dynamical linear nonstationary discrete-time finite-dimensional system a systemdescribed by the linear difference state equa-tion

x(k + 1) = A(k)x(k)+ B(k)u(k) (1)


y(k) = C(k)x(k)+D(k)u(k)

wherex(k) ∈ Rn is the state vector,u(k) ∈Rm is a control vector,y(k) ∈ Rq is an out-put vector, andA(k), B(k), C(k), andD(k)are matrices of appropriate dimensions withvariable coefficients. Solution of the differ-ence state equation (1) has the form

x(k, x(k0), u) = F(k, k0)x(k0)

+j=k−1∑j=k0

F(k, j + 1)B(j)u(j)

whereF(k, j) isn×n dimensional transitionmatrix defined for all

k ≥ j

in the following manner:

F(k, k) = Inxn for k ∈ ZF(k, j) = F(k, j + 1)A(j)

= A(k − 1)A(k − 2) . . . A(j + 1)A(j)

for k > j .

dynamical linear stationary continuous-time finite-dimensional system a sys-tem described by the linear differential state-equation

x′(t) = Ax(t)+ Bu(t) (1)


y(t) = Cx(t)+Du(t)

wherex(t) ∈ Rn

is the state vector,

u(t) ∈ Rm

is the input vector,

y(t) ∈ Rq

is the output vector,A, B, C, andD areconstant matrices of appropriate dimensions.The transition matrix of (1) has the formF(t, s) = eA(t−s).

dynamical linear stationary discrete-timefinite-dimensional system a system de-scribed by the linear difference state equation

x(k + 1) = Ax(k)+ Bu(k) (1)


y(k) = Cx(k)+Du(k)

wherex(k) ∈ Rn is a state vector,u(k) ∈ Rmis a control vector,y(k) ∈ Rq is an outputvector, andA, B, C, andD are constant ma-trices of appropriate dimensions. The tran-sition matrix of (1) has the formF(k, j) =Ak−j .


dynamical systems with delays a systemdescribed by the linear state equation

x′(t) = A0x(t)+ A1x(t − h)+ Bu(t) (1)

wherex(t) is n-dimensional vector,u(t) ism-dimensional control vector, andA0, A1,andB are constant matrices of appropriate di-mensions andh > 0 is a constant delay. Fora given admissible control and initial data,the above differential equation (1) with devi-ating argument has a unique solution derivedby the method of steps.

The state space for dynamical sys-tem (1),X = W

(2)1 ([−h,0], Rn), is infinite-

dimensional Sobolev space of absolutelycontinuous functions defined on[−h,0] withvalues in Rn and with square integrablederivatives.

Linear unbounded operator connectedwith the dynamical system (1) generates thesolution in the state spaceX and has in-finite number of eigenvalues each of finitemultiplicity. The corresponding eigenfunc-tions may form the basis in the infinite-dimensional state spaceX.

It should be stressed that it is possibleto consider another type of linear dynamicalsystems with delays, namely systems withmultiple delays, systems with delays in thecontrol or neutral dynamical systems with de-layed derivative.

dynamo a term used to describe any ofa variety of rotating machines that convertmechanical to electrical energy, or less com-monly, electrical to mechanical energy. Dy-namos typically consist of a stationary struc-ture, called the stator, supporting a rotatingelement called the rotor. Energy conversionoccurs via Faraday induction. A field wind-ing (or in some smaller machines, permanentmagnets) is mounted on one of the mechan-ical structures and produces a magnetic flux.An armature winding is mounted on the otherstructure, and rotation of the rotor producesrelative motion between the field flux andthe coils of an armature winding, inducinga Faraday voltage in the armature coil. ThisFaraday induced voltage is the source of elec-trical energy at the dynamo output.

dynamometer a rotating device used tomeasure the steady-state torque and poweroutput of rotating machines. Dynamometersgenerally provide precise control of the loadtorque applied to a test machine, and poweroutput is determined through precise speedmeasurements.


EE modes the wave solutions with the zeromagnetic field component along the directionof propagation. Also known as transversemagnetic (TM) modes.

E plane in measuring an antenna’s radia-tion pattern, the plane that contains the cur-rent in the element and therefore the electricfield intensity vector field. This plane is per-pendicular to the H plane cut.

E-beam excitation Seeelectron beamexcitation.

E-plane sectoral horn a horn antennawhere the aperture is formed by flaring thewalls in the direction of the E plane. The H-plane dimension is left unchanged from thatof the waveguide feed.

early stopping a technique applied to net-work training that is aimed at assuring goodgeneralization performance. Training on afinite set of data eventually leads to over-fitting, and this can be avoided by period-ically assessing generalization performanceon a set of test data. As training proceeds,network performance on both the training setand the test set gradually improves, but even-tually performance on the test set begins todeteriorate, indicating that training should bestopped. The set of network weights that givethe best performance on the test data shouldbe employed in the trained network.

early vision the set of (mainly perceptual)processes occurring at an early stage of thevision process, typically at the retinal level.

EAROM electrically alterable program-mable read-only memory.See alsoelectro-

nically erasable programmable read-onlymemory.

earth electrode system a network of elec-trically interconnected rods, plates, mats, orgrids installed for the purpose of establishinga low-resistance contact with earth. The de-sign objective for resistance to earth of thissubsystem should not exceed 10�.

Earth station the interface point for com-munications to and from a satellite. An earthstation (also known as a hub) consists of anantenna and transmit and receive subsystems.

earth wire an overhead wire which ismaintained at ground potential for purposesof lightning shielding and system grounding.

earthing Seegrounding.

EBCDIC See extended binary-coded-decimal inter-change code.

ebullient heat transfer the heat transferprocess associated with the formation and re-lease of vapor bubbles on a heated surface.

ECC Seeerror-correcting code.

ECG Seeelectrocardiography.

echo cancellation technique that re-moves the unwanted effects that result fromimpedance mismatch. Echo cancellation fre-quently makes use of adaptive filtering algo-rithms that can be realized by digital filters.

echo canceller in telephony, a filter thatremoves echoes caused by leakage of thetransmitted signal through a hybrid. The hy-brid is the interface between two wire (lo-cal) and four wire (long-distance) facilities,and separates inbound and outbound signals.Echo cancellers are used in analog telephonyto cancel echoes from signals that traversethe entire round-trip connection, and in full-duplex voiceband modems, where the echo isdue to leakage through the near-end hybrid.


echo width in a 2-D scattering problem,the width needed to capture the exact amountof incident power, that when the scatteringbody is replaced by a cylinder, and the cylin-der radiates the captured power, the amountof power received at a specified point is thesame as that received if the scattering body isnot replaced. Echo width is the 2-D analog ofradar cross section in 3-D. The units of echowidth are meters.

Eckert, John Presper Eckert is bestknown as one of the designers of ENIAC(Electronic Numerical Integrator And Calcu-lator), an early computer. Like many com-puter pioneers during World War II, Eckertwas looking for more efficient ways of calcu-lating trajectory tables for artillery and rang-ing systems for radar. Eckert graduated fromthe University of Pennsylvania and remainedthere to work with John Mauchly. Eckert andMauchly later formed a company and con-tinued to develop and refine their machine.Eckert eventually sold the company to whatwould become the Sperry Rand Corporation.Here he produced UNIVAC I (Universal Au-tomatic Computer), one of the first commer-cially successful computers. The chief im-provement of this machine over its predeces-sors was in its use of a stored memory.

ECL Seeemitter-coupled logic.

ecliptic plane of earth’s orbit around thesun.

economic dispatch a generation schemein which units are utilized such that the great-est profit is generated for the utility.

economic interchange an arrangementbetween interconnected electrical power sys-tems whereby a system can meet its load de-mand by buying power from one or moreof the other systems in the interconnectiongroup.

ectopic beat a heart beat that originatesfrom other than the normal site.

eddy current a circulating current in mag-netic materials that is produced as a result oftime-varying flux passing through a metallicmagnetic material.

eddy current brake a braking device inwhich energy is dissipated as heat by gener-ating eddy currents.

eddy current drive a magnetic drive cou-pled by eddy currents induced in an electri-cally conducting member by a rotating per-manent magnet, resulting in a torque that islinearly proportional to the slip speed.

eddy current loss the energy wasted insustaining undesirable eddy currents in anelectrical conductor.

EDFA Seeerbium doped fiber amplifier.

edge (1) a local intensity discontinuity,often corresponding to the boundary of anobject, in an image.See alsograph.

(2) a substantial change, over a small dis-tance, in the values of an image’s pixels —typically in the gray level values. Edges canbe curved or straight and are important be-cause they are often the boundaries betweenobjects in an image.

edge condition an electric/magnetic fieldperpendicular to a dielectric/metallic edgeshows a singular behavior referred to as theedge condition.

edge coupled microstrip lines the mi-crostrip lines that have the same ground planeand are on the same dielectric substrate andare parallel to each other. The coupling ismainly due to the fringing fields at the edgesof the lines.

edge detection the location of edges inan image by computer. See alsoCannyoperator, gradient edgedetector,Marr-Hildreth operator, Sobel operator,straight edge detection.


edge elements the basis functions that areassociated with the edges of the discretizingelements (such as triangles, tetrahedrals, etc.)in a numerical method such as finite elementmethod.

edge enhancement a type of image pro-cessing operation where edges are enhancedin contrast, such as by passing only the highspatial frequencies in an image.

edge guide two conductor transmissionlines in which one of the conductors is a thinsheet of width substantially larger than thegap to the second conductor. The guidedwave is largest at the edge or boundary ofthe thin sheet conductor, standing across tothe second conductor, and propagating in thedirection of the thin sheet conductor edge orboundary.

edge-sensitive pertaining to a bistable de-vice that uses the edge of a positive or neg-ative pulse applied to the control input, tolatch, capture, or store the value indicated bythe data inputs.

edge-triggered Seeedge-sensitive.

Edison, Thomas Alva (1847–1931)Born: Milan, Ohio, U.S.A.

Edison is best known as the holder of 1069patents secured during his lifetime. Amongthese were patents for the phonograph andthe incandescent filament lamp. Edison waslargely self-taught. His early interest in com-munication devices stemmed from his em-ployment as a telegraph operator. He usedthe profits from the sale of his first invention,a “stock ticker,” to set up a lab in Newark,New Jersey. Always a shrewd commercialdeveloper, he followed the invention of thelight bulb with work on developing efficientgenerators to power these bulbs. Edison isconsidered the archetypal American inven-tive genius.

Edison Electric Institute a trade groupof investor-owned public electric utilities inthe USA.

edit distance the edit distance betweentwo string A and B is defined as

Dedit = min{a + b + c}

whereB is obtained fromA with a replace-ments,b insertions, andc deletions. There isan infinite number of combinations{a, b, c}to achieve this. One of the ways to find theminimum from these is dynamic program-ming. Edit distance is also called Leven-shtein distance.

EEG Seeelectroencephalography.

EEPROM See electronically erasableprogrammable read-only memory.

effective address (1) the computed ad-dress of a memory operation.

(2) the final actual address used in a pro-gram. It is usually 32 or 64 bits wide.

The effective address is created from therelative address within a segment (that is, rel-ative to the base of a segment), which has hadapplied all address modification specified inthe instruction word. Depending on the con-figuration of the memory management unit,the effective address may be different fromthe real address used by a program, i.e., theaddress in RAM, ROM, or I/O space wherethe operation occurs.

(3) when a memory location is referencedby a machine instruction, the actual memoryaddress specified by the addressing mode iscalled the effective address. For example, ifan instruction uses an indirect mode of ad-dressing, the effective address is to be foundin the register or memory location whose ad-dress appears in the instruction.

effective aperture a figure of merit for anantenna giving the equivalent aperture sizerequired to intercept from a field of uniformpower density an amount of power equal to


that received by the antenna placed in thesame field. At a given wavelength, the ef-fective aperture is directly proportional to thegain.

effective dielectric constant (1) simpleand single dielectric constant used to describea complicated configuration of media with avariety of dielectric constants in an equiva-lent model.

(2) the resulting computational effectsof having two dielectric materials in a mi-crostrip transmission line.

effective isotropic radiated power theproduct of the total radiated power by the di-rective gain of the antenna.

effective isotropically radiated power(EIRP) in antenna theory, the amount ofpower needed by an isotropic radiator to pro-duce the same radiation intensity at a receiveras the original antenna in the main beam di-rection. EIRP, which is expressed in decibel-meters or decibel-watts, can be calculated bymultiplying power supplied to an antenna byits directive gain in the desired direction.Seealsoeffective radiated power.

effective length the ratio of the volt-age induced across an antenna terminat-ing impedance divided by the incident fieldstrength.

effective mass an approach whereby aparticular response is described using clas-sical equations by defining an effective masswhose value differs from the actual mass. Forexample, an electron in a lattice responds dif-ferently to applied fields than would a freeelectron or a classical particle.

effective permittivity a simple and singlepermeability constant used to describe a com-plicated configuration of media with a vari-ety of permeability constants in an equivalentmodel.

effective radiated power (ERP) the ef-fective power output from an antenna in aspecified direction, including transmitter out-put power, transmission line loss, and an-tenna power gain.

efficiency (1) the ratio of the input powerto the output power. It is a figure of merit forthe energy cost effectiveness of a device.

(2) in antennas, the ratio of the power ra-diated to the input power. This term is some-times defined with the mismatch loss (1−γ 2)included in the total efficiency of the antenna;other times, it is omitted from the calculation.

efficient estimator an unbiased estimatorwhich achieves the Cramer-Rao bound.SeealsoCramer-Rao bound.

EFIE Seeelectric field integral equation.

EGA Seeenhanced graphics adaptor.

EIA Electronics Industry Association.

eigenfunction the name given to an eigen-vector when the eigenvectors arise as solu-tions of particular types of integral equation.See alsoeigensystem, eigenvector.

eigenfunction expansion a method usedto expand a given field in terms of eigenfunc-tions. It is particularly used in modal analysisof waveguide discontinuities.

eigenstate a linear combination of quan-tum mechanical basis states that is constantin time. A quantum mechanical system start-ing in an eigenstate will remain unchanged intime except for an overall phase. The phasevaries as the product of the eigenvalue andtime. Quantum mechanical eigenstates areanalogous to normal modes of coupled oscil-lator systems in classical mechanics.

eigensystem a system where the output ofa system is the input function multiplied by aconstant.See alsoeigenfunction, eigenvalue,eigenvector.


eigenvalue the multiplicative scalar as-sociated with an eigenfunction or an eigen-vector. For example, ifAx = λx, thenλis the eigenvalue associated with eigenvec-tor x. If A is a covariance matrix, thenλrepresents the variance of one of the princi-pal components ofA. See alsoeigensystem,eigenvector, principal component.

eigenvalue assignment a technique that,given a set of desired eigenvalues, has theobjective to compute a constant state feed-back gainK such that the closed-loop stateequation has precisely these eigenvalues.

eigenvector for a linear systemA, anyvectorx whose direction is unchanged whenoperated upon byA. See alsoeigensystem,eigenvalue.

eight connected Seepixel adjacency.

eighteen connected Seevoxel adjacency.

Einstein coefficients coefficients intro-duced by Einstein to represent the sponta-neous decay rate,A, and the stimulated emis-sion or absorption coefficient,B, in the pres-ence of an electromagnetic energy density orintensity.

EIRP Seeeffective isotropically radiatedpower.

EISA See extended industry standardarchitecture.

elastic scattering (1) a scattering processin which energy is conserved.

(2) the state in which there is no en-ergy transfer between the light wave and thescattering medium, and hence no frequencychange.

elasto-optic effect the change in refractiveindex of a medium due to mechanical forces.Also known as the photoelastic effect.

elastomer a polymer material that can un-dergo reversible shape changes, expandingor contracting under influence of an appliedelectric field; often used for mechanical lightmodulation devices.

elbow in URD work, the termination of aburied cable where it attaches to the distribu-tion transformer.

electret a material similar to ferro-electrics, but charges are macroscopicallyseparated and thus are not structural. In somecases the net charge in the electrets is notzero, for instance when an implantation pro-cess was used to embed the charge.

electric charge a basic physical quantitythat is a source of electromagnetic fields. Theunits of electric charge are coulombs.

electric charge density the fundamental,macroscopic source of the electromagneticfield that quantifies the average number ofdiscrete electric charges per unit volume. SIunits are coulombs per cubic meter.

electric current density a source vec-tor in electromagnetics that quantifies theamount of electric charge crossing somecross–sectional area per unit time. The di-rection of the electric current density is inthe direction of electric charge motion. SIunits are amperes per square meter.

electric field in a region of space, if a testchargeq experiences a forceF then the re-gion is said to be characterized by an electricfield of intensityE given by

E = F

q

electric field integral equation (EFIE)the integral equation based on the boundarycondition of the total electric field. It is oftenused in moment method description of mi-crostrip circuits and metallic antennas.Seealsomoment method.


electric field intensity a force field that isa measure of the magnitude and direction ofthe force imparted upon a discrete charge nor-malized to the discrete charge’s value. De-pends on material characteristics. The unitsare volts per meter.

electric flux density basic electromag-netic field quantity used to describe the ef-fects of permeable matter to the electric field;it is expressed in SI units of coulombs persquare meter.

electric furnace a method of smeltingmetals and applying high heat for industrialprocesses which makes use of the heat froman electric arc struck between (typically) car-bon electrodes.

electric permittivity tensor relationshipbetween the electric field vector and the elec-tric displacement vector in a medium with nohysteresis; displacement divided by the elec-tric field in scalar media.

electric polarization vector an auxil-iary vector in electromagnetics that accountsfor the creation of atomic dipoles in a di-electric material due to an applied electricfield. Macroscopically, the electric polariza-tion vector is equal to the average numberof electric dipole moments per unit volume.Mathematically,P = D − ε0E, whereD isthe electric flux density,E is the electric fieldintensity, andε0 is the free space permittivity.SI units are coulombs per square meter.

electric susceptibility tensor relationshipbetween the electric field vector and the elec-tric polarization vector in a medium with nohysteresis. It is the polarization divided bythe permittivity of free space and the electricfield in scalar media.

electric vector potential a vector functionthat is used to derive solutions for electric andmagnetic fields.

electric vehicle a vehicle powered by anelectrical energy storage device such as bat-teries.

electrical breakdown Seebreakdown.

electrical degrees a convenient way ofrepresenting the distance around the circum-ference of a machine with two poles spanningthe entire 360◦ of the circumference,

Electrical Degree= Pairs of poles

× Mechanical Degree

electrical network a collection of inter-connected electrical devices.

electrical tree a microscopic crackingpattern which forms in the insulation of elec-tric power cables which are not exposed towaterSeetree, water tree.

electrically alterable read-only memory(EAROM) a PROM device that can beerased electronically. More costly than theEPROM (erasable PROM) device, whichmust be erased using ultra-violet light.

electrically erasable programmable read-only memory (EEPROM) a nonvolatilesemiconductor memory, it is used at theplace of PROM or EPROM. They can beprogramed and erased by electrical means.

electro-acoustic smart material one ofseveral materials that have self-adaptive char-acteristics in their acoustic behavior (suchas transmission, reflection, and absorption ofacoustical energy) in response to an externalstimulus applied as a function of the sensedacoustical response.

electro-optic coefficient a parameter thatdescribes the change in refractive index in amedium with application of an electric field.The linear electro-optic coefficient describes


media where the index change is linearly re-lated to applied field; the quadratic electro-optic coefficient describes media where theindex change is related to the square of theapplied field, i.e., the Kerr effect.

electro-optic device any device that usesthe electro-optic effect, such as for opticalbeam modulation and deflection.

electro-optic effect an optical phenome-non in certain types of crystals in which theapplication of an applied electric field resultsin a change in the dielectric tensor, or equiv-alently a change in the refractive index ofthe medium. In the linear electro-optic ef-fect, the resulting change in refractive indexis proportional to the electric field strength; inthe quadratic electro-optic effect, the result-ing change in refractive index is proportionalto the square of the electric field strength.The linear electric-optic effect was first stud-ied by F. Pockels in 1893.

electro-optic modulator a device thatuses the electro-optic effect to alter the tem-poral and spatial character of a light beam.

electro-optic scanner a device that usesthe electro-optic effect to scan a light beamacross a range of angles.

electro-optic smart materials materialsin which optical properties are changed self-adaptively with an external electric stimulusproportional to the sensed optical character-istics.

electro-optic switch a device that uses theelectro-optic effect to switch a light beam onor off.

electro-optic wave retarder a device thatwill retard the phase of an optical beam by afixed amount using the electro-optic effect;in particular, the retardation of one polariza-tion component of the light beam using theelectro-optic effect in a birefringent material.

electroacoustics science of interfacingbetween acoustical waves and correspondingelectrical signals. This includes the engineer-ing of transducers (e.g., loudspeakers and mi-crophones), but also parts of the psychologyof hearing, following the notion that it is notnecessary to present to the ear signal compo-nents that cannot be perceived.

electrocardiography (ECG) the device(electrocardiograph) or the output (elec-trocardiogram) depicting the body surfacerecording of the electrical activity of theheart.

electroencephalography (EEG) record-ings of the electrical potentials produced bythe brain.

electroluminescence a general term foroptical emission resulting from the passageof electric current.

electrolyte current-conducting solutionbetween two electrodes or plates of a capac-itor, at least one of which is covered by adielectric.

electrolytic capacitor a capacitor solu-tion between two electrodes or plates of acapacitor, at least one of which is covered bya dielectric.

electromagnet a magnet that employs anelectric current in a coil to produce a mag-netic field.

electromagnetic compatibility (EMC)the ability of a system or equipment tooperate within design tolerances in its in-tended environment, with adjacent systemsand equipment, and with itself, so that theeffect of any electromagnetic disturbancesproduced by the systems or equipment is re-duced.

electromagnetic energy energy con-tained in electromagnetic fields and associ-ated polarizable and magnetizable media.


electromagnetic environmental effectsencompasses all electromagnetic disciplines,including electromagnetic compatibility(EMC); electromagnetic interference (EMI);electromagnetic vulnerability (EMV); elec-tromagnetic pulse (EMP); radiation hazard(RADHAZ) (hazard of electromagnetic ra-diation to personnel, ordnance, and fuels(HERP, HERO, HERF)); lightning, p-static;electrostatic discharge (ESD), and emissioncontrol (EMCON).

electromagnetic interference (EMI) (1)any electromagnetic disturbance that inter-rupts, obstructs, or otherwise degrades orlimits the effective performance of electron-ics/electrical equipment. It can be inducedintentionally, as in some forms of electronicwarfare, or unintentionally, as a result of aspurious emissions and responses, intermod-ulation products, and the like. Additionally,EMI may be caused by atmospheric phenom-ena, such as lightning and precipitation staticand non-telecommunication equipment, suchas vehicles and industry machinery.

(2) unwanted high-frequency electricalsignals, also known as radio frequency in-terference (RFI), which can be generated bypower electronic circuits switching at highfrequencies. The signals can be transmit-ted by conduction along cables (450 kHz to30 MHz) or by radiation (30 MHz to 40 GHz)and can interfere with control or other elec-tronic equipment.

electromagnetic interference filter a fil-ter used to reduce or eliminate the electro-magnetic interference (EMI) generated bythe harmonic current injected back onto theinput power bus by switching circuits. Theharmonic current is caused by the switchaction that generates switch frequency rip-ple, voltage and current spikes, and high-frequency ringing. Generally called an EMIfilter.

electromagnetic pulse (EMP) a largeimpulsive-type electromagnetic wave gener-ated by nuclear or chemical explosions.

electromagnetic radiation an electro-magnetic wave created by the acceleration ordeceleration of charge.

electromagnetic smart materials ma-terials such as shielding materials, radar-absorbing materials (RAMs), and electro-magnetic surface materials, in all of whichsome electromagnetic properties can beadaptively controlled by means of an externalstimulus dictated by the sensed electromag-netic response.

electromagnetic spectrum the frequencyand wavelength of electromagnetic radiation.We have the following classification reportedin the figure, while the microwave frequencyband designations is reported in the table.

electromagnetic susceptibility a device’sfailure to perform appropriately if there is anelectromagnetic disturbance.

electromagnetic torque the torque pro-duced in a machine by the interaction of themagnetic fields and/or by the varying reluc-tance principle where the field attempts tomaximize its intensity in a machine duringelectromechanical energy conversion.

electromagnetic vulnerability (EMV)the inability of a device, equipment, or sys-tem to perform without degradation whensubjected to electromagnetic environment ofa specified power level and frequency range.

electromagnetic wave wave in which theelectric and magnetic variables are solutionsof the Maxwell–Heaviside equations.

electromagnetic wave propagation thephenomenon of electromagnetic energy prop-agating in the form of waves of the coupledelectric and magnetic field intensity vectors.

electromagnetically induced transparencya technique to render optically dense mediatransparent by using a long-lived quantum


Electromagnetic spectrum.

mechanical coherence. Depending on thequantum mechanical system, electromagnet-ically induced transparency can be viewed asa special case of effects such as Fano interfer-ence or coherent population trapping appliedto optically dense media.

electromagnetics the study of the effectof electric charges at rest and in motion.

electromechanical equation a basic non-linear equation that governs the rotational dy-namics of a synchronous machine in stabilitystudies. The equation is given by

2Hd2δ/ωsdt2

where H is a constant defined as the ra-tio of the kinetic energy in megajoules atsynchronous speed to the machine rating inMVA, ωs is the synchronous speed in radsper second, andδ is the load angle expressedin electrical degrees. Various forms of theswing equation serve to determine stabilityof a machine within a power system. So-lution of the swing equation yields the load

angleδ as a function of time. Examination ofall the Swing curves (plot ofδ w.r. to time)shows whether the machines will remain insynchronism after a disturbance.See alsoswing equation.

electromechanical relay a protective re-lay that uses electrical, magnetic, and me-chanical circuits to implement the operatinglogic.

electron band a range or band of energiesin which there is a continuum (rather than adiscrete set as in, for example, the hydrogenatom) of allowed quantum mechanical statespartially or fully occupied by electrons. Itis the continuous nature of these states thatpermits them to respond almost classicallyto an applied electric field.

electron beam excitation electron impactexcitation due to a free electron beam ratherthan, for example, the conduction electronsin a gas discharge; permits high-pressure ex-citation without arcing.


Microwave frequency band designationsFrequency Wavelength IEEE Radar Band Designation

Old New1–2 GHz 30–15 cm L D2–3 GHz 15–10 cm S E3–4 GHz 10–7.5 cm S F4–6 GHz 7.5–5 cm C G6–8 GHz 5–3.75 cm C H8–10 GHz 3.75–3 cm X I

10–12.4 GHz 3–2.42 cm X J12.4–18 GHz 2.42–1.67 cm Ku J18–20 GHz 1.67–1.5 cm K J

20–26.5 GHz 1.5–1.13 cm K K26.5–40 GHz 1.13 cm–7.5 mm Ka K40–300 GHz 7.5–1.0 mm mm

electron beam lithography lithographyperformed by exposing resist with a focusedbeam of electrons.

electron beam welding a welding pro-cess that produces coalescence of metals withthe heat obtained from a concentrated beamcomposed primarily of high-velocity elec-trons impinging on the surfaces to be joined.

electron collision frequency the averagenumber of collisions per second an electronhas with heavy particles in a medium such asplasma.

electron impact excitation excitation ofan atom or molecule resulting from collisionby an electron.

electron multiplication the phenomenonwhere a high-energy electron strikes a sur-face and causes additional electrons to beemitted from the surface. Energy from theincident electron transfers to the other elec-trons to cause this. The result is electron gainproportional to the incident electron energy.

electron oscillator model simplified clas-sical model for an atomic or molecularmedium in which the charges are assumedto be bound together by springs rather thanquantum mechanical potentials; provides

good qualitative explanation for absorptionand dispersion.

electron plasma a plasma medium inwhich electrons are the mobile charge car-riers and the ions form the stationary com-pensating positive charge background.

electron wave the wave described byBloch function solutions to the problem ofan electron in a periodic lattice of ions.

electron-beam lithography refers to alithographic (or photographic) process inwhich the exposure energy is provided by theenergy carried by a beam of focused electronsrather than by photons (light).

electronic bottleneck the factor limitingthe speed and capacity of a fiber optic com-munication network is ultimately the link be-tween photons of light and the electronicsrequired to transmit or receive and processthem. In order to exploit the full bandwidthcapacity of a single-mode optical fiber an op-tical network should minimize the number ofoptical to electrical conversions and insteadprocess the signals in the optical domain.

electronic brake a power electronic sys-tem designed to decelerate a motor. For aninduction machine, the brake supplies DCcurrent to the stator winding, producing a


stationary magnetic field, negative slip, andbraking torque.

electronic motor starter starter in whichsolid state devices provide reduced voltageto the motor for starting, thus limiting thestarting current.

electronic nonlinear response the non-linear optical response resulting from the mo-tion of bound electrons. It is characterizedby moderately large response and very short(several fs) response times.

electronic overload device an overloaddevice that employs an electronic circuitto sense motor voltage and current for thepurpose of providing precise motor over-load protection. See alsooverload heater,overload relay.

electronic switch an electronic circuit thatcontrols analog signals with digitary (binary)signals.

electronic transition alteration in theelectronic structure of a material such thatone electron temporarily changes its energylevel through the absorption or emission ofenergy.

electronic warfare contention for thecontrol of the electromagnetic (EM) spec-trum, to allow active and passive EM sensingand communications while denying the sameability to adversaries. Includes deceptive EMtechniques.

electronically commutated machine aDC machine with rotor-mounted permanentmagnets, and concentrated (square) statorwindings. The machine fundamentally be-haves like a DC machine with linear speedand torque characteristics, and hence is alsocalled a brushless DC machine. The ma-chine is distinguished from conventional DCmachines in its substitution of a six-switchinverter for brushes and copper commutator

bars and in its need for the rotor position in-formation in real time.

electronically erasable programmableread-only memory (EEPROM) a termused to denote a programmable read-onlymemory where the cells are electronicallyboth written and erased. Also known aselectrically alterable read-only memory.Seealso electronically programmable read-onlymemory.

electronically programmable read-onlymemory (EPROM) Programmable read-only memory that is electronically written butrequires ultraviolet light for erasure.

electroplastic effect plastic deformationof metals with the application of high-densityelectric current.

electroplastic smart material materialwith smart properties of elastic deformationchanges proportional to a controlled electriccurrent applied in proportion to the senseddeformation.

electrorheological property property ex-hibited by some fluids that are capable of al-tering their flow characteristics depending onan externally applied electric field.

electrorheological smart fluid fluid withsmart flow characteristics dictated to changeself-adaptively by means of an electric fieldapplied in proportion to the sensed flow pa-rameters.

electrorheological smart material ma-terial with smart properties of elastic defor-mation changes proportional to a controlledelectric current applied in proportion to thesensed deformation.

electroslag welding a welding processthat produces coalescence of metals withmolten slag that melts the filler metal and thesurfaces of the parts to be joined.


electrostatic discharge (ESD) the dis-charge of a body through a conducting pathbetween two pins of an IC. Circuits locatedat the inputs and outputs of ICs protect theinternal devices from ESD events.

electrostatic precipitator a method of ex-tracting dust from stack gases or ventilatingsystems in which ions are laid on the dustparticles by high-voltage electrodes and thenattracted electrostatically into a trap.

electrostatic voltmeter a voltmeter, typ-ically used for voltages in the kilovolt range,in which the pointer is moved by the elec-trostatic attraction of a pair of metal platesacross which the voltage to be measured isapplied.

electrostriction the tendency of materialsto become compressed in the presence of anapplied electric field. The change in densityis proportion to the square of the electric fieldstrength. This process leads to an increase inthe refractive index of the material, describ-able byδn = n2I , wheren2 is the (positive)coefficient of the nonlinear refractive indexandI is the intensity of the field in units ofpower per unit area. For condensed matter, atypical value ofn2 is 10−20 m2/W.

element factor in antenna theory, that partof the radiation pattern that is governed by thegeometrical shape of the antenna that con-strains the current.

Elias’ upper bound for any(n, k) blockcode, the minimum distance is boundedasymptotically as

dmin

n≤ 2A(1 − A)

where the parameterA is related to the coderate through the equation

k

n= 1 + A log2A+ (1 − A) log2(1 − A)

0 ≤ A ≤ 1

2

ellipse detection the detection of ellipsesin digital images, often with a view to lo-cating elliptical objects or those containingellipses; ellipse detection is also importantfor the location of circular features on realobjects following orthographic projection orperspective projection.

ellipsometry measurement of the changesof light polarization produced by scattering.

elliptic Seeelliptic filter.

elliptic filter (1) filter with an equal ripplepassband frequency response, but in whichthe stop band also exhibits an equal ripple(peaks) stopband response. Also known as aCaurer filter.

(2) member of a class of low pass, highpass, band pass and band stop filters with anequiripple characteristic, designed to achieveoptimally rapid rollof rates near cutoff fre-quencies (See alsoButterworth filter andChebyshev filter) at the expense of a lossof monotonicity in both the passbands andthe stopbands. For example, an elliptic lowpass filter design is equiripple in the passbandand stopband, and has a squared magnituderesponse of the form

|H(jw)|2 = 1

1 + ε2U2N(w/wc)

,

whereUN(w) is a Jacobian elliptic function.

elliptical polarization the polarizationstate of a radiated electromagnetic field inwhich the tip of the electric field vector tracesan ellipse as a function of time for a fixed po-sition. The sense of rotation of the electricfield vector is either right-hand or left-hand(clockwise or counter-clockwise). Circularpolarization and linear polarization are spe-cial cases of elliptical polarization.

embedded computer (1) a computingmachine contained in a device whose pur-pose is not to be a computer. For example,the computers in automobiles and householdappliances are embedded computers.


(2) a device, consisting of a micropro-cessor, firmware (often in EPROM), and/orFPGAs/EPLDs, which is dedicated to spe-cific functions, and becomes an inseparablecomponent of a device or system, in contrastto devices that are controlled by stand-alonecomputers. Embedded computers use em-bedded software, which integrates an oper-ating system with specific drivers and ap-plication software. Their design often re-quires special software–hardware codesignmethods for speed, low power, low cost, hightestability or other special requirements.

(3) software that is part of a larger systemand performs some of the requirements ofthat system; e.g., software used in an aircraftor rapid transit system. Such software doesnot provide an interface with the user.Seefirmware.

embedded computer system Seeembed-ded computer, embedded system.

embedded passives 3-D packaging so-lution, consisting of embedding passive el-ements into the mounting substrate, for in-creasing the packaging efficiency.

embedded system software that is partof a larger system and performs some of therequirements of that system; e.g., softwareused in an aircraft or rapid transit system.Such software does not provide an interfacewith the user.See alsofirmware.

EMC Seeelectromagnetic compatibility.

EMI Seeelectromagnetic interference.

EMI filter Seeelectromagnetic interfer-ence filter.

emission credit a scheme in the USA forthe control of nitrous oxide and sulfur diox-ide emissions from industrial plants. A utilitycan purchase from the government the rightto emit a certain quantity of pollutant. If itdoes not emit this amount, the credit can besold on the market to another plant which

may need to emit some pollutant. The resalevalue of the pollution credit forms a finan-cial incentive for a facility to emit as littlepollution as possible.

emissivity the fraction of the power in-cident on a material that is reradiated afterbeing absorbed by the material. For a mate-rial in thermal equilibrium, the emissivity isequal to the absorptivity.

emitter follower Seecommon collectoramplifier.

emitter-coupled logic (ECL) a very highspeed bipolar transistor logic circuit family.

emitter-coupled pair See differentialpair.

EMP Seeelectromagnetic pulse.

empirical distribution function Seehistogram.

empirical model mathematical modelbased on curve-fitting specific mathematicalfunctions to measured data, rather than ondevice physics. Empirical models generallyhave a low to midrange modeling valuationcoefficient.

EMTP the Electro-Magnetic TransientProgram, a computer program which simu-lates an electric power system such that itsresponse to disturbances may be accuratelypredicted.

emulate executing a program compiled toone instruction set on a microprocessor thatuses an incompatible instruction set, by trans-lating the incompatible instructions while theprogram is running.

emulation a model that accepts the sameinputs and produces the same outputs as agiven system. To imitate one system withanother.Contrast withsimulation.


emulation mode state describing the timeduring which a microprocessor is performingemulation.

emulator (1) the firmware that simulatesa given machine architecture.

(2) a device, computer program, or systemthat accepts the same inputs and produces thesame outputs as a given system.Comparewith simulator.

EMV Seeelectromagnetic vulnerability.

encapsulation property of a program thatdescribes the complete integration of datawith legal process relating to the data.

enclosure a box, cabinet, wall, fence,barrier, or other means designed to protectpersonnel from accidentally contacting ener-gized electrical parts and to protect the elec-trical parts from physical damage.

encoder (1) a device that directly createsa digital signal based on an analog value.

(2) a logic circuit with 2N inputs andNoutputs, the outputs indicating the number ofthe one input line that is asserted.

(3) a device used to obtain positional in-formation. The encoder uses a disk mountedon the shaft of a rotating machine and a lightsource. The combination provides lightspulses that can be decoded to provide the an-gular position of the motor.See alsoabsoluteencoderandincremental encoder.

encoding (1) the act of placing informa-tion to be transmitted in a form that can betransmitted over a particular medium and willbe recognizable by the receiver.See alsocoding.

(2) in computing systems, to representvarious pieces of information by a definedsequence of the binary digits 0 and 1, calledbits. To apply the rules of a code.

encryption the transformation employedto transform information to be transmitted(plaintext) into a format that is unintelligible

(ciphertext or a cryptogram). The ciphertextcan then be transmitted via a communicationchannel without revealing the contents of theplaintext. This is achieved by means of anencryption key. A system for performing en-cryption is also known as a cipher. The infor-mation to be encrypted is referred to as plain-text, and the encrypted message resultingfrom encryption is referred to as ciphertext.The intended receiver of the ciphertext alsohas the encryption key, and by having boththe ciphertext and the encryption key avail-able, the original plaintext can be recovered.See alsoencryption key, block cipher, streamcipher, public key cryptography, private keycryptography.

encryption key a codeword used for de-cryption of ciphertext into plaintext in en-cryption systems. Ciphertext can then betransmitted via a communication channelwithout revealing the contents of the plain-text. The plaintext can only be recovered bysomeone in possession of the encryption key.See alsoencryption, block cipher, streamcipher.

end bell the cap that forms the end of thestator housing for an electric machine with acylindrical frame.

end-around carry technique used inone’s complement arithmetic, in which a car-ryover of the result of an addition or sub-traction beyond the leftmost bit during addi-tion (or subtraction) is “wrapped around” andadded to the result.

end-effector a part found at the freeend of the chain of the links which forma manipulator. Position and orientation ofthe manipulator are referred to the end-effector. The frame attached the end-effectoris known as end-effector frame. The end-effector can be a gripper or it can be at-tached to the end-effector. The orthonormalend-effector frame consists of three unit vec-torsa-approach (it approaches the object)o-orientation (which is normal to the sliding


plane between the fingers of the gripper),andn-normal vector to two others so that theframe (n, o, a) is right-handed.

end-fire the pattern factor is maximum intheE plane (for a dipole antenna along thez axis, this is the plane whereθ = 0 or 180degrees).

end-fire array a linear phased array an-tenna in which the amplitudes and phases ofthe element excitations are adjusted such thatthe direction of maximum radiation is alongthe axis passing through the center of the ar-ray elements.

endian an adjective used with a qualifierto indicate how long values are constructedfrom smaller values that reside in memory. A“big-endian” machine constructs long valuesby placing the smaller parts from the loweraddresses at the left end of the value (the“big” end).

endoscope remote imaging system usedby physicians to examine a patient’s internalorgans. The cylindrical rod is placed eitherthrough an existing opening in the body orthrough a small incision. The rod contains anilluminating fiber optic bundle and an opticalsystem to deliver the image to a video mon-itor. In a conventional endoscope, the imageis transmitted through a series of lenses toa camera outside of the patient. In an elec-tronic endoscope the image is collected by aCCD camera positioned inside the patient.

endpoint detection the process of iso-lating spoken words, typically used subse-quently for word recognition.

ends a crimped-type wire connector.

energy that which does work or is capa-ble of doing work. In electrical systems, itis generally a reference to electrical energymeasured in kilo-watt hours.

energy band continuous interval of en-ergy levels that are allowed in the periodicpotential field of the crystalline lattice.

energy banking pertaining to the mainte-nance of a thermal unit on hot reserve.

energy compaction in a transformation,the concentration of the input signal energyinto a relatively small part of the transformedsignal. A linear transform of a random vectorcompacts the signal energy when the energyor variance of a small number of transformcoefficients is large relative to the variance ofthe other coefficients.

energy conservation the conservation ofenergy between the input and output of a sys-tem; i.e., the energy of the output signal isequal to that of the input signal to withina constant factor. A unitary or orthogonaltransform conserves energy in that the energyor magnitude of the output vector is equal tothat of the input vector.

energy gap the width of the energy inter-val between the top of the valence band andthe bottom of the conduction band.

energy level one of the specific valuespossible for the energy of an electron in anatom or molecule.

energy product the product of the mag-netic flux density,B and magnetic field inten-sity,H at any operating point on the normaldemagnetization curve, indicating the energydelivered by the magnet. The maximum en-ergy product is commonly used to designatevarying grades of materials.

energy relaxation time the characteristictime for energy loss to scattering processes.

energy signal a continuous time signalf (t) is an energy signal if∫ ∞

−∞f (t)2dt < ∞ .


For example, the signalf (t) = e−t , t ≥ 0 isan energy signal. The signalf (t) = 1 is notan energy signal. The definition has practi-cal significance if the signalf is a voltage,or current, applied to an electric circuit, asshown in the figure. The energy dissipatedby the circuit is given by

1

R

∫ inf ty

−∞v(t)2dt .

A discrete time signal is an energy signalif

∞∑k=−∞

f (k)2 < ∞ .

energy-density spectrum for a continu-ous time signalx(t) the function|X(w)|2,where X(w) is the Fourier Transform ofx. The function |X(w)|2 is referredto as an energy-density spectrum because1|X(w0)|2 is proportional to the energy inthe signalx in a bandwidth1 around thefrequencyw0. Analogously, a discrete timesignalx[n] has an energy-density spectrum|X(�)|2.

enhanced backscattering peak of the an-gular distribution of scattered intensity in theretroreflection direction, due to multiple scat-tering of light in dense volumes or very roughboundaries.

enhanced graphics adaptor (EGA) avideo adapter proposed by IBM in 1984. It iscapable of emulating CGA and MDA. It canreach 43 lines with 80 columns. In graphicmode, it can reach 640× 350 pixels (wide perhigh) with 16 colors selected from a pallet of64.

enhanced small device interface Seeenhanced small disk interface.

enhanced small disk interface (ESDI)mass storage device interface similar to MFMand RLL except that the clock recovery cir-cuits are in the peripheral device rather than

in the controller. Originally designed byMaxtor.

enhancement improvement of signalquality without reference to a model of sig-nal degradation. See restoration, imageenhancement.

enhancement mode an FET that is offwhen zero volts bias is applied from gate tosource.

ENI Seeequivalent noise current.

ENR Seeexcess noise ratio.

enriched uranium uranium which con-tains 1.5% to 5% of fissile U-235 and is thussuitable for use in power reactors.

enrichment the process of increasing theratio of one isotope of a chemical element,e.g. fissile uranium, to another, less desirableisotope, e.g. non-fissile uranium.

ensemble processor a parallel processorconsisting of a number of processing ele-ments, memory modules, and input/outputdevices under a single control unit. It hasno interconnection network to provide inter-processor or processor-memory communica-tions.

entity a software process that implementsa part of a protocol in a computer communi-cation network.

entity relationship diagram a diagramthat describes the important entities in a sys-tem and the ways in which they are interre-lated.

entropy information theoretic quantityrepresenting the amount of uncertainty inguessing the value of a random variable. Fora discrete random variableX, with probabil-ity mass functionp(x) defined on the support


setS, the entropy,H(X) is defined as

H(X) = −∑x∈S

p(x) logp(x)

Also writtenH(p), emphasizing the sole de-pendence upon the mass function.See alsodifferential entropy, relative entropy, mutualinformation.

entropy coding a term generally used asequivalent to lossless source coding. Thename comes from the fact that lossless sourcecoding can compress data at a rate arbitrarilyclose to the entropy of the source.

entropy estimation noiseless source cod-ing theorem states that the bit rate can bemade arbitrarily close to the entropy of thesource that generated the image. Entropy es-timation is the process of characterizing thesource using a certain model and then find-ing the entropy with respect to that model.The major challenge here is to approximatethe source structure as close as possible whilekeeping the complexity of the model and thenumber of parameters to a minimum.

entry mask bit pattern associated with asubroutine entry point to define which pro-cessor registers will be used within the sub-routine and, therefore, which should be savedupon entry to the subroutine. Some proces-sor designs perform this state saving duringthe execution of the instruction that calls asubroutine.

entry point instruction that is the first in-struction in a subroutine.

ENV Seeequivalent noise voltage.

envelope the imaginary waveform pro-duced by connecting the peak values of amodulated carrier wave. For AM, the am-plitude of the carrier sinusoid is a function oftime

e(t) = E(t) sin(ωc + θ)

whereE(t) is a function of the intelligence.In this case, the envelope of the resulting dou-

ble sideband AM waveform,e(t), is a scaledrepresentation of the intelligence.

envelope delay the time it takes for aspecific reference point on the envelope ofa modulated wave to propagate between twopoints in a circuit or transmission system. Ina time invariant system, the envelope delay isthe derivative (i.e., rate of change) of phase(in radians) with respect to angular frequency(also in radians) as the envelope of the sig-nal passes between two points in the system.Also referred to as time delay.

envelope detection a device that producesan output waveform that is proportional tothe real envelope of its input, i.e., when theinput isA(t)cos(2πfct + θ(t), the output isv(t) = CA(t) with C a constant.

envelope detector the optimum structurefor detecting a modulated sinusoid with ran-dom phase in the presence of additive whiteGaussian noise.

environment a set of objects outside thesystem, a change in whose attributes affects,and is affected by, the behavior of the system.

environmental dispatch a generationscheme in which units are committed so asto minimize disturbance to the natural envi-ronment.

epipolar line in stereo vision, the intersec-tion of an image plane with an epipolar plane,which is the plane identified by a pointP andthe centers of projection of the two cameras.

epitaxial layer a doped layer of semi-conducting material grown on the surface ofa prepared semiconductor substrate. Vari-ous methods are utilized, resulting in a verythin and dimensionally well controlled activelayer for fabricating semiconductor devices.“Epi” parts generally are more uniform andperform better than ion implanted devices,but are also more expensive to fabricate.


EPROM See erasable programmableread-only memory.

equal area criterion a criterion used of-ten in power system studies whenever thestability of a single machine system needsto be determined without actually solvingthe swing equation. It is a direct approach.See alsoelectromechanical equationorswingequation.

equal gain combining a method of di-versity combining in which the outputs ofseveral communication channels are first co-phased and then summed. The resulting sig-nal has increased mean power and reducedfade depth.See alsoangle diversity, antennadiversity.

equal ripple in-band power gain or powerloss vs. frequency response in which the min-imum dips are equal in power and maximumpeaks are equal in power.

equalization a method used in communi-cation systems to compensate for the channeldistortion introduced during signal transmis-sion.

equalizer a device used at the receiver(typically a digital filter) that attempts to sup-press or cancel intersymbol interference.

equalizer adaptive an equalizer that auto-matically adjusts its parameters to suppressor cancel intersymbol interference. In thecase of a digital filter, the parameters are thefilter coefficients, which are adapted to min-imize a cost criterion (e.g., mean-squared er-ror).

equalizer decision-feedback a nonlineardigital filter that consists of a linear prefilterP(z) for suppressing precursor intersymbolinterference, and a feedback filterF(z) forsuppressing postcursor intersymbol interfer-ence.

equalizer minimum mean squared error(MMSE) a digital equalizer (filter) in

which the coefficients are selected to min-imize the mean-squared error between thetransmitted symbols and the filter outputs.

equalizer transversal an equalizer that isimplemented as a finite impulse response dig-ital filter. Also called a “tapped-delay line.”

equalizer zero-forcing an equalizer inwhich the parameters are adjusted to elimi-nate intersymbol interference (at the expenseof enhancing the noise).

equalizing pulse interval series of 12sync pulses inserted in the vertical blank-ing interval of each field of the NTSC com-posite video signal. The equalizing pulsemakes a transition between the compositesignal blanking level and the sync level atone-half horizontal line intervals. The dura-tion at the sync level is 3.575± 0.425 percentof the horizontal line time (0.45 to 0.5 timesthe horizontal sync). There are three hor-izontal lines of equalizing pulses (6 pulses)preceding the vertical synchronization signaland 3 horizontal lines of equalizing pulses (6pulses) following the vertical synchroniza-tion interval. The equalizing pulse intervalrepeats every field of 262.5 horizontal lines.

equalizing pulses in an encoded videosignal, a series of 2× line frequency pulsesoccurring during vertical blanking, beforeand after the vertical synchronizing pulse.Different numbers of equalizing pulses areinserted into different fields to ensure thateach field begins and ends at the right timeto produce proper interlace. The 2× line ratealso serves to maintain horizontal synchro-nization during vertical blanking.

equiband a complex method of chromi-nance signal demodulation, that does not re-quire dissimilar filters and delay equilization;it most frequently utilizes a bandwidth of 800KHz.


equilibrium manifold Seesliding surface.

equilibrium point for a continuous-timesystemdx/dt = f (t, x), a constant solutionx∗ such thatf (t, x∗) = 0. For a discrete-time systemx(k + 1) = f (k, x(k)), a con-stant solutionx∗ such thatx∗ = f (k, x∗).

equilibrium solution consider a dynamicsystem described by a first-order vector dif-ferential equation of the form

x(t) = f (x(t), t, u(t))

If there exists a vectorxe such that

f (xe, t,0) = 0 ∀t

then wheneverx(t) = xe, the system is saidto be in equilibrium andxe is called an equi-librium state.

equilibrium state Seeequilibrium solution.

equipment hazard a possible source ofperil, danger, risk, or difficulty.

equiripple filter a filter designed using aniterative algorithm, for example, Remez ex-change algorithm, to minimize the maximaldeviation from the desired magnitude fre-quency response in both passband and stop-band for a given filter length. It has equalamplitude ripples within the passband andstopband.See alsoChebyshev filter.

equivalence in continuous-valued logicsimilar to equivalence in conventional logicequivalence in continuous-valued logic be-tween two variablesx andy, which are con-tinuous in the open interval(0,1), can be de-fined as

e(x, y) =max{min(x, y),min((1 − x), (1 − y))} .

Seeequivalence in logic.

equivalence in logic for two Boolean vari-ablesx andy, defined as

(x ≡ y) = (x ∧ y) ∨ (x ∧ y)

equivalence theorem an electromagnetictheorem: If the tangential magnetic and elec-tric fields are known everywhere on someclosed surfaceS, then these fields may be re-placed with equivalent electric and magneticsurface currents, respectively. These equiv-alent currents will produce the same fieldstructure exterior toS as the original fieldsand the null field internal toS.

equivalent circuit a combination of elec-tric circuit elements chosen to represent theperformance of a machine or device by es-tablishing the same relationships for voltage,current, and power.

equivalent control an algorithm usedto determine a system’s dynamics when re-stricted to a sliding surface. The methodentails combining the solution to an alge-braic equation involving the time derivativeof the function describing the sliding surfaceand the dynamical system’s model.See alsovariable structure systemand sliding modecontrol.

equivalent current a theoretical currentused to obtain the scattered field from a sur-face or discontinuity. The equivalent currentis formulated to represent the actual physi-cal currents so as to result in an equivalentscattered field.

equivalent impedance the impedance ofthe windings of an electromagnetic machinereflected to one side (component) of the ma-chine. For example, in a transformer, theequivalent impedance consists of the com-bined leakage reactances and resistances ofthe primary and secondary.

equivalent noise current (ENI) a noisecurrent source that is effectively in paral-lel with either the noninverting input termi-


nal (ENI+) or the inverting input terminal(ENI−) and represents the total noise con-tributed by the op amp if either input terminalis open circuited.

equivalent noise temperature an alter-native way of describing the noise propertiesof two port networks.

equivalent noise voltage (ENV) a noisevoltage source that is effectively in serieswith either the inverting or noninverting in-put terminal of the op amp and represents thetotal noise contributed by the op amp if theinputs were shorted.

equivalent reactance the reactance of thewindings of an electromagnetic machine re-flected to one side (component) of the ma-chine.See alsoequivalent impedance.

equivalent resistance the resistance of thewindings of an electromagnetic machine re-flected to one side (component) of the ma-chine.See alsoequivalent impedance.

equivalent source fictitious source usedin the equivalence theorem.

equivalent sphere illumination (ESI) thelevel of sphere illumination that would pro-duce task visibility equivalent to that pro-duced by a specific lighting environment.

equivalent system dynamics a dynamicalsystem model resulting from substituting theequivalent controlinto the plant’s modelingequation. The equivalent system’s trajectoryis confined to a surface that is parallel to thesliding surface if the system’s initial condi-tion is off the sliding surface. If the initialcondition is on the sliding surface, then theequivalent system’s trajectory will stay on thesliding surface.

erasable optical disk a magneto-opticaldisk that can be both read/written anderased. A thermo-magneto process is usedfor recording and erasure of information.

The recording process uses e.g., the “laserpower modulation” or the “magnetic fieldmodulation” technique.

erasable programmable read-only mem-ory (EPROM) a nonvolatile chip mem-ory, it is used in the place of PROM. EPROMspresents a glass on the case that allows oneto see the chip. They can be erased by ex-posing the chip at the ultraviolet light fortypically 20 minutes. Once erased they canbe reprogrammed. The programming hasto be performed by using a special algo-rithm and a supplementaryVpp. Also calledUVPROM. See alsoelectronically erasableprogrammable read-only memory.

erasure in a forward error control sys-tem, a position in the demodulated sequencewhere the symbol value is unknown. De-pending on the quality of the received signalwhen no decision is made on a particular bit,the demodulator inserts an erasure in the de-modulated data. Using the redundancy in thetransmitted data, the decoder attempts to fillin the positions where erasures occurred.Seebinary erase channel.

erbium doped fiber amplifier (EDFA)an optical amplifier based on the rare-eartherbium that can amplify optical signals inthe 1550 nm telecommunications window.The useful gain region extends from approx-imately 1530 nm to 1565 nm.

EREW Seeexclusive reads and exclusivewrites.

ergodic process a process that has all pos-sible ergodic properties.

For processX(t) when the time average

limT→∞ (1/T )

∫ T/2

−T/2X(t)dt

exists and equals the corresponding expectedvalueE [X(t)], it is said thatX(t) is ergodicin the mean. There are ergodic properties as-sociated with the mean, autocorrelation, and


power spectral density as well as all finite-order joint moments.

ergodicity stochastic processes for whichensemble averages can be replaced by tem-poral averages over a single realization aresaid to be ergodic. For a stochastic pro-cess to be ergodic, a single realization mustin the course of time take on configurationsclosely resembling the entire ensemble ofprocesses. Stationary filtered white noise isconsidered ergodic, while the sinusoidal pro-cessA cos(wt +φ)with random variablesAandφ is not.

Erlang B a formula (or mathematicalmodel) used to calculate call blocking prob-ability in a telephone network and in partic-ular in cellular networks. This formula wasinitially derived by A. K. Erlang in 1917, aDanish pioneer of the mathematical model-ing of telephone traffic, and is based on theassumption that blocked calls are forever lostto the network.See alsoErlang C.

Erlang C similar to Erlang B, the for-mula is based on a traffic model where thecall arrival process is modeled as a Poissonprocess, the call duration is of variable lengthand modeled as having an exponential dis-tribution. The system is assumed to have aqueue with infinite size that buffers arrivingcalls when all the channels in the switch areoccupied. The model is based on the assump-tion that blocked calls are placed in the queue.

Erlang capacity maximum number ofusers in the system which leads to the maxi-mum allowable blocking probability (for ex-ample, 2%).

erosion an important basic operation inmathematical morphology. Given a structur-ing elementB, the erosion byB is the oper-ator transformingX into the Minkowski dif-ferenceX B, which is defined as follows:1. If bothX andB are subsets of a spaceE,

X B = {z ∈ E | ∀b ∈ B, z+ b ∈ X}

2. If X is a gray-level image on a spaceE andB is a subset ofE, for everyp ∈ E we have

(X B)(p) = infb∈B X(p + b)

3. If bothX andB are gray-level images ona spaceE, for everyp ∈ E we have

(X B)(p) = infh∈E

[X(p + h)− B(h)

]with the convention∞ − ∞ = +∞ whenX(p + h), B(h) = ±∞. (In the two itemsabove,X(q) designates the gray-level of thepointq ∈ E in the gray-level imageX.) Seedilation, structuring element.

ERP Seeeffective radiated power.

error (1) manifestation of a fault at log-ical level. For example, a physical short orbreak may result in logical error of stuck-at-0or stuck-at-1 state of some signal in the con-sidered circuit.

(2) a discrepancy between a computed,observed, or measured value or condition andthe true, specified, or theoretically correctvalue or condition. See bug,exception.

error control coding Seechannel coding.

error-correcting code (ECC) code usedwhen communication data information in andbetween computer systems to ensure correctdata transfer. An error correcting code hasenough redundancy (i.e., extra informationbits) in it to allow for the reconstruction ofthe original data, after some of its bits havebeen the subject of error in the transmission.The number of erroneous bits that can be re-constructed by the receiver using this codedepends on the Hamming distance betweenthe transmitted codewords.See alsoerrordetecting code.

error correction capability of a code isbounded by the minimum distance and for an(n, k) block code, it is given byt = [(dmin−


1)/2], where[x] denotes the largest integercontained inx.

error detecting code code used whencommunication data information in and be-tween computer systems to ensure correctdata transfer. An error detecting code hasenough redundancy (i.e., extra informationbits) in it to allow for the detection of theoriginal data, after some of its bits have beenthe subject of error in the transmission. Thenumber of erroneous bits that can be detectedby the receiver using this code depends on theHamming distance between the transmittedcodewords. Seealso error-correcting code.

error detection the process of detectingif one or more errors have occurred during atransmission of information. Channel codesare suitable for this purpose. The family ofCRC-codes are an example of channel codesspecially designed for error detection.Seealsoerror detecting code.

error detection capability the capabilityof a code to detect error is bounded by theminimum distance, and for an(n, k) blockcode, it is given bydmin − 1.

error extension the multiplication of er-rors that might occur during the decoding of aline coded sequence, or during the decodingof a forward error control coded sequencewhen the number of symbol errors exceedsthe error correction capability of the code.

error function mathematical functionover some interval in which a calculated re-sult is compared to a known quantity (usuallydata) by utilizing a difference quantity to de-termine how well the mathematical functionreplicates the known quantity over that inter-val. Most error functions utilize an area orleast squares difference function.

error latency length of time between theoccurrence of an error and the appearance ofthe resulting failure.

error recovery process of regaining oper-ational status and restoring system integrityafter the occurrence of an error with the useof special hardware and software facilities.

error state diagram a diagram that il-lustrates all possible error events with re-spect to an assumed reference event, e.g., anerror sequence compared to a reference se-quence, both produced by a finite state ma-chine. All possible sequences that divergeand later merge with the reference sequenceare accounted for in the diagram. The trans-fer function of the error state diagram canbe used for performance evaluation of detec-tion/decoding algorithms working on noiseoutputs from a finite state machine.

error-correction the mechanism bywhich a receiving circuit is able to cor-rect errors that have occurred in an encodedtransmission. See alsoHamming code,error-correcting code.

ESD Seeelectrostatic discharge.

ESDI Seeenhanced small disk interface.

ESI Seeequivalent sphere illumination.

ESPRIT acronym for estimation of sig-nal parameters via rotational invariance tech-niques. A subspace-based estimation tech-nique based on two identical, displaced sen-sor arrays.

estimation in adaptive control, a key roleis played by on-line determination of processparameters. A recursive parameter estimatoris present in many adaptive control schemasuch as self-tuning regulator (explicitly) ormodel-reference adaptive controller (implic-itly).

Usually, parameters estimation is viewedin the broader context of system identifi-cation formed by selection of model struc-ture, experiment design, parameters estima-tion, and validation. In adaptive control, theparameters vary continuously, and it is nec-


essary to estimate them recursively. A basictechnique used for parameters estimation isthe Least Squares Method. This method isparticularly useful if the model has the prop-erty of being linear in the parameters.

estimator any function of the samplepoints. Hence, an estimator is a random vari-able. To be useful, an estimator must have agood relationship to some unknown qualitythat we are trying to determine by experi-ment.

etching a reactive process where mate-rial is removed from a semiconductor deviceor printed circuit board. Usually a photo-sensitive material is exposed through a pho-tomask, and either a wet chemical process ora dry plasma process is used to selectivelyremove material to leave a particular patternbehind after the etch process is completed.

Ethernet a standard for interconnectingdevices on a local area network (LAN).

Euclidean distance a distance measurebetween two real valued vectors(x1, x2, . . . ,

xn) and(y1, y2, . . . , yn) defined as

DEuclidean=√√√√ n∑

i=1

(xi − yi)2

Euclidean distance is the special case ofMinkowski distance whenλ = 2. See alsoMinkowski distance.

Euler number a topological invariant ofan object having an orientable surface. As-suming that the surface is endowed with thestructure of a graph with vertices, edges, andfaces (where two neighboring faces have incommon either a vertex or an edge with itstwo end-vertices, their interiors being dis-joint): the Euler number isV − E + F ,whereV ,E, andF are respectively the num-ber of vertices, edges and faces; this numberV − E + F does not depend on the choiceof the subdivision into vertices, edges, andfaces. For a bounded 2-D object in a Eu-

clidean or digital plane, the Euler numberis equal to the number of connected com-ponents of that object, minus the number ofholes in it. For 2-D binary digital figures ona bounded grid, the Euler number can easilybe computed by counting the number of oc-currences of some local configurations of onand off pixels. Also called genus.

eureka in a multiprocessor system, a co-ordination (synchronization) operation gen-erating a completion signal that is logicallyORed among all processors participating inan asynchronously parallel action. The in-terpretation and name come from its use insystems that delegate various portions of asearch space to different processing modules.A processor executes the eureka operationwhen it has found the desired object or value;this can serve as a signal for the others to aborttheir attempts to find a solution.

eutectic alloy composition with minimummelting temperature at the intersection of twosolubility curves.

eutectic alloy overload device an over-load device that employs a melting alloy asthe actuating element.See alsooverloadheater, overload relay.

EUV lithography lithography using lightof a wavelength in the range of about 5 to50 nm, with about 13 nm being the most com-mon. Also called soft X-ray lithography.

even function a real-valued functionx(t)in which x(−t) = −x(t) for all values oft .Compare withodd function.

even mode impedance characteristicimpedance of a transmission line when a sin-gle and certain even mode exists on it.

even order response a circuit gain orinsertion loss versus frequency response inwhich there are an even number of peaks inthe ripple pattern, due to an even number ofpaired elements in the circuit. Even order cir-


cuits exhibit a peak for each element pair anda loss equal toLamax (maximum attenuationloss across the band) at DC (low pass) or atω0 (band pass).

even signal a signal that has even sym-metry. If x(t) is an even signal, it satisfiesthe conditionx(t) = x(−t). See alsooddsignal.

even-mode characteristic impedancecharacteristic impedance of a circuit due toan even-mode current or voltage excitation.Often applied in the context of a transmis-sion line coupler where the even-mode ex-citation consists of applying equal amplitudevoltages or currents of identical phase on twoconductors. The resulting impedance underthis excitation is defined as the even modecharacteristic impedance.

even-order response circuit gain or inser-tion loss versus frequency response in whichthere are an even number of peaks in the rip-ple pattern, due to an even number of pairedelements in the circuit. Even-order circuitsexhibit a peak for each element pair and aloss equal toLamax (maximum attenuationloss across the band) at DC (low pass) or atωo (band pass).

event (1) a nonsequential change in thesequencing of macroinstructions in a com-puter. Events can be caused by a variety offactors such as external or internal interrupts(traps) or branch statements.

(2) a specific instance taken from somesample space, normally with an associatedprobability or probability density; also com-monly an idealized infinitesimal point in(x, y, z, t) space at which some occurrence istaken to happen. For example, a flash of lightat time t0 at position(x0, y0, z0) — event(x0, y0, z0, t0) — will lead at a later timet1to a wave of light passing a point(x1, y1, z1)

— namely event(x1, y1, z1, t1).

event table table listing all events, andtheir corresponding effects as well as re-

actions to them. Exception conditions/re-sponses table is a special type of event table.

exact absolute controllability a dy-namical system where the attainable setK∞ is equal to the whole state spaceW(2)1 ([−h,0], Rn).

exact coding coding methods that repro-duce the picture at the receiver without anyloss. This method is also called information-lossless or exact coding techniques. Fourmethods of exact coding are run-length cod-ing, predictive coding, line-to-line predictivedifferential coding, block coding. However,several coding schemes use these in a hybridmanner.

exact controllability of infinite dimen-sional system an infinite dimensional sys-tem where the attainable setK∞ is equal tothe whole state spaceX.

exception (1) an unusual condition aris-ing during program execution that causes theprocessor to signal an exception. This signalactivates a special exception handler that isdesigned to handle only this special condi-tion. Division by zero is one exception con-dition. Some vendors use the term “trap” todenote the same thing.

(2) an event that causes suspension of nor-mal program execution. Types include ad-dressing exception, data exception, operationexception, overflow exception, protection ex-ception, underflow exception.

exception handler a special block of sys-tem software code that reacts when a spe-cific type of exception occurs. If the excep-tion is for an error that the program can re-cover from, the program can recover from theerror and resume executing after the excep-tion handler has executed. If the programmerdoes not provide a handler for a given excep-tion, a built-in system exception handler willusually be called, which will result in termi-nating the process that caused the exception.Finally, the reaction to exception can be halt-


ing of the system. As an example, a bus errorhandler is the system software responsible forhandling bus error exceptions.

excess delay the arrival times of a compo-nent of the impulse response of a widebandcommunication channel relative to the firstarriving component. Hence the total excessdelay, the difference in arrival time betweenthe first and last significant components.

excess loss the ratio of the actual propa-gation loss between two antennas to the freespace loss for two antennas separated by thesame distance in a vacuum. Usually ex-pressed in decibels.

excess noise (1) thermal noise in excessor exceeding thermal noise at 290◦K. The ex-cess noise ratio (ENR) is the ratio of the ex-cess noise to the noise at 290◦K, expressedin decibels.

(2) Noise in excess of the thermal noise(ne), which is a function of the device, fre-quency and bias current, also known as 1/f ,1/fa , flicker or popcorn noise. This noiserapidly dies off such that above a few mega-hertz it becomes insignificant. It is a majorconcern in generating phase noise in oscilla-tors.

ENR = 10 log10

(T − 290◦K

290◦K

)ne = f

(device, 1/f a, Ibias

)excess noise ratio (ENR) a noise sourceused in noise figure measurements. ENR isthe ratio of the source’s noise power when it ison to the noise power when it is off. The ENRvalues are entered and stored in the noise fig-ure meter to calibrate a measurement.

excess-N representation method of rep-resenting floating point numbers, in which apositive or negative exponent is stored as apositive integer by adding the value N to it.For example, in excess-128 representation,an exponent of−40 would be stored as thevalue 88.

exchangeable disk Seeremovable disk.

excimer a molecule formed by the excitedstate of one atom and one or more other atomsthat remains a molecule only as long as theexcited state lifetime.

excimer laser laser using a gas or gasesto create an excited dimer (e.g., KrF), usuallyresulting in pulsed deep-UV radiation.

excitation population of excited states ofa laser medium at the expense of some energysource.

excitation system the DC voltage sourceand its accompanying control and protectionsystems connected to the synchronous gen-erator rotor.

exciter a DC source that supplies the fieldcurrent to produce a magnetic flux in an elec-tric machine. Often it may be a small DCgenerator, placed on the same shaft of theelectrical machine.

exciting current the current drawn by atransformer primary with its secondary opencircuited. It is the vector sum of the core losscurrentIc and the magnetizing branch cur-rent Im. The exciting currentIe is also thecurrent measured in the open circuit test ona transformer. The exciting current is cal-culated as the ratio of the primary inducedEMF and the impedance of the tank circuit.On load, it is equal to the difference betweenthe primary and reflected secondary currentsof the transformer.

exciton laser laser (or laser-like system)in which the amplified field consists of ex-citons rather than electromagnetic waves orphotons.

exclusive OR Boolean binary operatortypically used for comparing the status oftwo variables or signals. Sometimes written“XOR.” The truth table for⊕ ≡ X XOR Y


is as follows:

X Y X ⊕ Y

F F F

F T T

T F T

T T F

exclusive reads and exclusive writes(EREW) shared memory model, in whichonly exclusive reads and exclusive writes areallowed.

execution cycle sequence of operationsnecessary to execute an instruction.

execution time amount of time it takesa computation (whether an instruction or anentire program) to complete, from beginningto end. Time during which an instruction or aprogram is executed. The portion of one ma-chine cycle needed by a CPU’s supervisory-control unit to execute an instruction.

execution unit in modern CPU imple-mentations, the module in which actual in-struction execution takes place. There maybe a number of execution units of differenttypes within a single CPU, including inte-ger processing units, floating point process-ing units, load/store units, and branch pro-cessing units.

exhaustive search a search through allpossibilities before deciding on what actionto take. For example, for the maximum-likelihood (exhaustive search) detection ofa sequence ofk bits, all 2k possible bit se-quences are considered and the one with thelargest likelihood is selected.

expanded memory expanded memoryspecification, EMS, was born for addingmemory to PCs (the so-called LIM-EMS).PCs were limited in memory to 640 kb evenif the 8088/8086 CPU’s limit is 1 Mb. Thus,in order to overcome this limit, the additionalmemory was added by using a paging mech-anism: up to four windows of 16 kb of mem-

ory included into the 640 kb to seen up to8 Mb of memory divided into pages of 16 kb.To this end, special memory boards werebuilt. Currently, the MS-DOS is still limitedto 640 kb, but the new microprocessors canaddress even several gigabytes over the firstmegabyte. Thus, to maintain the compatibil-ity with the previous version and the adoptionof the MS-DOS, the presence of expandedmemory is simulated by means of specificdrivers.

expectation the integral of a function withrespect to some probability measure. Iff (·)is a deterministic function andx is random,governed by probability densityp(x), then

E[f (x)] =∫ ∞

−∞f (x)p(x)dx.

expected value of a random variable en-semble average value of a random variablethat is given by integrating the random vari-able after scaling by its probability densityfunction (weighted average) over the entirerange.

expert system a computer program thatemulates a human expert in a well-boundeddomain of knowledge.

explicit value a value associated with thebit string according to the rule defined by thenumber representation system being used.

explosion-proof machine National Elec-trical Manufacturers Association (NEMA)classification describing an electrical ma-chine that is totally enclosed and whose en-closure is designed to withstand an internalexplosion of a specified gas or vapor that mayaccumulate within the enclosure. The speci-fication also requires that the design preventignition of the specified gas or vapor sur-rounding the enclosure due to sparks, flashes,or explosions of the specified gas within theenclosure.


exponent (1) the field within a floating-point format that determines the power towhich the mantissa should be raised.

(2) a shorthand notation for representingrepeated multiplication of the same base. 24

is exponential notation to multiply two byitself four times : 22 = 2 · 2 · 2 · 2 = 16. 4is called the exponent, indicating how manytimes the number 2, called the base, is usedas a factor.

(3) the component of a binary floating-point number that signifies the integer powerof two by which the significand is multipliedin determining the value of the representednumber.

exponential distribution a probabilitydensity function having the following expo-nential behavior:

f (x) ={λe−λx x ≥ 0

0 x < 0

where λ > 0. This distribution can de-scribe a number of physical phenomena, suchas the time for a radioactive nucleus to de-cay, or the time for a component to fail.See alsoprobability density function,Cauchydistribution, Gaussian distribution.

exponential stability (1) the property ofan asymptotically stable equilibrium solutionthat guarantees an exponentially decreasing(to zero) norm in time of the difference be-tween the solution and the equilibrium point.

(2) a special case of uniform asymptoticstability of an equilibrium point ofx =f (t, x).

exposure the process of subjecting a resistto light energy (or electron energy in the caseof electron beam lithography) for the purposeof causing chemical change in the resist.

exposure dose Seeexposure energy.

exposure energy the amount of energy(per unit area) that the photoresist is subjectedto upon exposure by a lithographic exposuresystem. For optical lithography, it is equal

to the light intensity times the exposure time.Also called exposure dose.

exposure field the area of a wafer that isexposed at one time by the exposure tool.

express feeder a feeder to which lateralsare connected only at some distance from thesubstation. These thus traverse areas fed byother feeders and are used to supply concen-trated loads or new subdivisions.Seefeeder,lateral.

expulsion fuse a fuse used on primary dis-tribution lines which extinguishes the arc thatresults when it blows by explosively ejectingthe fuse wire from its enclosure.

expulsion tube arrester a gapped light-ning arrester which establishes the power-follow arc in a tube lined with a substancewhich generates a sufficient quantity of gaswhen heated to blow out the arc.Seepowerfollow, lightning arrestor.

expurgated code a code constructed fromanother code by deleting one or more code-words from the original code.

extended binary-coded-decimal inter-change code (EBCDIC) character codedeveloped by IBM and used in mainframecomputers. It is closely related to the Hol-lerith code for punched cards.

extended code a code constructed fromanother code by adding additional symbolsto each codeword. Thus an(n, k) originalcode becomes an(n + 1, k) code after theadding of one redundant symbol.

extended industry standard architecture(EISA) a bus architecture designed forPCs using an Intel 80386, 80486, or Pentiummicroprocessor. EISA buses are 32 bits wideand support multiprocessing. The EISA buswas designed by IBM competitors to com-pete with micro channel architecture (MCA).EISA and MCA are not compatible with each


other, the principal difference between EISAand MCA is that EISA is backward compat-ible with the ISA bus (also called the ATbus), while MCA is not. Therefore, com-puters with an EISA bus can use new EISAexpansion cards as well as old AT expansioncards, while computers with an MCA bus canuse only MCA expansion cards.

extended Kalman filter state estimationmethod based on linearization (see the defi-nition) of nonlinear system and measurementequations about the current state estimate.The forms of both the filter gain and esti-mation error covariance equations are similarto those in the Kalman filter (see the defini-tion). However, due to the linearization aboutthe current state estimate, both equations aredependent on the current state estimate andcannot be calculated off-line.

extended memory in PCs, the memorylocated over the first megabyte of memory.This kind of memory is seen in the PC ascontinuous. Operating systems such as Win-dows NT, LINUX, and OS/2 are capable ofworking in protected mode, and thus at 32bits. When MS-DOS is adopted to allow theexploitation of this memory as EMS, specialsoftware drivers have to be used. This over-comes the limit of 640 kb imposed by thereal-mode and adopted by MS-DOS.

extended source a (light) source in whichrays are emitted from a large source area.Compare withpoint source.

extended space for a space of func-tionsX , another space of functions, denotedXe, defined as the space of all functionswhose truncation belongs toX . See alsotruncation.

extended storage Seesolid state disk.

extension principle a basic identity forextending the domain of nonfuzzy or crisprelations to their fuzzy counterparts.

external cavity klystron a klystron de-vice in which the resonant cavities are locatedoutside the vacuum envelope of the tube.

external event event occurring outside theCPU and I/O modules of a computer systemthat results in a CPU interrupt. Examples in-clude power fail interrupts, interval timer in-terrupts, and operator intervention interrupts.

external fragmentation in segmentation,leaving small unusable areas of main mem-ory that can occur after transferring segmentsinto and out of the memory.

external interrupt a signal requesting at-tention that is generated outside of the CPU.

external memory secondary memory ofa computer.

external modulation modulation of theoptical intensity using an optical intensitymodulator to modulate a constant powerlaser.

external space the space of allowable po-sitions and orientations of the end-effectorof the manipulator. In general the exter-nal space, with the special Euclidean groupSE(3) ∼= SO(3)× R3, consists of rotations(SO(3))and translations(R3) inR3. The ex-ternal space is called as Cartesian space, test-oriented space, or operational space by manyroboticians. Some people mean by Carte-sian space the space in which the position ofa point is given with three numbers, and inwhich the orientation of a body is given bythree numbers. This is because in order toposition and orient a body in Cartesian spacewe need 6 coordinates. In general orienta-tion is specified by 3× 3 orientation matrixthat forms orthonormal vectors. Orientationcan be specified in terms of a minimal repre-sentation describing the rotation of the end-effector frame with respect to the referenceframe, e.g., Euler angles or rotation aboutan arbitrary axis and equivalent angle. Thenmanipulator position and orientation ism×1


vector, withm ≤ n (n denotes number of de-grees of freedom). Using minimal represen-tation of the orientation matrix, direct kine-matics equation can be written in the follow-

ing form: x =[φ

p

]= k(q) whereφ is a set

of vectors for minimal representation andpdenotes three position coordinates.

external stability stability concepts re-lated to the input–output behavior of the sys-tem.

externally vented machine classificationdescribing an electrical machine constructedwith an open frame in which ventilation airis forced through the machine by blower(s)mounted outside the machine enclosure.

extinction angle time in electrical degreesfrom the instant the current in a valve reacheszero (end of conduction) to the time the valvevoltage changes sign and becomes positive.

extinction cross section the sum of thescattering and the absorption cross sections.

extrapolation one of several methods toestimate the values of a sequencer(k) forlags |k| > p from the given values ofr(k)for |k| ≤ p.

extrinsic associated with the outside orexterior. In devices and device modeling,extrinsic refers to that part of the device ormodel associated with the passive structuresthat provide interconnects and contacts toother components, but are still considered apart of the device.

extrinsic fiber optic sensor a fiber opticsensor where the fiber delivers light to andfrom a sensing element external to the fiber.Chemical sensors are an example where thesensing element exhibits a change in opti-cal property such as absorption, fluorescenceor phosphorescence upon detection of thespecies to be measured.


FfH common notation for higher band edgefrequency in hertz.

fL common notation for lower band edgefrequency in hertz.

Fabry–Perot etalon interferometer con-sisting of two highly reflecting flat or spher-ical mirrors; only resonant frequencies aretransmitted.

Fabry–Perot laser a laser source wherethe gain medium is placed within a Fabry–Perot cavity, which provides feedback intothe laser medium. Several simultaneous las-ing modes are supported in such cavities.

Fabry–Perot resonator any open (as op-posed to a cavity) resonator, usually the as-sembly of two parallel plates resembling theoptical Fabry–Perot interferometer.See alsostanding-wave resonator.

Fabry–Perot structure Fabry–Perotetalon or interferometer that has an opticallynonlinear medium in its cavity.

facsimile the process of making an ex-act copy of a document through scanning ofthe subject copy, electronic transmission ofthe resultant signals modulated by the subjectcopy, and making a record copy at a remotelocation.

facsimile encoding a bilevel coding meth-od applied to the encoding and transmissionof documents. Facsimile systems may in-clude support for grayscale image coding too,which is described under still image coding.

FACTS See flexible AC transmissionsystem.

fading channel signal fluctuation causedby multiple propagation paths over a radiochannel is called fading. May be categorizedas fast fading or slow fading, depending onthe rate of fading with respect to the informa-tion symbol rate. May also be categorized asfrequency selective or frequency nonselec-tive, depending on the transfer function ofthe radio channel.See alsofading margin,fading Rayleigh, fading Rician.

fading margin the margin by which theaverage signal-to-noise ratio (SNR) in a ra-dio communications link is over-designed,in order to compensate for variations in theshort-term SNR that occur due to fading ofthe signal. The signal fading is typically dueto multipath propagation, which arises dueto the presence of multiple reflectors in theradio link. Utilizing diversity schemes andtransmitter power control are typically em-ployed in order to reduce the fading marginrequired in a radio system.

fading rate the rate at which the receivedsignal level crosses the median signal levelin a downward direction (i.e., with a negativeslope). It is usually expressed in fades persecond or fades per minute, depending on theactual rate of fading.

fading Rayleigh in mobile wireless com-munications, wide fluctuations in receivedsignal strength (e.g., swings of 30 – 40 dB)and phase caused by scattering of the trans-mitted signal off of surrounding objects. Thescattering induces a Gaussian distribution onthe in-phase and quadrature signal compo-nents, so that the received signal envelope hasa Rayleigh distribution.See alsoRayleighdistribution.

fading Rician similar to Rayleigh fading,the only difference being that a direct line-of-sight component is present in the receivedsignal in addition to the scattered signal. Thereceived signal envelope has a Rician distri-bution. See alsoRice distribution.


fail safe pertaining to a circuit, for a set offaults, if and only if for any fault in this set andfor every valid input code either the output iscorrect or assumes some defined safe state.

fail-stop processor a processor that doesnot perform incorrect computation in theevent of a fault. Self-checking logic is of-ten used to approximate fail-stop processing.

failure manifestation of an error at systemlevel. It relates to execution of wrong actions,nonexecution of correct actions, performancedegradation, etc.

failure mechanism a physical or chem-ical defect that results in partial degradationor complete failure of a product.

fairness (1) the concept of providingequivalent or near-equivalent access to ashared resource for all requestors.

(2) a fair policy requires that tasks,threads, or processes are allowed access toa resource for which they compete.

(3) the degree to which a scheduling or al-location policy is equitable and nondiscrimi-natory in granting requests among processescompeting for access to limited system re-sources such as memory, CPU, or networkbandwidth.

fall time (1) in digital electronics, the pe-riod of duration of the transition of a digitalsignal from a stable high-voltage level to astable low-voltage level.

(2) in optics, the time interval for thefalling edge of an optical pulse to transitionfrom 90% to 10% of the pulse amplitude. Al-ternatively, values of 80% and 20% may beused.

falling edge (1) the region of a waveformwhen the wave goes from its high state to itslow state.

(2) the high-to-low transition in voltage ofa time-varying digital signal.

false color the replacement of a color ina colored image by a different color, usuallynot present in the original image. Used tohighlight regions or distinguish pixels of sim-ilar colors.See alsopseudo color.

false sharing the situation when morethan one processor accesses different parts ofthe same line in their caches but not the samedata words within the line. This can causesignificant performance degradation becausecache coherence protocols consider the lineas the smallest unit to be transferred or inval-idated.

fan beam reconstruction reconstructionof a computed tomography image from pro-jections created by a point source that emitsa fan- or wedge-shaped beam of radiation.Fan beam reconstruction enables data to begathered much more quickly than by usinga linear beam to produce parallel projec-tions.See alsocomputed tomography, imagereconstruction, projection, Radon transform.

fan-in (1) the number of inputs to a mod-ule. This is usually used in connection withlogic gates.

(2) multiple inputs of a channel. If thechannel is a bus, only an input is allowedat a time. WhenN light beams are com-bined with an optical element, only 1/N ofthe power of each beam finds its way into thecombined beam. Large fan-in is quite im-practical for VLSI, because it uses a uniquediscrete channel for each input and currentflows are limited by transistor capacitance.

fan-out (1) the limit to the number of load-ing inputs that can be reliably driven by adriving device’s output.

(2) multiple outputs of a bus. A signalis distributed into multiple channels. With afan-out ofN , each channel receives only 1/Nof the light power. Large fan-out is quite im-practical for VLSI, because it uses a uniquediscrete channel for each output and currentflows are limited by transistor capacitance.


Fano algorithm a sequential decoding al-gorithm for decoding of trellis codes.

Fano mode a bound nonradiative surfacemode that propagates along an interface anddecays in a nonoscillatory manner in a direc-tion perpendicular to the propagation. Sucha mode occurs when one of the media is aplasma medium and has a negative dielectricfunction. A metal at optical frequency is anexample of such a medium.

Fano’s inequality information theoreticinequality bounding the probability of incor-rectly guessing the value of one random vari-able based on observation of another. IfPeis the probability of incorrectly guessing arandom variableX ∈ X , based upon obser-vation of the random variableY , then

H (Pe)+ Pe log(|X − 1) ≥ H(X|Y )

Named after its discoverer, R. M. Fano(1952). Used in proving the weak converse tothe channel coding theorem (Shannon’s sec-ond theorem).

Fano’s limit theoretical limit relating theachievable gain and bandwidth of a givenpassive lossless matching network when ter-minated in an arbitrary load impedance.

far field (1) that region of space in whichthe electric field and magnetic field compo-nents of an electromagnetic wave are relatedby the impedance of free space. The far fieldis generally considered to begin no closerto the NIER source than a distance of sev-eral wavelengths or several times the antennaaperture.

(2) that region of the field of a certainsource where the angular field distributiondepends in a known way from the distanceof the source. Generally, in free space, if weconsider as a source an antenna with max-imum overall dimensionD, assumed largecompared to the wavelength, the far-field re-gion is commonly taken to exist at distancesgreater than 2D2/λ from the antenna,λbeing

the wavelength. Also called the Fraunhoferregion.

far pointer a pointer to a far segment. In80 × 86 architecture, a far pointer specifiesthe segment address and the offset.

far-field pattern graph or chart repre-senting the absolute or normalized antennagain as a function of angle (typically azimuthor elevation) and used to describe the direc-tional properties of an antenna in the far field(Fraunhofer region).

far-infrared (FIR) spectral region oftenconsidered to range from about 10 to 100 mi-crometers.

farad the basic unit of measure in capac-itors. A capacitor charged to 1 volt with acharge of 1 coulomb (1 ampere flowing for1 second) has a capacitance of 1 farad.

Faraday effect the rotation of the plane ofpolarization of a high-frequency signal (mi-crowave RF, optical field) in the presence ofa magnetic field.

Faraday rotation (1) rotation in the di-rection of polarization experienced by a wavetraveling through an anisotropic medium.Important examples of media in which thephenomenon occurs include the earth’s iono-sphere and ferrites biased by a static magneticfield.

(2) depolarization caused in a plasma(e.g., the ionosphere) resulting from interac-tion between the ions of the plasma and themagnetic field of the wave.

(3) rotation in the polarization vectorexperienced by a wave after it propagatesthrough a gyromagnetic medium.

Faraday rotator a magneto-optical de-vice that changes the orientation plane ofpolarized light when it passes parallel to amagnetic field through a substance with pro-nounced absorption lines.


Faraday shield an electrostatic (E field)shield made up of a conductive or partiallyconductive material or grid. A Faraday cageor screen room is effective for protecting in-side equipment from outside radiated RF en-ergies.

Faraday’s law one of Maxwell’s equa-tions that describes the fundamental relation-ship between induced voltage and a time-varying magnetic field. For a conductingcoil, the induced voltage is proportional to thetime rate of change in the magnetic flux link-ing the coil. This change may be producedeither by actual variation of field strength orby relative motion between coil and field.SeealsoMaxwell’s equations.

Faraday, Michael (1791–1867) Born:Newington, Surrey, England

Faraday is best known as the greatest ex-perimental physicist of the 19th century. Itwas Faraday who invented the electric mo-tor, generator, and transformer, and first de-scribed electromagnetic induction and thelaws of electrolysis. Faraday had no for-mal schooling, although he attended manylectures. The most inspirational of these lec-tures were by the famed chemist Sir HumphryDavy. Faraday became Davy’s assistant andthus began an extraordinary career as an ex-perimentalist. Faraday’s contributions arerecognized by the use of his name as the unitof electrical capacitance, the farad in the SIsystem, and the Faraday constant in electrol-ysis.

farm Seeprocessor farm.

fast Fourier transform (FFT) a compu-tational technique that reduces the numberof mathematical operations in the evaluationof the discrete Fourier transform (DFT) toN log2N .

fast neutrons neutrons emitted fromfission reactions which travel at velocitieshigher than the average velocity of atoms un-der random thermal motion.

fast packet networks networks in whichpackets are transferred by switching at theframe layer rather than the packet layer. Suchnetworks are sometimes called frame relaynetworks. Current thinking views frame re-lay as a service, rather than transmission,technology.

fast reactor a reactor which maintains acritical chain reaction with fast neutrons andwhich does not require a moderator.

father wavelet the scaling function in thecoarsest resolution in wavelet analysis.

fault (1) in hardware, a physical defector imperfection of hardware. Typical circuitfaults are shorts opens in conductor, defectsin silicon, etc.See alsodisturbance.

(2) in software, the manifestation of anerror.

fault avoidance a technique used to pre-vent or limit fault occurrence (for example,with signal shielding, fan-out limitation, andpower dissipation decrease).

fault confinement technique that limitsthe spread of fault effects to some area ofthe system and prevents propagation of theseeffects to other areas.

fault coverage the measure of test qual-ity expressed as the percentage of detectedfaults.

fault detection the process of locating dis-tortions or other deviations from the ideal,typically during the process of automated vi-sual inspection, e.g., in products undergoingmanufacture.

fault indicator a small indicating unitequipped with a permanent magnet and piv-oting pointer which is hung on a transmissionline suffering intermittent faults of unknownorigin. After a fault occurs, fault indicatorsare inspected. Each shows the presence and


direction of a fault, thus allowing the defectto be located.

fault kva fault kilovolt-amps (kva) is thefault level expressed in terms of volt-ampsrather than amps. One advantage of usingvolt-amps rather than amps is that the sameflow is experienced on both sides of a trans-former when expressed in volt-amps, whilethe flow changes due to the transformer turnsratio when it is expressed in amps. Volt-amps for a three-phase fault are expressedas 1.73× rms line-line voltage× rms sym-metrical fault current. Volt-amps for a singlephase fault are defined as 1.73× rms line-line voltage× rms symmetrical current in thefaulted phase.

fault latency the length of time betweenthe occurrence of a fault and the appearanceof an error.

fault masking a technique that hides theeffects of faults with the use of redundantcircuitry or information.

fault mva fault megavolt-amps (mva) isthe fault level expressed in terms of volt-ampsrather than amps. One advantage of usingvolt-amps rather than amps is the same flowis experienced on both sides of a transformerwhen expressed in volt-amps, while the flowchanges due to the transformer turns ratiowhen it is expressed in amps. Volt-amps fora three-phase fault are expressed as 1.73×rms line-line voltage× rms symmetrical faultcurrent. Volt-amps for a single phase fault aredefined as 1.73× rms line-line voltage× rmssymmetrical current in the faulted phase.

fault prevention any technique or processthat attempts to eliminate the possibility ofhaving a failure occur in a hardware deviceor software routine.

fault resistance the resistance that occursat the point of fault due to voltage drop acrossan arc or due to other resistance in the faultpath.

fault secure pertaining to a circuit, withrespect to a set of faults, if and only if forany fault in this set, and any valid input codethe output is a non-code or correct code (theoutput is never an invalid code). The circuitis considered to operate properly if the outputis a code word.

fault simulation an empirical methodused to determine how faults affect the op-eration of the circuit and how much testing isrequired to obtain the desired fault coverage.

fault tolerance correct execution of aspecified function in a circuit (system), pro-vided by redundancy despite faults. The re-dundancy provides the information needed tonegate the effects of faults.

fault tree the identification and analysisof conditions and factors that cause or con-tribute to the occurrence of a defined undesir-able event, usually one that significantly af-fects system performance, economy, safety,or other required characteristics.

fault-tolerant control system a systemthat exhibits stability and acceptable perfor-mance in the presence of component faults(failures) or large changes in the system thatresemble failures.

fax abbreviation for a machine that makesand transmits facsimiles.Seefacsimile.

FCA Seefixed channel assignment.

FCM Seefuzzy cognitive mapor fuzzyc-means.

FCT Seefield controlled thyristor.

FDD Seefrequency division duplex.

FDDI Seefiber distributed data interface.

FDM Seefrequency division multiplexing.


FDMA Seefrequency division multipleaccess.

FDTD Seefinite difference time domain.

feature a measurable characteristic of anobject in an image. Simple examples wouldbe area, perimeter, and convexity. More com-plex features use vectors; examples includemoments, Fourier descriptors, projections,and histogram based features. Features arefrequently used to recognize classes of ob-ject, and sets of simple features can be col-lected into a vector for this purpose. Usingboth area and perimeter, for instance, one canquickly distinguish between a circle and a tri-angle. May also refer to a characteristic ofa whole image: such a feature could then beused in image database analysis.

feature detection the detection of smallerfeatures within an image with a view to in-ferring the presence of objects. This typeof process is cognate to pattern recognition.Typically, it is used to locate products readyfor inspection or to locate faults during in-spection. A feature can be detected by find-ing points having optimal response to a givencombination of local operations such as con-volutions or morphological operators.Seealsoobject detection.

feature extraction a method of trans-forming raw data, which can have very highdimensionality, into a lower dimensional rep-resentation that still contains the importantfeatures of the data.

feature map a fixed geometrical struc-ture (often two dimensional) for unsuper-vised learning that maps the input patternsto different output units in the structure sothat similar input patterns always triggernearby output units topographically.Seealso self-organizing system, self-organizingalgorithm.

feature measurement the measurementof features, with the aim of recognition or

inspection to determine whether products arewithin acceptable tolerances.

feature orientation measurement of theorientation of features, either as part of therecognition process or as part of an inspectionor image measurement process.

feature recognition the process of lo-cating features and determining what typesof features they are, either directly or indi-rectly through the location of sub-featuresfollowed by suitable inference procedures.Typically, inference is carried out by appli-cation of Hough transforms or associationgraphs.

feature size the characteristic size of elec-tronic components on a die.

FEC Seeforward error correction.

feedback (1) signal or data that is sentback to a commanding unit from a controlprocess output for use as input in subsequentoperations. In a closed-loop system, it is thepart of the system that brings back informa-tion on the process condition under control.

(2) the provision of a path from the outputto the input of a system, such that the outputmay be made a function of both the input andthe previous outputs of the system.

(3) the technique of sampling the outputof an amplifier and using that information tomodify the amplifier input signal. A portionof the output is “fed back” to the input. Posi-tive feedback occurs when the output is addedto the input; negative feedback occurs whenthe output is subtracted from the input. Neg-ative feedback, invented by communicationsengineer Harold Black in 1928, usually re-sults in a gain–bandwidth tradeoff: decreas-ing and stabilizing the amplifier gain, whileincreasing the bandwidth. According to Nor-bert Wiener, feedback is a method of control-ling a system by reinserting into it the resultsof its past performance.


feedback amplifier a circuit configura-tion of amplifiers that has a feedback path.A negative feedback configuration is com-monly used in amplifiers for its stable perfor-mance, where a portion of the output signalis added to the input signal at 180 degrees outof phase. There are two types of basic feed-back configurations: parallel feedback and aseries feedback.

feedback control the regulation of a re-sponse variable of a system in a desired man-ner using measurements of that variable in thegeneration of the strategy of manipulation ofthe controlling variable.

feedback decoding a majority-logic de-coding method for decoding of convolutionalcodes.

feedback linearization a method of us-ing feedback to cancel out nonlinearities ina dynamical system model so that the result-ing closed-loop system model is linear. Themethod uses tools from differential geome-try. See alsoLie derivative.

feedback oscillator electronic circuits de-signed to provide specific signals of desiredwaveshape. The feedback oscillator shouldbe envisioned as two basic network subsys-tems: the active portion (amplifier) and thepassive (or feedback, FB) portion. The out-put of one subsystem is connected to the in-put of the other and vice versa. The passivenetwork is lossy but provides the importantfunction of establishing the necessary elec-trical phase shift to establish oscillation atthe desired frequency.See alsoBarkhausencriterion.

feeder overhead lines or cables that areused to distribute the load to the customers.They interconnect the distribution substa-tions with the loads.

feeder circuit an electrical circuit de-signed to deliver power from the serviceequipment or separately derived system to

the branch circuit panelboard(s) on a facil-ity. For large systems, there may be morethan one level of feeder circuits.See alsobranch circuit.

feedforward amplifier a circuit config-uration of low distortion amplifiers where adistorted voltage extracted from the outputsignal is fed again to the output port so as tocancel the distorted voltage included in theoutput signal.

feedforward control a form of compen-sation in which the measurement of a dis-turbance is used to take preventative controlaction before the effect of the disturbance isnoted at the system output (and then onlycompensated by feedback control action). Atypical feedforward control configuration isshown in the figure. The controller is de-signed to have the same effect as the cascadecombination of the process and process D.

A typical feedforward control configuration.

feedforward network Seefeedforwardneural network.

feedforward neural network one of twoprimary classifications of neural networks(the other is recurrent). In a feedforwardneural network thex-vector input to the sin-gle functional layer of processing elements(this input typically occurs via a layer of in-put units) leads to they-vector output in asingle feedforward pass. An example of thefeedforward neural network is a multilayerperceptron.


female connector a connector present-ing receptacles for the insertion of the corre-sponding male connector that presents pins.

fenestration any opening or arrangementof opening (normally filled with media forcontrol) for the admission of daylight.

Ferranti, Sebastian Ziani de (1864–1930) Born: Liverpool, Lancashire, England

Ferranti is best known for developing sys-tems of high-voltage AC power systems. Thegenerating and transmission systems he de-signed still form the basis for most modernpower systems. As a principal in the LondonElectric Supply Corporation, Ltd., Ferrantidemonstrated that high-voltage AC currentcould be distributed and then stepped downfor use in a more efficient and economicalsystem than the smaller DC current systemsthen operating. Ferranti invented a numberof other devices and systems as a consultantand in his own company. He was named pres-ident of the Institute of Electrical Engineersin 1911.

ferrite a term applied to a large groupof ceramic ferromagnetic materials usuallyconsisting of oxides of magnesium, iron, andmanganese. Ferrites are characterized bypermeability values in the thousands and areused for RF transformers and high Q coils.

ferrite beads small toroids made of fer-rites which are slipped over a conductor inorder to suppress RF currents. The beads actas RF chokes at high frequencies.

ferrite core a magnetic core made upof ferrite (compressed powdered ferrromag-netic) material, having high resistivity andlow eddy current loss.

ferrite core memory Seemagnetic corememory.

ferrite material a material that has verylow conductivity (σ ) and very large perme-ability (µ). Its properties can be altered when

an external magnetic field is applied. It isused in ferrite loaded loop antennas, for ex-ample, to increase the flux through the loopantenna.

ferroelectric material a polar dielectricin which the crystallographic orientation ofthe internal dipole moment can be changedby the application of an electric field.

ferrofluid iron based solution employedin voice coil/pole piece gap improving mag-netic flux and power handling capacity.

ferromagnetic materials in which inter-nal magnetic moments spontaneously line upparallel to each other to form domains, re-sulting in permeabilities considerably higherthan unity (in practice, 1.1 or more); exam-ples include iron, nickel, and cobalt.

ferroresonance a resonant phenomenoninvolving inductance that varies with satura-tion. It can occur in a system through theinteraction of the system capacitance withthe inductance of, for example, that of anopen-circuited transformer. Ferroresonanceresembles, to some extent, the normal reso-nance that occurs wherever L-C circuits areencountered. If the capacitance is apprecia-ble, ferroresonance can be sustaining or re-sult in a limited over voltage enough to dam-age the cable or the transformer itself.

ferroresonant transformer a trans-former that is designed to operate as a tunedcircuit by resonating at a particular fre-quency.

Fessenden, Reginald Aubrey (1866–1932) Born: East Bolton, Quebec, Canada

Fessenden is best known as a radio pio-neer who described the principle of ampli-tude modulation and the heterodyne effect.Fessenden became the chief scientist at Edi-son’s laboratory and then spent two yearswith Westinghouse. He later held teachingposts at Purdue University and Western Uni-versity of Pennsylvania (now the U. of Pitts-


burgh.) During his career, Fessenden was theholder of over 500 patents — only Edison hadmore.

FET Seefield-effect transistor.

fetch cycle the period of time duringwhich an instruction is retrieved from mem-ory and sent to the CPU. This is the part of thefetch–decode–execute cycle for all machineinstruction.

fetch on demand Seefetch policy.

fetch on miss Seefetch policy.

fetch policy policy to determine when ablock should be moved from one level of ahierarchical memory into the next level closerto the CPU.

There are two main types of fetch poli-cies: “fetch on miss” or “demand fetch pol-icy” brings in an object when the object isnot found in the top-level memory and is re-quired; “prefetch” or “anticipatory fetch pol-icy” brings in an object before it is required,using the principle of locality. With a “fetchon miss” policy, the process requiring the ob-jects must wait frequently when the objectsit requires are not in the top-level memory.A “prefetch” policy may minimize the waittime, but it has the possibility of bringing inobjects that are never going to be used. Italso can replace useful objects in the top-levelmemory with objects that are not going to beused. See alsocacheand virtual memory.The prefetching may bring data directly intothe relevant memory level, or it may bring itinto an intermediate buffer.

fetch-and-add instruction for a multi-processor, an instruction that reads the con-tent of a shared memory location and thenadds a constant specified in the instruction,all in one indivisible operation. Can be usedto handle multiprocessor synchronizations.

fetch–execute cycle the sequence of stepsthat implement each instruction in a com-

puter instruction set. A particular instructionis executed by executing the steps of its spe-cific fetch–execute cycle. The fetch part ofthe cycle retrieves the instruction to be exe-cuted from memory. The execution part ofthe cycle performs the actual task specifiedby the instruction. Typically, the steps in afetch–execute cycle are made up of variouscombinations of only three operations:

1. the movement of data between variousregisters in the machine,

2. the addition of the contents of two reg-isters or the contents of a register plus a con-stant with the results stored in a register, and,less frequently,

3. shift or rotate operations upon the datain a register.

fetching the process of reading instruc-tions from a stored program for execution.

FFT Seefast Fourier transform.

FH-CDMA frequency hopping code divi-sion multiple access.Seefrequency hoppingandcode division multiple access.

FIB Seefocused ion beam.

fiber Bragg grating a distributed Braggreflector written by ultraviolet light in thecore of a photosensitive optical fiber. Multi-ple weak Fresnel reflections coherently addin phase to produce a strong reflection over awell defined narrow band of wavelengths.

fiber cladding the region of an opticalfiber having a lower index of refraction thanthe core region, to allow confinement of lightin the core.

fiber distributed data interface (FDDI)an American National Standards Institutestandard for 100 megabits per second fiber-optic local-area networks. Incorporates to-ken processing and supports circuit-switchedvoice and packetized data. For its physicalmedium, it uses fiber optic cable, in a dualcounter-rotating ring architecture.


fiber optic bundle an optical system todeliver the image to a video monitor. In aconventional endoscope, the image is trans-mitted through a series of lenses to a cam-era outside of the patient. In an electronicendoscope the image is collected by a CCDcamera positioned inside the patient.

fiber optic sensor a sensor employing anoptical fiber to measure chemical composi-tion or a physical parameter, such as tem-perature, pressure, strain, vibration, rotationor electromagnetic fields. Light is launcheddown a fiber to a transducer or sensing ele-ment that alters the properties of the light inresponse to the parameter being measured.The altered light is returned back down a fiberto a detector.

fiber Raman amplifier provides amplifi-cation of signals through stimulated Ramanscattering in silica fibers. Raman amplifica-tion differs from stimulated emission in thatin stimulated emission, a photon of the samefrequency as the incident photon is emittedwhere as in the case of Raman amplification,the incident pump photon looses energy tocreate another photon at a lower frequency.

fiber-optic cable a glass fiber cable thatconducts light signals and can be used in to-ken ring local area networks and metropoli-tan area networks. Fiber optics can providehigher data rates than coaxial cable. They arealso immune to electrical interference.

fiber-optic interconnect interconnectthat uses an optical fiber to connect a sourceto a detector. An optical fiber is used forimplementing a bus. The merits are largebandwidth and high speed of propagation.

fidelity a qualitative term used to de-scribe how closely the output amplitude ofa device faithfully reproduces that of its in-put. Faithful reproduction here refers to thepreservation of those characteristics of the in-put signal (e.g., amplitude, frequency, and/orphase shift) essential to proper operation of

the device under question. Waveform fidelitydegradation is most often characterized by achange in signal shape (time domain), or achange in the relative scaling of its frequencycomponents as the signal propagates throughthe circuit or system (frequency domain).Seealsoamplitude linearity.

field (1) the member of an electrical ma-chine that provides the main magnetic flux,which then interacts with the armature caus-ing the desired machine operation (i.e., motoror generator).

(2) a description of how a physical quan-tity varies as a function of position and pos-sibly time.

See alsoelectric field, finite field.

field circuit a set of windings that pro-duces a magnetic field so that the electro-magnetics induction can take place in electricmachines.

field controlled thyristor (FCT) a thyris-tor controlled by change in the magnitude ofthe field current.

field current control a method of con-trolling the speed of a DC motor by varyingthe field resistance, thus producing a changein the field current.

field discharge resistor a resistor used todissipate the energy stored in the inductanceof a field winding. It may be a standard powerresistor that is connected across the windingjust prior to opening the supply switch, ora permanent non-linear resistive device thathas high resistance at normal voltage but lowresistance when voltage rises at switching.

field loss protection a fault-tolerantscheme used in electric motors. Some DCmotor control circuits provide field loss pro-tection in the event the motor loses its shuntfield. Under a loss of field, DC motors mayoverspeed causing equipment damage and/orpersonal injury. In a motor controller thathas field loss protection, a sensor determines


when the shunt field has lost current flow,then secures the motor before an overspeedcondition occurs.

field memory video memory required tostore the number of picture elements for onevertical scan (field) of video information ofan interlaced scanned system.

The memory storage in bits is computedby multiplying the number of video samplesmade per horizontal line times the numberof horizontal lines per field (vertical scan)times the number bits per sample. A sam-ple consists of the information necessary toreproduce the color information.

Storage requirements can be minimizedby sampling the color video information con-sisting of the luminance (Y) and the two colordifference signals, (R - Y) and (B - Y). Thecolor signal bandwidth is less than the lumi-nance bandwidth that can be used to reducethe field memory storage requirements. Foursamples of the luminance (Y) signal is com-bined with two samples of the (R - Y) signaland two samples of the (B - Y) signal. Thepreceding video sampling technique is desig-nated as 4:2:2 sampling and reduces the fieldmemory size by one-third. Field memory forNTSC video sampled at 4 times the color sub-carrier frequency at 8 bits/pixel would require3.822 megabits of RAM when 4:2:2 samplingis used.

field mill a measuring instrument forelectric fields consisting of a rotary arrayof blades which revolves near a capacitivepickup. The AC variation of the pickup’soutput is proportional to the electric field inthe region near the instrument in the directionof the rotor axis.

field orientation a term related tothe vector control of currents and volt-ages that enables direct control of ma-chine torque in AC machines. Vectorcontrol refers to the control of both theamplitude and phase angle, and the controlresults in a desired spatial orientation of theelectromagnetic fields in the machine.

field oxide an insulating silicon oxidelayer used in integrated circuits to electricallyisolate components.

field propagator the analytical descrip-tion of how electromagnetic fields are relatedto the sources that cause them. Commonfield propagators in electromagnetics are thedefining Maxwell equations that lead to dif-ferential equation models, Green’s functionsthat produce integral equation models, opti-cal propagators that lead to optics models,and multipole expansions that lead to modalmodels.

field rate the rate at which a field of videois generated. The field rate for NTSC televi-sion is 59.94 fields per second consisting ofalternating even and odd fields. Each fieldconsists of 262.5 lines of video, generatingthe even or odd numbered lines of video.

field reversing a method of achieving areversal of rotation of a DC motor by reserv-ing the field flux.

field singularity Seeedge condition.

field strength in general terms the mag-nitude of the electric field vector (in voltsper meter) or the magnitude of the magneticfield vector (in ampere-turns per meter). Asused in the field of EMC/EMI, the term isapplied only to measurements made in thefar field and is abbreviated as FS. For mea-surements made in the near field, the termelectric field strength (EFS) or magnetic fieldstrength (MFS) is used, according to whetherthe resultant electric or magnetic field, re-spectively, is measured.

field weakening a method of achievingspeed increase in DC motors by reducing thefield flux (increasing field circuit resistance).

field-by-field alignment a method ofalignment whereby the mask is aligned to thewafer for each exposure field (as opposed toglobal alignment).


field-effect transistor (FET) a majority-carrier device that behaves like a bipolar tran-sistor with the important difference that thegate has a very high input impedance andtherefore draws no current.

An active device with three terminals —gate, source, and drain — in the active (am-plifier) mode of operation, the drain current isrelated to the gate-source voltage. The rela-tionship is usually approximated by a squarelaw, but there are significant deviations fromthe square law depending on factors such asdevice geometry. The FET can also be usedas a switch, with the gate-source voltage con-trolling the “on/off” state of the conductingchannel between source and drain terminals.The input resistance at the gate is extremelyhigh (usually of order tens of megaohms) andthe gate current is negligibly small (usuallyof order picoamperes or less).

There are various families of FETs, in-cluding MOSFETs and JFETs. Within eachfamily, there are two types of FET, n-channeland p-channel (named for the sign of the ma-jority carriers that form the current conduct-ing path between source and drain).

Some FETs also have a fourth terminal,the “substrate” or “body” terminal. The p-n junctions between the substrate and thedrain and source terminals should be reverse-biased to insure proper device operation.

field-oriented control speed control ofan induction motor obtained by varying themagnitude and orientation of the airgap mag-netic field. This is also referred to as vectorcontrol and requires sensing of the rotor po-sition. Vector controllers allow the inductionmotor to operate very much like a DC mo-tor, including development of rated torque atzero speed.

field-programmable gate array (FPGA)(1) a programmable logic device that consistsof a matrix of programmable cells embeddedin a programmable routing mesh. The com-bined programming of the cell functions androuting network define the function of the de-vice.

(2) a gate array with a programmablemulti-level logic network. Reprogramma-bility of FPGAs make them generic hard-ware and allow them to be reprogrammed toserve many different applications. FPGAsconsist of SRAMS, gates, latches, and pro-grammable interconnects.

FIFO Seefirst-in-first-out.

FIFO memory commonly known as aqueue. It is a structure where objects aretaken out of the structure in the order theywere put in. Compare this with a LIFO mem-ory or stack. A FIFO is useful for bufferingdata in an asynchronous transmission wherethe sender and receiver are not synchronized:the sender places data objects in the FIFOmemory, while the receiver collects the ob-jects from it.

figure of merit performance evaluationmeasure for the various target and equipmentparameters of a sonar system. It is a subset ofthe broader sonar performance given by thesonar equations, which includes reverbera-tion effects.

file format the structure of the computerfile in which an image is stored. Often theformat consists of a fixed-size header fol-lowed by the pixel values written from thetop to the bottom row and within a rowfrom the left to the right column. However,it is also common to compress the image.SeealsoGraphicsInterchangeFormat, header,image compression, tagged image file format(TIFF).

Filippov method a definition of a solutionto a system of first-order differential equa-tions with discontinuous right-hand side,

x = f (t, x) ,

proposed by A. F. Filippov. A vector func-tion x(t) defined on the interval[t1, t2] isa solution to the above system of differen-tial equations in the sense of Filippov, if itis absolutely continuous and for almost all


t ∈ [t1, t2],x(t) ∈ F(t, x) ,

whereF(t, x) is an appropriately constructedconvex set.See alsodifferential inclusion.

fill-in when solving a set of sparse lin-ear equations using Gaussian elimination, itis possible for a zero location to becomenonzero. This new nonzero is termed a fill-in.

FILO Seefirst-in-last-out.

filter (1) a network, usually composedof inductors and capacitors (for lumped cir-cuit), or transmission lines of varying lengthand characteristic impedance (for distributedcircuit), that passes AC signals over a cer-tain frequency range while blocking signalsat other frequencies. A bandpass filter passessignals over a specified range (flow to fhi),and rejects frequencies outside this range.For example, for a DBS receiver that is to re-ceive satellite transmitted microwave signalsin a frequency range of 11 GHz to 12 GHz,a band-pass filter (BPF) would allow sig-nals in this frequency range to pass throughwith minimum signal loss, while blocking allother frequencies. A low-pass filter (LPF)would allow signals to pass with minimumsignal loss as long as their frequency wasless than a certain “cutoff frequency” abovewhich significant signal blocking occurs.

(2) an operator that transforms image in-tensityIx of pixel x into a different intensityIx, depending on the values of a set of (usu-ally neighboring) pixels (which may or maynot includex). Filtering is performed to en-hance significant features of an image or toremove nonsignificant ones or noise.

filter bank a set of filters consisting of abank of analysis filters and a bank of syn-thesis filters. The analysis filters decom-pose input signal spectra into a number ofdirectly adjacent frequency bands for furtherprocessing, and the synthesis filters recom-bine the signal spectra from different fre-quency bands.

filtered backprojection an algorithm forimage reconstruction from projections. Inthe filtering part of the algorithm, the pro-jections are measured, their Fourier trans-forms computed, and the transforms aremultiplied (filtered) by a weighting func-tion. In the backprojection part, the in-verse Fourier transforms of the weightedprojections are computed and summed toyield the reconstructed image. Filtered back-projection is the reconstruction algorithmcurrently used by almost all commercialcomputed tomography scanners.See alsoFourier transform, image reconstruction,projection, Radon transform, reconstruction,tomography.

filtering (1) an estimation procedure inwhich the present value of the state vector(see the definition) is estimated based on thedata available up to the present time.

(2) the process of eliminating object, sig-nal or image components which do not matchup to some pre-specified criterion, as in thecase of removing specific types of noise fromsignals. More generally, the application of anoperator (typically a linear convolution) to asignal.

fin efficiency a thermal characteristic of anextended surface that relates the heat transferability of the additional area to that of thebase area.

final test electrical test performed afterassembly to separate “good” devices from“bad.”

finesse measure of the quality of a Fabry-Perot interferometer; free spectral range di-vided by linewidth (full width at half maxi-mum).

finger stick an insulated stick like a hot-stick used to actuate a disconnect-switch atopa pole.

finite difference method a numericaltechnique for solving a differential equation


wherein the differential equation is replacedby a finite difference equation that relates thevalue of the solution at a point to the valuesat neighboring points.

finite difference time domain (FDTD) anumerical technique for the solution of elec-tromagnetic wave problems that involves themapping of the Maxwell equations onto a fi-nite difference mesh and then following thetime evolution of an initial value problem.This technique is widely used to investigatethe performance of a complex RF structures.

finite differences a method used to nu-merically solve partial differential equationsby replacing the derivatives with finite incre-ments.

finite element a numerical technique forthe solution of boundary value problems thatinvolves the replacement of the set of differ-ential equations describing the problem un-der consideration with a corresponding setof integral equations. The area or volume ofthe problem is then subdivided with simpleshapes such as triangles and an approxima-tion to the desired solution with free parame-ters is written for each subregion and the re-sulting set of equations is minimized to findthe final solution. This approach is usefulfor solving a variety of problems on complexgeometries.

finite field a finite set of elements and twooperations, usually addition and multiplica-tion, that satisfy a number of specific alge-braic properties. In honor of the pioneeringwork by Evariste Galois, finite fields are of-ten called Galois fields and denotedGF(q),whereq is the number of elements in the field.Finite fields exist for allq which are primeor the power of a prime.

finite impulse response (FIR) filter anyfilter having an impulse response that isnonzero for only a finite period of time (there-fore having a frequency response consistingonly of zeros, no poles). For example, every

moving average process can be written as theoutput of a FIR filter driven by white noise.See also impulse function, movingaverage, infinite impulse response (IIR) filter.Seerecursive filter.

finite state machine (FSM) a mathemat-ical model that is defined in discrete time andhas a finite number of possible states it canreside in. At each time instance, an input,x, is accepted and an output,y, and a transi-tion from the current state,Sc, to a new state,Sn, are generated based on separate functionsof the input and the current state. A finitestate machine can be uniquely defined by aset of possible states,S, an output function,y = f (x, Sc), and a transition function,Sn =g(x, Sc). An FSM describes many differentconcepts in communications such as convo-lutional coding/decoding, CPM modulation,ISI channels, CDMA transmission, shift-register sequence generation, data transmis-sion and computer protocols. Also known asfinite state automata (FSA), state machine.

finite state VQ (FSVQ) a vector quan-tizer with memory. FSVQ form a subset ofthe general class of recursive vector quanti-zation. The next state is determined by thecurrent stateSn together with the previouschannel symbolun by some mapping func-tion.

Sn+1 = f (un, Sn) , n = 0,1, . . .

This also obeys the minimum distortion prop-erty

α(x, s) = −1mind(x, β(u, s))

with a finite stateS = [α1, α2, . . . , αk, suchthat the stateSn can only take on values in S.The states can be called by names in gener-ality.

finite wordlength effect any perturbationof a digital filter output due to the use of fi-nite precision arithmetic in implementing the


filter calculations. Also called quantizationeffects.

finite-extent sequence the discrete-timesignals with finite duration. The finite-extentsequence{x(n)} is zero for all values ofnoutside a finite interval.

FIR See far-infrared, finite impulseresponse filter.

firing angle time in electrical degreesfrom the instant the valve voltage is positiveto the application of firing pulse to the valve(start of conduction). Also called delay an-gle.

firm power an amount of electric powerintended to be available at all times to a com-mercial customer, regardless of system con-ditions.

firm real-time Seefirm real-time system.

firm real-time system a real-time systemthat can fail to meet one or more deadlineswithout system failure.Compare withsoftreal-time, hard real-time.

firmware software that cannot be modi-fied by the end user.

first difference for a sequence{x(n)} thesequence obtained by simply subtracting its(n− 1)th element from itsnth element, i.e.,

y(n) = x(n)− x(n− 1)

first order hold (FOH) for a signalf (k),the sequence of straight lines connecting thesample points off (k). It interpolates thevalues between two adjacent samplesf (k)andf (k + 1) using a linear approximationgiven by

x(t) = x(kTs)+ t − kTs

Ts(x((k + 1)Ts)− x(kTs)).

first order system the system that canbe described by a linear first-order differenceequation. The output of the first-order systemy(n) is equal to a linear combination of thepast output valuey(n−1) and the input valuex(n), i.e.,

y(n) = αx(n)+ βy(n− 1).

first-fit memory allocation a memoryallocation algorithm used for variable-sizeunits (e.g., segments). The “hole” selected isthe first one that will fit the unit to be loaded.This hole is then broken up into two pieces:one for the process and one for the unusedmemory, except in the unlikely case of anexact fit, there is no unused memory.

first-in-first-out (FIFO) a queuing dis-cipline whereby the entries in a queue areremoved in the same order as that in whichthey joined the queue.

first-in-last-out (FILO) a queuing rulewhereby the first entries are removed in theopposite order as that in which they joinedthe queue. This is typical of Stack structuresand equivalent to last-in-first-out (LIFO).

first-swing stability criterion to deter-mine transient stability by use of the swingequation. The rotor angle immediatelyfollowing a severe disturbance usually in-creases. The criterion states that if the ro-tor angle swings back and decreases a shorttime after the disturbance, then the system isfirst-swing stable.

Fisher information a quantitative mea-surement of the ability to estimate a specificset of parameters. The Fisher informationJ (θ) is defined by

J (θ) = Eθ

(∂ ln fθ (y)

∂θ

)2

= − Eθ

(∂2 ln fθ (y)

∂2θ

)


whereY is aN -dimensional vector indexedby a vector of parametersθ . See alsoCramer-Rao bound.

fissile material an isotope which has asignificant probability of undergoing nuclearfission, e.g., U235, plutonium-239, thorium-232, and enriched uranium.

fission the nuclear reaction in which a sin-gle heavy nucleus is split into two or morelighter nucleii called "daughter" products andemit highly energetic sub-atomic particlesplus energy in the process.

fixed channel assignment (FCA) a tech-nique of assigning radio channels in a com-munications system in a fixed and prede-termined way in accordance with predictedrather than actual interference and propaga-tion conditions. Such assignments are notchanged during radio transmission.

fixed losses that component of the copperlosses in DC shunt, short-shunt, and long-shunt machines’ field circuit, that does notvary with change in the load current. Witha fixed field power supply, it is an acceptedindustry agreement to not consider the lossesin the field circuit rheostat in computing theefficiency and hence consider the field lossesas fixed losses.

fixed point Seeequilibrium point.

fixed reference D/A converter the analogoutput is proportional to a fixed (nonvarying)reference signal.

fixed resolution hierarchy an image pro-cessing scheme in which the original and re-constructed image are of the same size. Pixelvalues are refined as one moves from levelto level. This is primarily used for progres-sive transmission. Tree-structured VQ andtransform-based hierarchical coding are twoof the fixed resolution hierarchies.

fixed shunt Seeshunt capacitor.

fixed station (FS) that part of a radio com-munications system that is permanently lo-cated in a given geographical location. An-other name for the base station in a cellularradio telephone network.

fixed termination a broadband termina-tion of a specific nominal impedance, usually50 ohms, used to terminate the signal path ofa transmission line. Fixed terminations areused in the calibration of network analyzersand general microwave measurements.

fixed-gap head Seedisk head.

fixed-head disk a disk in which oneread/write head unit is placed at every trackposition. This eliminates the need for po-sitioning the head radially over the correcttrack, thus eliminating the “seek” delay time.Rarely used today because modern disks con-sist of hundreds of tracks per disk surface,making it economically infeasible to place ahead unit at every track.

fixed-length instruction the machine lan-guage instructions for a computer all have thesame number of bits.

fixed-point processor a processor capa-ble of operating on scaled integer and frac-tional data values.

fixed-point register a digital storage el-ement used to manipulate data in a fixed-point representation system whereby eachbit indicates an unscaled or unshifted binaryvalue. Common encoding schemes utilizedin fixed-point registers include Unsigned Bi-nary, Sign-Magnitude, and Binary CodedDecimal representations.

fixed-point representation (1) a num-ber representation in which the radix pointis assumed to be located in a fixed posi-tion, yielding either an integer or a fractionas the interpretation of the internal machinerepresentation.Contrast withfloating-pointrepresentation.


(2) a method of representing numbers asintegers with an understood slope and ori-gin. To convert a fixed-point representationto its value, multiply the representation bythe slope (called the DELTA in Ada) andadd the origin value. Fixed-point representa-tion provides fast computations for data itemsthat can be adequately represented by thismethod.

FLA Seefull load amperage.

flag (1) a bit used to set or reset some con-dition or state in assembly language or ma-chine language. For instance, the inheritanceflag, and the interrupt flag. As an example,each maskable interrupt is enabled and dis-abled by a local mask bit. An interrupt isenabled when its local mask bit is set. Whenan interrupt’s trigger event occurs, the pro-cessor sets the interrupt’s flag bit.

(2) a variable that is set to a prescribedstate, often “true” or “false,” based on theresults of a process or the occurrence of aspecified condition. Same as indicator.

flag register (1) a register that holds aspecial type of flag.

(2) a CPU register that holds the controland status bits for the processor. Typically,bits in the flag register indicate whether a nu-meric carry or overflow has occurred, as wellas the masking of interrupts and other excep-tion conditions.

flash EEPROM Seeflash memory.

flash memory a family of single-transistor cell EPPROMs. Cell sizes areabout half that of a two-transistor EEPROM,an important economic consideration. Bulkerasure of a large portion of the memory arrayis required.

The mechanism for erasing the memory iseasier and faster that needed for EEPROM.This allowed their adoption for making mem-ory banks on PCMCIA for replacing harddisks into portable computers. Recently, alsoused for storing BIOS on PC main boards. In

this way, the upgrading of BIOS can be madeby software without opening the computer bynon-specifically skilled people.

flashover arcing between segments of thecommutator of a DC machine. Flashovermay occur due to distortion of the airgap fluxas a result of heavy overloads, rapidly chang-ing loads, or operation with a weak mainfield.

flat address a continuous address specifi-cation by means of a unique number. Operat-ing systems such as Windows NT, OS/2 andMac OS use such a type of address. MS-DOSadopts the real-mode that adopts a segmentedmemory.

flat fading See frequency nonselectivechannel.

flat pack a component with two straightrows of leads (normally on 0.050-in centers)that are parallel to the component body.

flat panel display a very thin displayscreen used in portable computers. Nearly allmodern flat-panel displays use liquid crystaldisplay technologies. Most LCD screens arebacklit to make them easier to read in brightenvironments.

flat topping a distortion mechanism in anamplifying stage in which the stage is unableto faithfully reproduce the positive peaks ofthe output signal. Reasons for such problemsinclude poor regulation of the plate voltagesupply.

flat voltage start the usual initial as-sumption made when beginning a power-flow study. All voltages are assumed to be1.0 p.u., and all angles are assumed to bezero.

flat-compounded characteristic of certaincompound-wound DC generator designs inwhich the output voltage is maintained essen-


tially constant over the entire range of loadcurrents.

Fleming, John Ambrose (1849–1945)Born: Lancaster, Lancashire, England

Fleming is best known as the inventor ofthe thermionic diode valve, also known as theFleming valve. This valve was later adaptedby Lee DeForest into a form of vacuum tubehe called the Audion Triode. Fleming wasan avid experimentalist as well as a solid the-oretician. He began his studies under JamesClerk Maxwell and ended up consulting forEdison, Swan, and Ferranti, as well as Mar-coni. Thus Fleming can be considered one ofthe earliest pioneers in the field of radio andtelevision electronics.

Fletcher-Powell algorithm an iterativealgorithm of Fletcher and Powell (1963) forfinding a local minimum of the approxima-tion error. In its original form, the Fletcher-Powell algorithm performs an unconstrainedminimization.

flexible AC transmission system a trans-mission scheme in which each power line ismaintained at its optimal impedance, gener-ally by means of thyristor-controlled seriescompensators.

flexible link opposite to rigid link, issubject to elastic deformations of its struc-ture. Basic assumption is that its model isdescribed by the so-called Euler–Bernoullibeam equations of motion in which of secondor higher order in the deformation variablesare neglected.

flexible manipulator member of a classof robot manipulators with flexible links.Seeflexible link.

flexible manufacturing systems (FMS)manufacturing systems that deal with high-level distributed data processing and auto-mated material flow using computer-control-led machine tools, robots, assembly robotsand automated material handling, etc., with

the aim of combining the benefits of a highlyproductive transfer line, and a highly flexiblejob shop.

flexure node standing wave pattern inloudspeaker membrane caused by applied en-ergy.

flicker the apparent variation of lightingluminance over time to an observer.

(2) the apparent visual interruptions pro-duced in a TV picture when the field (onehalf of a TV frame) frequency is insuffi-cient to completely synchronize the visualimages. Occurs when the field rate is too lowand the biologic characteristic of the humaneye known as the persistence of vision doesnot give the illusion of continuous motion.Flicker is eliminated if the field rate exceeds50 to 60 Hz.

(3) repetitive sags and swells in the elec-tric service voltage, often accompanied byperiodic harmonic distortion.

flicker fusion the perception by the hu-man visual system of rapidly varying lights(flicker) as being steady (fused). Flicker fu-sion is why fluorescent lights and scenes inmovies and television appear to have constantillumination.

flicker noise occurs in solid-state devicesand vacuum tubes. Its power varies inverselywith frequency; therefore, it is also called1/f noise.

flip-chip for microwave a packaging in-terconnection, in which the chips are bondedwith contact pads, face down by solder bumpconnection. The solder bumps support theweight of the chip and control the height ofthe joint. It is a very good approach for high-frequency packages.

flip-flop a basic digital device capable ofstoring one bit of information (1 or 0). Inan “edge-triggered” flip/flop, information isstored in the device at the transition (positiveor negative) of a clock-signal. A flip/flop is


typically constructed using two “latches” inseries. The first one (the “master”) opens andcloses by one clock-phase, and the second(the “slave”) opens and closes using another(non-overlapping) phase. A flip/flop may beimplemented using a static or dynamic logicstyle. In the latter case, the size is only halfof the static version, but the information islost (due to leakage) after some time, need-ing “refresh” (or storing new data) for properoperation.

float switch a switch that is operated by afluid level in a tank or process channel.

floating point characteristic part of anumber that represents the exponent.

floating point operation arithmetic oper-ation (e.g., add or multiply) involving float-ing point numbers (i.e., numbers with deci-mal points).

floating-point operations per second(FLOP) a measure of processing speed, asin megaFLOPs or gigaFLOPs.

floating-point register a register thatholds a value that is interpreted as being infloating-point format.

floating-point representation in floating-point notation, a number is represented as afractional part times a selected base (radix)raised to a power. This is the counterpartof scientific notation used in digital systems.The decimal equivalent of the floating-pointvalue can be writtenN.n = f × re wheref is the fraction or mantissa ande is a pos-itive or negative integer called the exponent.Contrast withfixed-point representation.

floating-point unit a circuit that performsfloating point computations, which is gener-ally addition, subtraction, multiplication, ordivision.

FLOP Seefloating-point operations persecond.

floppy disk a flexible plastic disk coatedwith magnetic material. Enclosed in a card-board jacket having an opening where theread/write head comes into contact with thediskette. A hole in the center of the floppyallows a spindle mechanism in the disk driveto position and rotate the diskette. A floppydisk is a smaller, simpler, and cheaper formof disk storage media, and is also easily re-moved for transportation.

Floquet mode a solution to Maxwell’sequations that can be supported by an infinite,periodic structure.

Floquet’s theorem a basic theorem un-derlying the theory of wave propagation inperiodic structures.

flow dependency Seetrue data dependency.

flow diagram Seeflowchart.

flowchart a traditional graphic represen-tation of an algorithm or a program, in usingnamed functional blocks (rectangles), deci-sion evaluators (diamonds), and I/O symbols(paper, disk) interconnected by directionalarrows which indicate the flow of process-ing. Also called flow diagram.

flower pot a cover for the bushing of apad-mount transformer.

fluidized bed combustion a method ofsolid-fuel combustion in which the fuel, usu-ally coal, is pulverized and mixed with a bal-lasting substance and burned on a bed of pres-surized air. If the ballasting agent is crushedlimestone, sulfur from the coal is absorbedand carried out as solid ash.

fluorescence emission of light from anelectronically excited state that was producedby absorption of radiation with a wavelengthshorter than the emitted light. Fluorescenceemission is a quantum mechanically allowedtransition between electronic levels of the


Flyback converter.

same spin state, resulting in emission of lightwith a very short lifetime, typically nanosec-onds.

fluorescent lamp typically a lamp madeby exciting a low pressure discharge in mer-cury vapor and other gases; mercury, whenexcited in the discharge, predominantly emits257 nm radiation (ultraviolet) which is ab-sorbed by a phosphor on the inside wall of thelamp tube; the phosphor fluoresces, emittinga white light spectrum in the visible.

fluoroscopy a mounted fluorescent screenon which the internal structure or parts ofan optically opaque object may be viewedas shadows formed by the transmission of X-rays through the object.

flush the act of clearing out all actions be-ing processed in a pipeline structure. Thismay be achieved by aborting all of those ac-tions, or by refusing to issue new actions tothe pipeline until those present in the pipelinehave left the pipeline because their process-ing has been completed.

flux (1) lines that indicate the intensityand direction of a field. Intensity is usuallyrepresented by the density of the lines.

(2) a measure of the intensity of free neu-tron activity in a fission reaction, closely re-lated to power, the product of neutron den-

sity and neutron velocity, e.g., neutrons persquare cm per second.

flux density lines of magnetic flux per unitarea, measured in tesla; 1T = 1Wb/m2.

flux line Seedirection line.

flux linkage quantity that indicates theamount of flux associated with a coil. Fluxlinkage is denoted by the symbolλ andexpressed in Webers (Wb) or Weber-Turns(Wb-t). For a single turn coil, flux linkage isthe same as the flux. Flux linkages of anNturn coil areNφ Wb-t.

fluxmeter an instrument that measures thechange in magnetic flux within a coil by in-tegrating the induced voltage with respect totime.

flyback converter the isolated version ofthe buck-boost converter. In this case, thetransformer is also an inductor that stores en-ergy when the transistor is on and releasesenergy to the output when the transistor is off.The primary and the secondary of the trans-former conduct alternately. The advantage ofthis topology is that through the addition ofa second winding on the input inductor, i.e.,to form a transformer, electrical isolation isachieved. This type of transformer is called aflyback transformer or flyback inductor.Seealsobuck-boost transformer.


Flynn’s TaxonomySingle-Data Stream Multiple-Data Stream

Single-Instruction Stream von Neumann architecture/uniprocessorsRISC

Systolic processorsWavefront processors

Multiple-Instruction Stream Pipelined architecturesVery long instruction word processors

Dataflow processorsTransputers

flyback inductor Seeflyback converter.

flyback transformer Seeflyback converter.

flying head disk a disk storage device thatuses a read/write head unit “flying” over (i.e.,very closely above) the disk surface on a thinair bearing. Used e.g., in “Winchester” disks(a sealed “hard disk”). This is in contrast to,e.g., floppy disks, where the head unit is ac-tually in physical contact with the disk whenreading or writing.See alsodisk head.

Flynn’s taxonomy a classification systemthat organizes computer processor types aseither single-instruction stream or multiple-instruction stream and either single-datastream or multiple-data stream. The fourresultant types of computer processors areknown as SISD, MISD, SIMD, and MIMD(see table below). Due to Michael J. Flynn(1966).

flywheel a heavy wheel placed on the shaftof an electrical machine for storing kineticenergy. A flywheel may be used to help dampspeed transients or, to help deliver energy toimpact loads such as a punch press.

FM Seefrequency modulation.

FMS Seeflexible manufacturing systems.

FM sound a carrier wave whose instanta-neous frequency is varied by an amount pro-portional to the instantaneous amplitude ofthe sound input modulating signal.

focal length distance from a lens or mirrorat which an input family of parallel light rayswill be brought to a focus.

focus the position of the plane of best fo-cus of the optical system relative to some ref-erence plane, such as the top surface of theresist, measured along the optical axis (i.e.,perpendicular to the plane of best focus).

focus of attention the center of the regiondetected by a visual attention mechanism,which is the first stage of a pattern recog-nition system that is aimed to detect regionswhere a target object is likely to be found.

focus of expansion (1) in optical flowcomputation, the point into which the veloc-ity field lines of a translating object converge.

(2) the point in an image from whichfeature points appear to be diverging whenthe camera is moving forward or objects aremoving toward the camera.

focus-exposure matrix the variation oflinewidth (and possibly other parameters) asa function of both focus and exposure energy.The data is typically plotted as linewidthversus focus for different exposure energies,called the Bossung plot.

focused ion beam (FIB) a lithographytechnique similar to electron beam lithog-raphy in which a stream of charged ions israster-scanned to produce an image in a re-sist.

focuser a focusing electrode for an elec-tron steam in a vacuum tube.

FOH Seefirst order hold.


fold-over distortion the names given tothe phenomena associated with the genera-tion of an erroneously created signal in thesampling process are called “fold-over” dis-tortion or aliasing. Sampling a signal ex-hibits many of the same properties that a“mixer” circuit in RF communications pos-sesses. The mathematical relationship for amixer circuit and a sampling circuit is ex-pressed by the trigonometric identity

sinA·sinB = .5 cos(A−B)−.5 cos(A+B)

From the equation it is evident that if theA radian frequency is not twice theB radianfrequency, then theA−B term will producea signal whose radian frequency is less thanB ’s. This created signal will appear withinthe original frequency bandwidth. These er-ror signals can be minimized by incorporat-ing an anti-aliasing filter (i.e., a low-pass fil-ter) on the input to the sample-and-hold cir-cuit. The anti-aliasing filter bandlimits theinput frequencies so that fold-over distortionor aliasing is minimized.

folded sideband an FM method used fre-quently in video tape recording (VTR) sys-tems. The term refers to the fact that there isno corresponding upper sideband to the oneexisting sideband.

folding the technique of mapping manytasks to a single processor.

foldover a form of distortion in a digitalcommunications system which occurs whenthe minimum sampling rate is less that twotimes the analog input signal frequency.

footing impedance the electrical imped-ance between a steel tower and distant earth.

foot-candle the unit of illuminance whenthe foot is taken as the unit of length. It is theilluminance on a surface one square foot inarea on which there is a uniformly distributedflux of one lumen.

force manipulability ellipsoid a modelthat characterizes the end-effector forcesand/or torques,F , that can be generatedwith the given set of joint forces and/ortorques,τ , with the manipulator in a givenposture. Mathematically assuming that thejoint forces and/or torques, satisfy the equa-tion τT τ = 1, the ellipsoid (taking into ac-count kineto-static duality) can be written asfollows: FT (JJ T )F = 1. Here,J denotegeometric Jacobian of the manipulator.Seealsokineto-static duality.

forced commutation the use of externalcircuitry to artificially force a current zero,thus allowing a diode or thryistor to turn off.

forced interruption an interruption inelectric supply caused by human error, inap-propriate equipment operation, or resultingfrom situations in which a device is quicklytaken out of service by automatic or manualswitching operations.

forced outage the unscheduled interrup-tion of electric power to a portion of a powersystem due to equipment failure, weatherconditions, or other mishaps.

forced outage rate a measure of perfor-mance usually applied to generation units. Itis the ratio of equipment down-time vs. thetotal time that the unit is available for opera-tion.

forced system a dynamic system is said tobe forced if it is excited by a nonzero externalsource.

foreground (1) a process that is currentlythe process interacting with the user in a shell,as opposed to background processes, whichget suspended when they require input fromthe user.

(2) a term describing a Unix command thatis using your terminal for input and output,so that you will not get another Unix promptuntil it is finished.


(3) the foreground of a window is a singlecolor or a pattern. The foreground text orgraphics is displayed against the backgroundof the window.

fork (1) the action of a process creatinganother (child) process without ending itself.

(2) in Unix, fork is a process that can cre-ate a new copy of itself (a child) that has itsown existence until either it terminates (or iskilled) or its parent terminates. The preparehandler is called before the processing of thefork subroutine commences. The parent han-dler is called after the processing of the forksubroutine completes in the parent process.The child handler is called after the process-ing of the fork subroutine completes in thechild process.

form term used to indicate the structureand dimensions of a multiterm equation with-out details within component terms.

formants the main frequencies that resultfrom typical short-time spectral analysis ofvowels. Depending on the application, twoor three formants are commonly extracted.

forming gas a mixture of hydrogen andnitrogen (typically 5% hydrogen and 95% ni-trogen), typically used in the post-metal an-nealing process in wafer fabrication.

forward channel also referred to as for-ward link, it is the radio link (channel) in acellular or microcellular network where thebase station is the transmitter and the mobileterminal is the receiver. In the case of cel-lular networks, the term is synonymous withdown-link. The mobile to base station link isknown as the reverse link or up-link.

forward converter an isolated version ofa buck converter. The primary and the sec-ondary of the transformer conduct at the sametime. The transformer does not store energy.An additional winding with a diode is oftenused to reset the magnetizing energy whenthe transistor is off.

forward error correction (FEC) an er-ror control system used for simplex chan-nels (only a forward channel) where thechannel code is used to determine the mostlikely transmitted sequence of informationsymbols. Compare withautomatic repeatrequest.

forward error recovery a technique (alsocalled roll-forward) of continuing processingby skipping faulty states (applicable to somereal-time systems in which occasional missedor wrong responses are tolerable).

forward voltage the voltage across the de-vice when the anode is positive with respectto the cathode.

forward wave in a traveling wave tube,microwaves propagate in the same directionwith electrons in the electron beam. Electro-magnetic waves travel in the positive direc-tion of the coordinate system.

forward–backward procedure an effi-cient algorithm for computing the probabil-ity of an observation sequence from a hiddenMarkov model.

forwarding (1) sending a just-computedvalue to potential consumers directly, with-out requiring a write followed by a read.

(2) to immediately provide the result ofthe previous instruction to the current instruc-tion, at the same time that the result is writtento the register file. Also called bypass.

Foster’s reactance theorem states thatthe driving-point impedance of a networkcomposed of purely capacitive and inductivereactances is an odd rational function of fre-quency.

Foster–Seeley discriminator similar tothe balanced slope detector. A circuit usedto convert an incoming FM signal to a re-lated AM signal to provide a demodulationprocess.


Forward converter.

foundry for semiconductor manufactur-ing, a vendor providing wafer processing.

four connected Seepixel adjacency.

four-level laser laser in which the mostimportant transitions involve only four en-ergy states; usually refers to a laser in whichthe lower level of the laser transition is sepa-rated from the ground state by much morethan the thermal energy kT (contrast withthree-level laser).

four-point starter a manual motor starterthat requires a fourth terminal for the holdingcoil. Because of its independent holding coilcircuit, it is possible to vary the current in thefield circuit independently of the holding coilcircuit. The disadvantage is that the motorstarter holding relay will not drop out withloss of the field; however, proper overcurrent

protection should shut down the motor in theevent of field loss.

four-quadrant operation (1) a signedrepresentation of electrical or mechanicalvariables in the phase plane in order to sit-uate the different modes for energy transfer.This term can be used both for power elec-tronics and electrical machines. For electri-cal variables, the four-quadrant operation isdefined by the voltage–current (or current–voltage) characteristic with the two variablesexpressed as instantaneous or mean values.For mechanical variables, the term is definedby the torque–speed (or speed–torque) char-acteristic with the same time-domain repre-sentation as previously explained. The en-ergy transfer is defined with electrical powerin the voltage–current curve and with me-chanical power in the torque–speed curve.The four-quadrant operation is related to re-


versible power in electromechanical systems.For example, in the case of power electronics,the four-quadrant operation is defined onlywith electrical variables to visualize the wayof energy transfer in a static converter. Forelectrical machines, the same operation canbe defined with electrical variables at the in-put (motor) or at the output (generator) andalso with mechanical variables at the output(motor) or at the input (generator).

(2) the four combinations of forward/reverse rotation and forward/reverse torqueof which a regenerative drive is capa-ble. These are: motoring: forward rota-tion/forward torque; regeneration: forwardrotation/reverse torque; motor: reverse ro-tation/reverse torque; and regeneration: re-verse rotation/forward torque.

four-wave mixing a nonlinear opticalphenomenon in which four optical beams in-teract inside nonlinear media or photorefrac-tive crystals. When four beams of coherentelectromagnetic radiation intersect inside anonlinear or photorefractive medium, theywill, in general, form six interference pat-terns and induce six volume refractive indexgratings in the medium. The presence of theindex gratings will affect the propagation ofthese four beams. This may lead to energycoupling. The coupling of the four opticalbeams is referred to as four-wave mixing. Inone of the most useful four-wave mixing con-figurations, the four beams form two pairsof counterpropagating beams. In this par-ticular configuration, some of the refractiveindex gratings are identical in their gratingwavevectors. This leads to the generation ofphase conjugate waves. Four-wave mixingis a convenient method for the generation ofphase conjugated waves.

four-way interleaved splitting a resourceinto four separate units that may be accessedin parallel for the same request (usually in thecontext of memory banks).

Fourier amplitude the amplitudeangle (modulo 2π ) taken by the Fourier trans-

form. Specifically, given the Fourier trans-form F(f ) = A(f )(u)exp(i 8(f )(u)),then A(f ) represents the amplitude.SeeFourier transform.

Fourier filter a filter, or mask, placed inthe Fourier transform plane.

Fourier binary filter a filter placed inthe Fourier plane of an optical system con-structed with only two amplitude or twophase values.

Fourier integral the integral that yieldsthe Fourier transform of an absolutely inte-grable functionf over n-dimensional Eu-clidean space:

F(f )(u) =∫f (x) exp[−i u · x] dx

where the frequencyu is expressed in radi-ans.See alsoFourier transform.

Fourier optics optical systems that utilizethe exact Fourier transforming properties ofa lens.

Fourier optics relay lens a lens systemthat produces the exact Fourier transform ofan image. Two such relay lens will reproducean image without any phase curvature.

Fourier phase the phase angle (modulo2π ) taken by the Fourier transform. Specif-ically, given the Fourier transformF(f ) =A(f )(u)exp(i 8(f )(u)), then8(f ) repre-sents the phase.SeeFourier transform.

Fourier phase congruence for a 1-D real-valued signalf and a pointp, the Fourierphase that the signalf would have if the ori-gin were shifted top; in other words, it is theFourier phase off translated by−p. Thecongruence between the phases atp for thevarious frequencies — in other words the de-gree by which those phases atp are close toeach other — can be measured by

f (p)2 + H(f )(p)2,


whereH(f ) is theHilbert transformof f .SeeFourier amplitude, Fourier phase, Hilberttransform.

Fourier plane a plane in an optical systemwhere the exact Fourier transform of an inputimage is generated.

Fourier series Let f (t) be a continuoustime periodic signal with fundamental periodT such that

f (t) =∞∑

n=−∞cne

jω0t ,−∞ < t < ∞

whereω0 is the fundamental frequency,c0is a real number, andcn, n 6= 0 are com-plex numbers. This representation off (t) iscalled theexponential Fourier seriesof f (t).The coefficientscn are called the Fourier co-efficients and are given by

cn = 1

T

∫T

f (t)e−jω0t dt, n = 0,±1,

±2, . . .

The functionf (t) can also be expressed as

f (t) = c0+∞∑n=1

(an cos(nω0t)+bn sin(nω0t))

with Fourier coefficientsan and bn deter-mined from

an = 2

T

∫T

f (t) cos(nω0tdt), n = 1,

2,3, . . .

bn = 2

T

∫T

f (t) sin(nω0tdt), n = 1,

2,3, . . .

This representation off (t) is called thetrigonometric Fourier seriesof f (t). A sig-nal f (t) has a Fourier series if it satisfiesDirichlet conditions, given by (i)f (t) is ab-solutely integrable over any period, (ii)f (t)is piecewise continuous over any period, and(iii) d

dtf (t) is piecewise continuous over any

period. Virtually all periodic signals used in

engineering applications have a Fourier se-ries.

Fourier transform a linear mathematicaltransform from the domain of time or spacefunctions to the frequency domain. The dis-crete version of the transform (DFT) can beimplemented with a particularly efficient al-gorithm (fast Fourier transform or FFT). Thediscrete Fourier transform of a digital imagerepresents the image as a linear combinationof complex exponentials.

The Fourier transform of a continuoustime period signalf (t) is given by

F(ω) =∫ ∞

−∞f (t)e−jωtdt,−∞ < ω < ∞ .

If the signalf (t) is absolutely integrable andis well behaved, then its Fourier transformexists. For example, the rectangular pulsesignal

f (t) ={

1,−T2 ≤ t < T

20,otherwise

has Fourier transformF(w) = 2ω

sin ωT2 .

The inverse Fourier transform of a signalis given by

f (t) = 1

2π

∫ ∞

−∞F(ω)ejωtdω

See alsooptical Fourier transform, two-di-mensional Fourier transform.

Fourier, Jean Baptiste Joseph (1768–1830) Born: Auxerre, France

Fourier is best known for the developmentof new mathematical tools. The Fourier se-ries, which describes complex periodic func-tions, and the Fourier integral theorem, whichallows complex equations to be broken intosimpler trigonometric equations for easier so-lution, are named in honor of their discoverer.Fourier was an assistant lecturer under thegreat mathematicians Joseph Lagrange andGaspard Monge at the Ecole Polytechniquein Paris. He also served in a number of posi-tions in Napoleon’s government, eventually


resigning in protest. His contributions to thefield of electrical engineering are well recog-nized.

Fowler–Nordheim oscillation oscilla-tions in internal field emission tunneling cur-rent due to quantum interference.

Fox and Li method iterative diffractionintegral method for calculating the electro-magnetic modes of a laser resonator.

FPC coefficients the parameters of thelinear system that performs the linear predic-tion of a given stochastic process.

FPGA See field-programmable gatearray.

FQO Seefour-quadrant operation.

fractal from the Latinfractus, meaningbroken or irregular, a fractal is a rough ge-ometric shape that is self-similar over mul-tiple scales, i.e., its parts are approximatecopies of larger parts, or ultimately of thewhole. Magnifying a fractal produces moredetail. Fractals are good models of many nat-ural phenomena such as coast lines, clouds,plant growth, and lightning as well as arti-ficial items such as commodity prices andlocal-area network traffic.See alsofractalcoding.

fractal coding the use of fractals tocompress images. Regions of the imageare represented by fractals, which can thenbe encoded very compactly. Fractal cod-ing provides lossy compression that is in-dependent of resolution and can have ex-tremely high compression ratios, but it isnot yet practical for general images. Someof the methods are covered by patents.Seealsocompression, compression ratio, fractal,image compression, image resolution, lossycompression, model based image coding,spatial resolution, vector image.

fractional arithmetic mean radian band-width dimensionless ratio of the band-width (bw = ωH −ωL) to the arithmetic ra-dian center frequency (ωoa). The band edgesare usually defined as the highest (ωH ) andlowest (ωL) frequencies within a contigu-ous band of interest at which the loss equalsLAmax, the maximum attenuation loss acrossthe band.

bwa = ωH − ωL

ωoa= 2 (ωH − ωL)

ωH + ωL

fractional Brownian motion a nonsta-tionary generalization of Brownian motion;it is a zero-mean Gaussian process havingthe following covariance:

E [BH(t)BH (s)] =σ 2

2

(|t |2H + |s|2H − |t − s|2H

).

This process has interesting self-similarityand spectral properties; it is a fractal withdimensionD = 2 − H . See alsoBrownianmotion.

fractional discrimination coding a pre-processing scheme where the images are pro-cessed with fractional discrimination func-tions and then coded. This method of codingenables efficient local feature based encodingof pictures while preserving global features.This is a type of contextual coding scheme.

fractional geometric mean radian band-width dimensionless ratio of the band-width (bw = ωH − ωL) to the geometric ra-dian center frequency (ωoa). The band edgesare usually defined as the highest (ωH ) andlowest (ωL) frequencies within a contigu-ous band of interest at which the loss equalsLAmax, the maximum attenuation loss acrossthe band.

bwa = ωH − ωL

ωoa= (ωH − ωL)√

ωH + ωL

fractional horsepower National Elec-trical Manufacturers Association (NEMA)


classification describing any “motor built ina frame smaller than that having a continuousrating of 1 horsepower, open type, at 1700 to1800 rpm.”

fractional rate loss for frame-by-frametransmitted convolutionally encoded data,the fraction of overhead (compared to the sizeof the frame) needed to put the encoder intoa known state.

fragmentation waste of memory spacewhen allocating memory segments for pro-cesses. Internal fragmentation occurs whenmemory blocks are rounded up to fix blocksizes, e.g., allocated in sizes of power of 2only. E.g., if 35 K of data is allocated a 64 Kblock, the difference (64− 35 = 29K) iswasted. External fragmentation occurs be-tween allocated segments, as a result of allo-cating different sized segments for processesentering and leaving memory. This latterfragmentation is also called checkerboard-ing.

frame (1) a set of four vectors giving po-sition and orientation information.

(2) the basic element of a video sequence.The standard frame rate for TV standardsis 25 frames/s (European standards, e.g.,PAL and SECAM) or 30 frames/s (U.S. andJapanese standards, e.g., NTSC).

(3) in paging systems, a memory blockwhose size equals the size of a page. Framesare allocated space according to alignedboundaries, meaning that the last bits of theaddress of the first location in the frame willend withn zeros (binary), wheren is the ex-ponent in the page size. Allocating framesfor pages makes it easy to translate addressesand to choose a frame for an incoming page(since all frames are equivalent).

(4) time interval in a communication sys-tem over which the system performs some pe-riodic function. Such functions can be mul-tiple access functions (e.g., TDMA multi-ple access frame) or speech-processing func-tions (e.g., speech coding frame, interleavingframe, or error control coding frame).

(5) single image; component of a se-quence of images which, displayed rapidlyin succession, give the illusion of a movingpicture. In video a frame represents a singlecomplete scan of the image; it often consistsof two interlaced fields.

frame grabber a device that is attached toan electronic camera and which freezes andstores images digitally, often in gray-scale orcolor format, typically in one or three 8-bitbytes per pixel respectively.

frame memory video memory requiredto store the number of picture elements forone complete frame of electronically scannedvideo information. The memory storage inbits can be computed by multiplying thenumber of video samples made per horizontalline, times the number of horizontal lines perfield (vertical scan), times the number bits persample, times the number of fields/frame. Asample consists of the information necessaryto reproduce the color information.

The NTSC television system consists oftwo interlaced fields per frame. Storage re-quirements are usually minimized by sam-pling the color video information consistingof the luminance (Y ) and the two color differ-ence signals, (R−Y ) and (B−Y ). The colorsignal bandwidth is less than the luminancebandwidth that can be used to reduce the fieldmemory storage requirements. Four samplesof the luminance (Y ) signal is combined withtwo samples of the (R − Y ) signal and twosamples of the (B − Y ) signal. The preced-ing video sampling techniques is designatedas 4:2:2 sampling and reduces the field mem-ory size by one third.

The number of memory bits that are re-quired to store one NTSC frame is two timesthe number of bits required to store oneNTSC field. Field memory for NTSC videosampled at 4 times the color subcarrier fre-quency at 8 bits/pixel would require 7.644megabits of RAM when 4:2:2 sampling isused.


frame processing image processing ap-plied to frames or sequences of frames.

frame rate in analog and digital video, therate per second at which each frame or im-age is refreshed. The NTSC (National Tele-vision Systems Committee) rate for analogtelevision is 30 frames per second.

frame store a device that stores imagesdigitally, often in gray-scale or color format,typically in one or three 8-bit bytes per pixelrespectively; a frame store often incorpo-rates, or is used in conjunction with, a framegrabber.

frame synchronization a method to ob-tain rough timing synchronization betweentransmitted and received frames. The degreeof synchronization obtained depends on thelevel of frame synchronization (super-frame,hyper-frame). The level of frame synchro-nism also determines the actual method andthe place where the synchronism takes place.

frame synchronizer (1) a time-base cor-rector effective over one frame time. Videodata may be discarded or repeated in one-frame increments; in the absence of an input,the last received frame is repeated, thus pro-ducing a freeze-frame display.

(2) a device that stores video information(perhaps digitally) to reduce the undesirablevisual effect caused by switching nonsyn-chronous sources.

Franklin, Benjamin (1706–1790) Born:Boston, Massachusetts, U.S.A.

Franklin was best known as the great-est American statesman, scientist, and en-trepreneur of his day. He was the personwho described electric charge, introducedthe terms “positive” and “negative” as de-scriptive of charges, and described the fun-damental nature of lightening. Franklin re-ceived no formal training in the sciences.His printing business gave him the finan-cial security necessary to carry out numer-ous fundamental experiments on the nature

of electricity. The results of these experi-ments were later given theoretical validationby Michael Faraday. Franklin’s political con-tributions to the United States were also ex-traordinary. Amazingly, this additional re-sponsibility did not deter Franklin from nu-merous other experiments and inventions ina variety of fields.

Fraunhofer region Seefar field.

free distance the minimum Hamming dis-tance between two convolutionally encodedsequences that represent different valid pathsthrough the same code trellis. The free dis-tance equals the maximum column distanceand the limiting value of the distance profile.

free electron laser laser in which the ac-tive medium consists of electrons that aresubject to electric and magnetic fields but arenot associated with atoms or molecules.

free input a quantity influencing thecontrolled process from outside, generatedwithin the process environment, and beingbeyond direct or indirect influence of the con-troller; a free input is defined by a continuoustrajectory over given time interval or by a se-quence of values at given time instants.

free input forecasting process of predict-ing — at a given time — future values of thefree input; these future values can be pre-dicted in different forms: single future freeinput trajectory (scenario) over a specifiedtime interval (forecasting interval); bunch ofpossible scenarios over specified time inter-val, (or intervals) — eventually with attachedweights; stochastic model valid at the consid-ered forecasting time and other. Free inputforecasting can be essential for making mostof the decisions concerning current values ofthe manipulated inputs.

free input model model, usually statedin form of a set of differential or differenceequations, describing the behavior of the freeinputs; for example, ARMA process is often


used as the free input model. Usually free in-put model is driven by a white noise stochas-tic process or by a sequence of set bounded— but otherwise time-independent — quan-tities.

free running frequency the frequency atwhich an oscillator operates in the absenceof any synchronizing input signal. The fre-quency is usually determined by a time con-stant from externally connected resistanceand capacitance.

free space loss the propagation loss expe-rienced between two isotropic antennas in avacuum resulting from spreading of powerover spherical wavefronts centered at thetransmitting antenna and from the finite aper-ture of the receiving antenna. The free spaceloss increases proportional to the square ofthe distance from the antenna and the squareof the frequency.

free system Seeunforced system.

free-page list a linked-list of informationrecords pointing to “holes” (i.e., free pageframes) in main memory.

free-space interconnect interconnect us-ing optical elements such as lenses, gratings,and holograms, in which optical signals maycross each other in the space. The mainadvantages are high interconnection densityand parallelism, and dynamic reconfigura-tion.

free-wheeling diode a diode connected inparallel with the load of a half-wave rectifierto prevent the return of energy from the loadto the source. Due to the stored energy inthe inductive load, the current must continueafter the source voltage becomes negative.The free-wheeling diode provides a path forthe current to circulate and allows the diodeor SCR in the rectifier to turn off.

Fremdhold integral equation a linear in-tegral equation wherein the limits of integra-tion are fixed.

frequency the repetition rate of a periodicsignal used to represent or process a commu-nication signal. Frequency is expressed inunits of hertz (Hz). 1 Hz represents one cy-cle per second, 1 MHz represents one millioncycles per second, and 1 GHz represents onebillion cycles per second.

frequency chirp a monotonic change inoptical frequency with time; often used forlaser radar ranging in analogy with conven-tional radar and for ultrashort pulse gener-ation via pulse compression or autocorrela-tion. See alsochirp signal.

frequency compensation the modifica-tion of the amplitude–frequency response ofan amplifier to broaden the bandwidth or tomake the response more nearly uniform overthe existing bandwidth.

frequency converter an equipment or cir-cuit that converts an RF signal to an interme-diate (IF) signal in receivers. It converts anIF signal to an RF signal in transmitters.

frequency correlation function a func-tion characterizing the similarity of a re-ceived signal with respect to a shift in fre-quency.

frequency deviation in a frequency mod-ulated system, the number of hertz the carrieris varied during the modulation process.

frequency domain representation of asignal by frequency components, such as itsFourier transform.

frequency domain sampling a proce-dure that is a dual of the time-domain sam-pling theorem, whereby a time-limited signalcan be reconstructed from frequency-domainsamples.


frequency distortion caused by the pres-ence of energy storage elements in an am-plifier circuit. Different frequency compo-nents have different amplifications, resultingin frequency distortion, and the distortion isspecified by a frequency response curve.

frequency division duplex (FDD) a tech-nique based on the allocation of two separatefrequency bands for the transmission in bothdirections in a link. FDD typically requires aguard band between the two frequency bandsin order to eliminate interference between thetransmitter and receiver at a terminal. In afull duplex communication system, informa-tion flows simultaneously in both directionsbetween two points.

frequency division multiple access(FDMA) a multiple-access techniquebased on assigning each user a unique fre-quency band upon which transmission takesplace.See alsotime division multiple access,code division multiple access.

frequency division multiplexing (FDM)refers to the multiplexing of signals by shift-ing each signal to a different frequency band.See alsofrequency division multiple access.

frequency domain storage an opticaldata storage technique in which individ-ual bits are stored by saturating or bleach-ing spectral holes into an inhomogeneouslybroadened absorption spectrum. Usually,this hole burning is accomplished using elec-tromagnetic fields of discrete frequencies.The maximum frequency domain storagedensity is given by the ratio of inhomoge-neous width to homogeneous width of theabsorption spectrum.

frequency hopping a system where thecarrier frequency used for modulation ischanged according to some predeterminedhopping pattern. The receiver hops accord-ing to the same pattern in order to retrieve theinformation. If the rate of change is lower orequal to the data rate, the process is known

as slow frequency hopping, otherwise as fast.Frequency hopping is used for counteractingintentional interference, frequency selectivefading, and as a multiple access technique(FH-CDMA).

frequency independent antenna antennawith very large bandwidths. Upper fre-quency can be about 40 times higher thanthe lower frequency. Examples: spiral andlog periodic antennas.

frequency modulation (FM) angle mod-ulation in which the instantaneous frequencyof a sine-wave carrier is caused to depart fromthe carrier frequency by an amount propor-tional to the instantaneous value of the mod-ulating wave.

frequency nonselective channel a radiochannel that has a uniform transfer functionover the bandwidth of information signal. Insuch a channel, fading leads to reduced SNRat the receiver, but does not lead to distortionof the symbol shape and hence intersymbolinterference.

frequency pulling shift of the frequencyof a laser oscillation mode from the emptycavity or nondispersed cavity frequency to-ward the center frequency of the amplifyingtransition.

frequency regulation the change in thefrequency of an unloaded generator with re-spect to its frequency in a fully-loaded state.Typically applied to small, isolated powersystems such as emergency power units.

frequency relay a protective relay whichmonitors the frequency of the electric powersystem.

frequency resolution a measure of theability of a system to resolve different fre-quencies in a signal. As the frequency resolu-tion increases, more finely-spaced frequencycomponents can be resolved. The time res-olution of a system is roughly inversely pro-


portional to the frequency resolution; the un-certainty principle places a lower bound onthe time-frequency resolution product.

frequency response consider a systemwith transfer function given by

H(s) = Y (s)

F (s).

The termfrequency responseis used to de-noteH(jω), (commonly writtenH(ω)). Thefrequency response is often used in describ-ing the steady-state (stable) system responseto a sinusoidal input. For example, in thecase of a stable continuous-time system withinput signalf (t) = A cos(ωt), the steady-state output signal is given by

yss(t) = A|H(jω) cos(ωt + 6 H(jω) .

The term|H(jω)|,−∞ < ω < ∞, iscalled themagnitude response. The term6 H(jω),−∞ < ω < ∞, is called thephase response. The magnitude and phaseresponses are typically represented in theform of a Bode plot. See alsoBode plot,transfer function.

frequency reuse a way to increase the ef-fective bandwidth of a satellite system whenavailable spectrum is limited. Dual polariza-tions and multiple beams pointing to differentearth regions may utilize the same frequen-cies as long as, for example, the gain of onebeam or polarization in the directions of theother beams or polarization (and vice versa)is low enough. Isolations of 27–35 dB aretypical for reuse systems.

frequency reuse cluster a group ofcells in a cellular communications networkwherein each frequency channel allocated tothe network is used precisely once (i.e., in onecell of the group). The size of the frequencyreuse cluster is one of the major factors thatdetermines the spectral efficiency and ulti-mately the capacity of a cellular network. Fora given channel bandwidth, the smaller thecluster size, the higher the network capacity.

frequency selective filter a filter thatpasses signals undistorted in one or a setof frequency bands and attenuate or totallyeliminate signals in the remaining frequencybands.

frequency selective surface (FSS) fil-ter made of two-dimensional periodic arraysof apertures or metallic patches of variousshapes. Several layers can be used to obtaina structure with a set of desired spectral prop-erties.

frequency shift keying (FSK) (1) an en-coding method where different bits of infor-mation are represented by various frequen-cies; used for spread-spectrum signal encod-ing for security and reduced interchannel in-terference.

(2) a digital modulation technique inwhich each group of successive source bitsdetermines the frequency of a transmitted si-nusoid.

frequency space the transformed spaceof the Fourier transform.

frequency synchronization the processof adjusting the frequency of one source sothat it exactly matches that of another source:more specifically, so thatn periods of onefrequency are exactly equal tom periods ofthe other frequency, for integraln andm. Seealsophase-locked loop.

frequency synthesizer an oscillator thatproduces sinusoidal wave with arbitrary fre-quency. In common cases, generated fre-quencies are allocated with a frequency spac-ing called the channel step.

frequency variation a change in the elec-tric supply frequency.

frequency-modulation recording Seemagnetic recording code.


fresh fuel nuclear fuel which has neverparticipated in a nuclear reaction and is thusonly slightly radioactive.

Fresnel region the region in space aroundan antenna at which the fields have both trans-verse and radial components and the antennapattern is dependent on the distance from theantenna. The Fresnel, or near-field, regionis typically taken to ber < 2D2/λ, wherer is the distance from the antenna,D is themaximum dimension of the antenna, andλ isthe wavelength.

Fresnel zone an indicator of the signif-icant volume of space occupied by a radiowave propagating along a line-of-sight pathbetween the transmitter and receiver. At anarbitrary point which is at distanced1 fromthe transmitter and at distanced2 from thereceiver, along the axis joining the trans-mitter and the receiver, a radio wave withwavelengthλ occupies a volume, which atthat point between the transmitter and the re-ceiver, has a radius which is given by the ra-dius of the first Fresnel zone. The radius ofthe first Fresnel zone is given by

R =√λd1d2

d1 + d2

From this equation one can see that the sig-nificant volume occupied by a radio wave isessentially ellipsoidal. In order for a trans-mitter and receiver to communicate success-fully, it is important for this volume to bekept free from any obstruction, especially atmicrowave frequencies.

fricative a phoneme pronounced whenconstricting the vocal tract so as the airflow becomes turbulent and the correspond-ing frequency spectrum looks like a broad-band noise. The utterances with fricationcan occur with or without phonation. Whenphonation also occurs, the presence of for-mants can be detected in the spectrum.

Friis transmission formula a formulaused to compute the received signal power in

a radio communication link given the amountof power transmitted, the effective aperturesor gains of the transmit and receive antennas,the distance between the antennas, and thewavelength.

fringing the portion of the flux at the airgap in a magnetic circuit that does not followthe shortest path between the poles.

frit a relatively low softening point mate-rial of glass composition.

Fritchman model N -state Markoff modeldescribingN error states of a channel.

frog the top of a tower.

front end (1) an initial processing unit thatprovides a user interface and/or reformatsinput data for subsequent computations ona special-purpose or high-performance backend processor.

(2) the portion of the compiler that doesmachine-independent analysis.

front of a motor the end of a motor thatis opposite to the major coupling or drivingpulley.

front porch video blanking level durationof approximately 1.27 microseconds con-tained within the horizontal blanking inter-val of the composite NTSC signal. The frontporch duration is 2% of the total horizontalline time, starting at the end of a horizontalline of video signals and ending at the begin-ning of the horizontal line sync signal.

frontside bus the term used to describethe main bus connecting the processor to themain memory (cf. backside bus).

frustrated total reflection Seeattenuatedtotal reflection.

FS Seefixed station.


FSA finite state automata.Seefinite statemachine.

FSK Seefrequency shift keying.

FSM Seefinite state machine.

FSS Seefrequency selective surface.

FSVQ Seefinitestatevector quantizer.

fuel cell an electrochemical device inwhich the chemical energy of a conventionalfuel converted directly and efficiently intolow voltage, direct current electrical energy.

fuel cycle (1) the operating time of a nu-clear reactor between shutdowns for refuel-ing.

(2) the life cycle of nuclear fuel, e.g., rawmaterials, to fresh fuel, to irradiated fuel, tospent fuel, to recycling, and finally to residualwaste.

fuel pool a large vat of water used to storeboth fresh and spent nuclear fuel at the reactorsite.

fuel rod long, thin canister, typically madeof zirconium or other metal, which containsthe radioactive fuel in a nuclear reactor core.

fulgurite a vitrified tube of fused sand cre-ated by power fault arcs or lightning strikesto sandy soil.

full adder a combinational logic circuitthat produces a two-bit sum of three one-bitbinary numbers.

full bridge amplifier a class-D amplifierbased on a full-bridge inverter configuration.

full load amperage (FLA) a value, foundon the nameplate of an induction motor, in-dicating the expected current drawn by themotor when operating at rated voltage andload.

full permutation symmetry Seenonlinearsusceptibility.

full rank Jacobian a Jacobian matrixwhere dim(R(J )) = r, dim(N(J )) = n− r.At a singular configuration of the Jacobian,one has dim(R(J )) + dim(N(J )) = n in-dependently of the rank of the matrixJ .Seenull space of the Jacobianandrange ofJacobian.

full-load amperage the current measuredat the input of an electrical apparatus whichhas a rated variable at the output. The full-load amperage is also defined as a value thatpermits the system to operate in a safe con-dition if it is equal to the rated value. Thefull-load amperage is equal to the rated valueif and only if the voltage is at its rated value.The full load is defined with the electricalpower in the case of transformers, genera-tors, or power electronics converters and it isdefined with the mechanical power for elec-trical motors.

full-load speed the speed of a motor thatproduces rated power when operating at ratedvoltage and, for AC motors, frequency.

full-load torque the torque of a motor thatis producing rated power at rated speed.

full-wave analysis the rigorous compu-tation of electromagnetic fields without ap-proximations (apart for numerical discretiza-tion) is often referred to as full-wave analysis.

full-wave control both the positive andnegative half cycle of the waveforms are con-trolled.

full-wave rectifier a device that passespositive polarity portions of a signal and re-verses negative polarity portions of an ACsignal. Ideally, for a sinusoidal inputvi(t) =Vm cos(ωt), the output of an ideal full-waverectifier isvo(t) = |Vm cos(ωt)|.


full width at half max (FWHM) thewidth of a distribution function (i.e., the hor-izontal distance between two points havingthe same vertical coordinate) defined at avertical position halfway between the max-imum and minimum. This value exists forcertain functions that are of interest as pos-sible point spread functions (e.g., Gaussian,Moffat function) and gives a rough idea ofthe amount of blur to be expected.

fully associative cache a cache blockmapping technique where a main memoryblock can reside in any cache block position.A number of bits from the address receivedfrom the CPU is compared in parallel withthe cache tags to see if the block is presentin the cache. Although costly to implement,and impractical for large cache sizes, it isthe most flexible mapping method.See alsocache, direct mapped cache, and associativ-ity.

function a programming construct thatcreates its own frame on a stack, accepting ar-guments, performing some computation, andreturning a result.

function sensitivity in a circuit describedby the functionF(s, x)wheres is a complexvariable, andx is a passive or active elementin the circuit realization of the function, theequation

SF(s,x)

x = ∂F (s, x)

∂x

x

F (s, x)

Under conditions of sinusoidal steady state,whens = jω, the functionF(jω, x) is de-scribed by the sensitivity

SF(jω,x)

x = ReSF(jω,x)

x + j Im SF(jω,x)x

= S|F(jω,x)|x + j ∂argF(jω, x)

∂x/x

Thus, the real part of the function sensitivitygives the relative change in the magnituderesponse, and the imaginary part gives thechange in the phase response, both with re-spect to a normalized element change.

function test a check for correct deviceoperation generally by truth table verifica-tion.

function-oriented design a design method-ology based on a functional viewpoint of asystem starting with a high-level view andprogressively refining it into a more detaileddesign.

functional block ASICs designed usingthis methodology are more compact than ei-ther gate arrays or standard cells becausethe blocks can perform much more complexfunctions than do simple logic gates.

functional decomposition the division ofprocesses into modules.

functional redundancy checking an er-ror detection technique based on checkingbus activity generated by a processor (mas-ter) with the use of another slave processorwhich monitors this activity.

functional unit a module, typically withina processor, that performs a specific limitedset of functions. The adder is a functionalunit.

fundamental component the first ordercomponent of a Fourier series of a periodicwaveform.

fundamental frequency for a continuoustime periodic signal,f (t) with fundamentalperiodT (seconds),ω0 = 2π/T (rad/sec).See alsoperiodic signal.

fundamental mode (1) operating mode ofa circuit that allows only one input to changeat a time.

(2) lowest order mode of a laser cavity;usually refers to a Gaussian transverse fielddistribution without regard to the longitudi-nal mode order.

fuse an overcurrent device that employsone or more fusible elements in series. The


fusible element(s) are designed to melt andinterrupt the circuit when current above athreshold value flows in the circuit.

fuse coordination the process of match-ing the fuse or circuit breaker interruption ca-pability to overload current and short-circuitcurrent and insuring that the protective deviceclosest to a fault opens first so as to minimizethe service interruption.

fuse cutout a primary distribution voltagelevel fuse that employs a replaceable fusiblelink and provides a means of disconnect. Theinterrupting rating of a fuse cutout can besomewhat lower than that of a power fuse,however.

fuse link used in nonerasable programma-ble memory devices. Each bit in the mem-ory device is represented by a separate fuselink. During programming, the link is either“blown” or left intact to reflect the value ofthe bit. See alsofusible link ROM.

fuse reducer an adapter that allows fusesto be installed in fuseclips designed for largerfuses.

fuse saving the practice of tripping distri-bution line reclosers or circuit breakers ona fast trip to beat sectionalizing fuses thatare protecting laterals. The following re-close will restore all load. If the fault re-mains following the fast trip(s), subsequentslow trip(s) will allow the fuse to operate ona permanent fault on the lateral. Fuse savingdecreases the customer outage rate but causesmore sags to customers in a fault situation.

fused disconnect a disconnect switch thatalso employs fuse(s) for the purpose of over-current protection.

fusible link ROM read-only memory us-ing fuse links to represent binary data.Seealsofuse link.

fusion a nuclear reaction in which twolight nuclei are combined into a heavier nu-cleus with a release of energy. Fusion powerreactors have been proposed but as yet neversuccessfully constructed.

Futurebus a bus specification standard-ized by the IEEE, originally defined for CPU-memory data transfers.

fuzz stick a short hot-stick.

fuzzification a procedure of transform-ing a crisp set or a real-valued number into afuzzy set.

fuzzifier a fuzzy system that transformsa crisp (nonfuzzy) input value in a fuzzy set.The most used fuzzifier is thesingleton fuzzi-fier, which interprets a crisp point as a fuzzysingleton. It is normally used in fuzzy controlsystems. See alsofuzzy inference system,fuzzy singleton.

fuzziness the degree or extent of impreci-sion that is naturally associated with a prop-erty, process or concept.

fuzzy adaptive control adaptive con-trol involving fuzzy logic concepts or fuzzycontrol involving adaptation. Examples ofsuch control are self-organizing fuzzy con-trol, model reference fuzzy adaptive control,and fuzzy self-tuning control.

fuzzy aggregation network artificial neu-ral network that can be trained to produce acompact set of fuzzy rules with conjunctiveand disjunctive antecedents. An aggregationoperator is used in each node.

fuzzy algorithm an ordered set of fuzzyinstructions which upon execution yield anapproximate solution to a specified problem.Examples of fuzzy algorithms include fuzzyc-mean clustering, fuzzy-rule-based classifi-cation, etc.

fuzzy AND Seetriangular norm.


fuzzy approximation an action or pro-cess of using a fuzzy system to approximatea nonlinear function or mapping.

fuzzy associative memory a look-up ta-ble constructed from a fuzzy IF-THEN rulebase and inference mechanism to define a re-lationship between the input and output of afuzzy controller, or a fuzzy system.See alsofuzzy inference system.

fuzzy automata based on the concept offuzzy sets, a class of fuzzy automata is formu-lated similar to Mealy’s formulation of finiteautomata.

fuzzy basis functions a set of fuzzy mem-bership functions or their combinations in afuzzy system that form a basis of a functionspace, normally for function approximation.

fuzzy behavioral algorithm a relationalalgorithm that is used for the specific purposeof approximate description of the behavior ofa system.

fuzzy C the C programming language in-corporating fuzzy quantities and fuzzy logicoperations.

fuzzy c-means (FCM) a fuzzy versionof the commonly used K-means clusteringalgorithm. The main feature of FCM is thata type of membership function is utilized incomputing a distance measure. Also calledfuzzy ISODATA.

fuzzy clustering a method used to clusterdata into subsets based on a distance or simi-larity measure that incorporates fuzzy mem-bership functions.

fuzzy cognitive map (FCM) nonlineardynamical systems whose state trajectoriesare constrained to reside in the unit hypercube[0,1]n. The components of the state vectorof an FCM stand for fuzzy sets or events thatoccur to some degree and may model events,actions, goals, etc. The trajectory of an FCM

can be viewed as an inference process. Forexample, if an FCM’s trajectory starting froman initial conditionx(0) converges to an equi-librium stateA, then this can be interpretedas the FCM providing an answer to a question“What if x(0) happens?” In this sense it canbe interpreted that the FCM stores the rule“IF x(0) THEN the equilibriumA.” FCMswere introduced by B. Kosko.

fuzzy complement the complement of afuzzy setA is understood as NOT (A).

fuzzy concentration the concentration ofa fuzzy set produces a reduction by squaringthe membership function of that fuzzy set.

fuzzy conditional statement an IF-THEN statement of which either the an-tecedents or consequent(s) or both may belabels of fuzzy sets. Also know as fuzzy rule,linguistic fuzzy model, linguistic rules.Seefuzzy IF-THEN rule.

fuzzy control the application of fuzzylogic and fuzzy inference rules utilizingknowledge elicited from human experts ingenerating control decisions for the control ofprocesses. More specifically, a means of ex-pressing an operator’s knowledge of control-ling a process with a set of fuzzy IF-THENrules and linguistic variables.

fuzzy control system stability stability ofa fuzzy control system that normally includesa plant to be controlled and a fuzzy controller.

fuzzy controller devices for implement-ing fuzzy control, normally including the fol-lowing components: a fuzzifier to transforma crisp real-valued number to a fuzzy set; afuzzy rule base to define IF-THEN controlrules; a fuzzy inference engine to combinethe IF-THEN rules; and a defuzzifier to trans-form a fuzzy set to a crisp real-valued num-ber.

fuzzy controller design a process ofdetermining a fuzzy controller including a


fuzzifier, a fuzzy rule base, a fuzzy inferenceengine, and a defuzzifier.

fuzzy decision rule a decision rule withfuzzified antecedents.

fuzzy decision tree a decision tree withfuzzy decision functions.

fuzzy decision tree algorithm an algo-rithm to generate a fuzzy decision tree, suchas the branch-bound-backtrack algorithm.

fuzzy decisional algorithm a fuzzy algo-rithm that serves to provide an approximatedescription of a strategy or decision rule.

fuzzy definitional algorithm a finite setof possible fuzzy instructions that define afuzzy set in terms of other fuzzy sets (andpossibly itself, i.e., recursively).

fuzzy digital topology an extension of thetopological concepts of connectedness andsurroundness in a digital picture using fuzzysubsets.

fuzzy dilation an operator that increasesthe degree of belief in each object of a fuzzyset by taking the square root of the member-ship function.

fuzzy dynamic model model involvingfuzzy logic concepts and system dynamics.A typical example is a hierarchical modelwith a higher level of fuzzy inference rulesand a lower level of analytical linear dynamicequations.

fuzzy dynamic system system involvingfuzzy logic concepts and system dynamics.A typical example is a fuzzy control systemwith a dynamic plant and a fuzzy controller.

fuzzy entropy the entropy of a fuzzy setis a functional to measure the degree of fuzzi-ness of a fuzzy set based on Shannon’s func-tion. It has been applied to provide a quanti-

tative measure of ambiguity to the problemsof gray-tone image enhancement.

fuzzy estimation an action or processof deducing state variables from given inputand output measurements in a fuzzy dynamicstate space model.

fuzzy expert system a rule-based sys-tem for approximate reasoning in which ruleshave fuzzy conditions and their triggering aredriven by fuzzy matching with fuzzy facts.See alsofuzzy inference system.

fuzzy filter a filter involving fuzzy logicconcepts. A typical example is a filterconstructed from a number of local filtersthrough fuzzy membership functions.

fuzzy gain scheduling gain schedulinginvolving fuzzy logic concepts. A typicalexample is the use of a number of fuzzy IF-THEN rules to adjust a controller gain.

fuzzy generational algorithm a fuzzy al-gorithm serves to generate rather than definea fuzzy set. Possible applications of genera-tional algorithms include generation of hand-written characters and generation of speech.

fuzzy geometry a way to describe an im-age by fuzzy subsets.

fuzzy global control a fuzzy control foran overall plant or a control based on a fuzzyglobal model that consists of a number oflocal models smoothly connected through aset of membership functions.

fuzzy global model a fuzzy dynamicmodel consisting of a number of local linearmodels or simple nonlinear models smoothlyconnected through a set of nonlinear mem-bership functions.

fuzzy grammar an attributed grammarthat uses fuzzy primitives as the syntacticunits of a language.


fuzzy H-infinity control H-infinity con-trol involving fuzzy logic concepts or fuzzycontrol to achieve an H-infinity controllerperformance index.

fuzzy H-infinity filter an H-infinity filterinvolving fuzzy logic concepts or a fuzzy fil-ter to achieve an H-infinity filter performanceindex.

fuzzy hierarchical systems fuzzy sys-tems of hierarchical structures. A typicalexample is a two-level fuzzy system witha higher level of fuzzy inference rules andlower level of analytical linear models.

fuzzy identification a process of deter-mining a fuzzy system or a fuzzy model. Atypical example is identification of fuzzy dy-namic models consisting of determination ofthe number of fuzzy space partitions, deter-mination of membership functions, and de-termination of parameters of local dynamicmodels.

fuzzy IF-THEN rule rule of the form

IF x isA THEN y isB

whereA andB are linguistic values definedby fuzzy sets on universe of discourseX andY respectively (abbreviated asA −→ B).The statement “x is A” is called the an-tecedentor premise,while “y isB” is calledtheconsequenceor conclusion.A fuzzy IF-THEN rule can be defined as a binaryfuzzyrelation. The most common definition of afuzzy ruleA −→ B is asA coupled with B,i.e.,

R = A −→ B = A× B ,

or

µA×B(x, y) = µA(x) ? µB(y) ,

where? is a triangular norm.See alsofuzzy relation, fuzzy set, linguistic

variable, triangular norm.

fuzzy implication Seefuzzy IF-THENrule.

fuzzy inference a fuzzy logic principle ofcombining fuzzy IF-THEN rules in a fuzzyrule base into a mapping from a fuzzy set inthe input universe of discourse to a fuzzy setin the output universe of discourse. A typicalexample is a composition inference.

fuzzy inference engine a device or com-ponent carrying out the operation of fuzzy in-ference, that is, combining fuzzy IF-THENrules in a fuzzy rule base into a mappingfrom a fuzzy set in the input universe of dis-course to a fuzzy set in the output universe ofdiscourse.See alsoapproximate reasoning,fuzzy inference system, fuzzy rule.

fuzzy inference system a computingframework based on fuzzy set theory, fuzzyIF-THEN rules, and approximate reasoning.There are two principal types of fuzzy infer-ence systems:

1. Fuzzy inference systems mappingfuzzy sets into fuzzy sets (pure fuzzy infer-ence systems) that are composed of a knowl-edge base containing the definitions of thefuzzy sets and the database of fuzzy rulesprovided by experts; and a fuzzy inferenceengine that performs the fuzzy inferences.

2. Fuzzy inference systems performingnon-linear mapping from crisp (nonfuzzy) in-put data to crisp (nonfuzzy) output data. Inthe case of a Mamdani fuzzy system, in ad-dition to a knowledge base and a fuzzy in-ference engine, there is a fuzzifier that rep-resents real-valued inputs as fuzzy sets, anda defuzzifier that transforms the output set toa real value. In the case of a Sugeno fuzzysystem, special fuzzy rules are used, giving acrisp (nonfuzzy) conclusion, and the outputof the system is given by the sum of thosecrisp conclusions, weighted on the activationof the premises of rules. Some fuzzy sys-tems of this type hold the universal functionapproximation property.

See also defuzzifier, fuzzifier, fuzzyset, fuzzy IF-THEN rule, fuzzy reasoning,Sugeno fuzzy rule, universal function appro-ximation property.


Pure fuzzy inference system.

Mamdani fuzzy system.

fuzzy input–output model input–outputmodels involving fuzzy logic concepts. Atypical example is a fuzzy dynamic modelconsisting of a number of local linear trans-fer functions connected by a set of nonlinearmembership functions.

fuzzy integral an aggregation operatorused to integrate multiattribute fuzzy infor-mation. It is a functional defined by usingfuzzy measures, which corresponds to prob-ability expectations. Two commonly usedfuzzy integrals are Sugeno integral and Cho-quet integral.

fuzzy intensification an operator that in-creases the membership function of a fuzzyset above the crossover point and decreasesthat of a fuzzy set below the crossover point.

fuzzy intersection the fuzzy intersectionis interpreted as “A AND B,” which takesthe minimum value of the two membershipfunctions.

fuzzy ISODATA Seefuzzy c-means.

fuzzy local control fuzzy control in asense of local region of plant operation range,which is a key feature of fuzzy control. Atypical example is a local control based on alocal fuzzy dynamic model.

fuzzy local model a fuzzy model ina sense of local region of plant operation

range. A typical example is a fuzzy model ofTakagi–Sugeno–Kang type.

fuzzy local system fuzzy system in a senseof local region of plant operation range. Atypical example is a fuzzy system of Takagi–Sugeno–Kang type.

fuzzy logic introduced by Zadeh (1973),gives us a system of logic that deals withfuzzy quantities, the kind of information hu-mans manipulate that is generally impre-cise and uncertain. Fuzzy logic’s maincharacteristics is the robustness of its in-terpolative reasoning mechanism.See alsoapproximate reasoning, linguistic variable,modifier, generalizedmodusponens.

fuzzy logic controller controller based ona fuzzy logic system.See alsofuzzy logicsystem.

fuzzy logic system a system of logic thatdeals with fuzzy quantities, that are unpreciseand uncertain.

fuzzy measure a subjective scale used toexpress the grade of fuzziness similar to theway probability measure is used to expressthe degree of randomness. An example offuzzy measure isgλ-fuzzy measure proposedby Sugeno.

fuzzy membership function functionof characterizing fuzzy sets in a universeof discourse which takes values in the in-terval [0,1]. Typical membership func-tions include triangular, trapezoid, pseudo-trapezoid, and Gaussian functions.

fuzzy minus the minus operation appliedto a fuzzy set to give an intermediate effectof dilation. See alsofuzzy plus.

fuzzy model See alsofuzzy inferencesystem.

fuzzy model-based control fuzzy controlbased on various kinds of mathematical or


linguistic models. A typical example is con-trol of dynamic systems based on dynamicfuzzy models, which consist of a numberof local linear models smoothly connectedthrough a set of nonlinear membership func-tions.

fuzzy modeling combination of availablemathematical description of the system dy-namics with its linguistic description in termsof IF-THEN rules. In the early stages offuzzy logic control, fuzzy modeling meantjust a linguistic description in terms of IF-THEN rules of the dynamics of the plantand the control objective. Typical exam-ples of fuzzy models in control applicationincludes Mamdani model, Takagi–Sugeno–Kang model, and fuzzy dynamic model.

fuzzy neural control a control system thatincorporates fuzzy logic and fuzzy inferencerules together with artificial neural networks.

fuzzy neural network artificial neuralnetwork for processing fuzzy quantities orvariables with some or all of the followingfeatures: inputs are fuzzy quantities; outputsare fuzzy quantities; weights are fuzzy quan-tities; or the neurons perform their functionsusing fuzzy arithmetic.

fuzzy neuron a McCulloch–Pitts neuronwith excitatory and inhibitory inputs repre-sented as degrees between 0 and 1; output isa degree to which it is fired.

fuzzy nonlinear control nonlinear con-trol involving fuzzy logic concepts or fuzzycontrol with application to nonlinear sys-tems.

fuzzy number a convex fuzzy set of thereal line such that

1. it exists exactly one point of the real linewith membership 1 to the fuzzy set;

2. its membership function is piecewisecontinuous.

In fuzzy set theory, crisp (nonfuzzy) num-bers are modeled as fuzzy singletons.

See alsoconvex fuzzy set, fuzzy singleton.

fuzzy observer a device to estimate thestates of a dynamic system, involving fuzzylogic concepts. A typical example is an ob-server constructed from a number of localobservers through fuzzy membership func-tions.

fuzzy operator logical operator used onfuzzy sets for fuzzy reasoning. Examples arethe complement (NOT), union (OR), and theintersection (AND).

fuzzy optimal control optimal control in-volving fuzzy logic concepts or fuzzy controlto achieve an optimal control performance in-dex.

fuzzy optimal filter optimal filter involv-ing fuzzy logic concepts or fuzzy filter toachieve an optimal filter performance index.

fuzzy OR Seetriangular co-norm.

fuzzy output feedback fuzzy controlbased on feedback of a plant output. Thisis closely related to fuzzy dynamic models.

fuzzy parameter estimation a methodthat uses fuzzy interpolation and fuzzy ex-trapolation to estimate fuzzy grades in a fuzzysearch domain based on a few cluster center-grade pairs. An application of this method isto estimate mining deposits.

fuzzy partition partition of a plant oper-ating space based on fuzzy logic concepts. Atypical example is a partition of a state spaceby overlapping subspaces which are charac-terized by a set of fuzzy membership func-tions.

fuzzy pattern matching a pattern match-ing technique that applies fuzzy logic to dealwith ambiguous or fuzzy features of noisypoint or line patterns.


fuzzy plus the operation of plus appliedto a fuzzy set to give an intermediate effectof concentration.See alsofuzzy minus.

fuzzy PROLOG the PROLOG program-ming language incorporating fuzzy quanti-ties and fuzzy logic operations.

fuzzy proposition a proposition in whichthe truth or falsity is a matter of degree.

fuzzy reasoning approximate reasoningbased on fuzzy quantities and fuzzy rules.See alsoapproximate reasoning.

fuzzy relation a fuzzy set representing thedegree of association between the elementsof two or more universes of discourse.Seealsofuzzy set.

fuzzy relational algorithm a fuzzy al-gorithm that serves to describe a relation orrelations between fuzzy variables.

fuzzy relational matrix a matrix whoseelements are membership values of the corre-sponding pairs belonging to a fuzzy relation.

fuzzy relaxation a relaxation techniquewith fuzzy membership functions applied.

fuzzy restriction a fuzzy relation thatplaces an elastic constraint on the values thata variable may take.

fuzzy rule Seefuzzy IF-THEN rule.

fuzzy rule bank a collection of all fuzzyrules that are arranged in N-dimensionalmaps.

fuzzy rule base a set of fuzzy IF-THENrules. It is a central component of a fuzzysystem and defines major functions of a fuzzysystem.

fuzzy rule minimization a technique tosimplify the antecedent and consequent partsof rules and to reduce the total number ofrules.

fuzzy rule-based system system basedon fuzzy IF-THEN rules, or another namefor fuzzy systems.See alsofuzzy inferencesystem.

fuzzy rules of operation a system of rela-tional assignment equations for the represen-tation of the meaning of a fuzzy proposition.

fuzzy self-tuning self-tuning based onfuzzy logic concepts or fuzzy systems withsome kind of self-tuning strategy or adapta-tion.

fuzzy set introduced by L. Zadeh (1965).A fuzzy setA in a universe of discourseX ischaracterized by amembership functionµAwhich maps each element ofX to the inter-val [0,1]. A fuzzy set may be viewed as ageneralization of a classical (crisp) set whosemembership function only takes two values,zero or unity.See alsocrisp set, membershipfunction.

fuzzy singleton a fuzzy set of member-ship value equal to 1 at a single real-valuedpoint and 0 at all other points in the universeof discourse.See alsomembership function,fuzzy set.

fuzzy sliding mode control a combina-tion of available mathematical description ofthe system dynamics with its linguistic de-scription in terms of IF-THEN rules. In theearly stages of fuzzy logic control, fuzzymodeling meant just a linguistic descriptionin terms of IF-THEN rules of the dynamicsof the plant and the control objective.SeealsoTSK fuzzy modelandfuzzy system.

fuzzy space the region containing thefuzzy sets created by set theoretic operations,as well as the consequent sets produced byapproximate reasoning mechanisms.

fuzzy state feedback fuzzy control basedon feedback of plant states.


fuzzy state space model fuzzy dynamicmodels in state space format. A typical ex-ample is a set of local linear state space mod-els smoothly connected through a set of non-linear membership functions.

fuzzy system a fuzzy system is a set ofIF-THEN rules that maps the input space,sayX, into the output spaceY . Thus, thefuzzy system approximates a given functionF : X → Y by covering the function’s graphwith patches, where each patch correspondsto an IF-THEN rule. The patches that over-lap are averaged.See alsofuzzy inferencesystem.

fuzzy thresholding algorithm a thresh-olding algorithm with a threshold selectedbased on fuzzy measure.

fuzzy tuning the tuning of a control sys-tem using fuzzy logic and fuzzy rules.

fuzzy union an operator that take the max-imum value of the membership grades.

FWHM Seefull width at half max.


GG (giga) a prefix indicating a quantity of109. For instance, a gigabyte (GB) of storageis 1,000,000,000 (typically implemented as230) bytes.

GCR common notation for compression.Seecompression.

gd common notation for DC drain con-ductance.

GI common notation for current gain. GIis dimensionless.

gm common notation for DC transconduc-tance.

GP common notation for power gain indecibels.

GT common notation for transducer gainin amperes/volt.

GV common notation for voltage gain.GV is dimensionless.

G-line a line of the mercury spectrum cor-responding to a wavelength of about 436 nm.

GaAlSb/InGaAlAsSb nearly latticematched semiconductor heterostructure sys-tem capable of “staggered” band lineups inwhich the electrons and holes congregate inseparate layers. Unique transport proper-ties are utilized in optical devices and tunnelstructures.

GaAs periodic table symbol for galliumarsenide.Seegallium arsenide.

GaAs laser a semiconductor laser forwavelengths in the near infrared. The active

medium is a gallium-arsenide semiconductoralloy.

GaAs/AlGaAs most commonly grownsemiconductor epitaxial heterostructure dueto its lattice match, common anion, and ex-isting technology base of GaAs devices.

Gabor transform a short-time Fouriertransform in which the window function isthe Gaussian function.

gain (1) the ratio of the output variable ofa device to its input variable. For calculationpurposes, the dimensionality of the gain issimply the unit of the output variable dividedby the unit of the input variable. The gain ofa device is a dimensionless value only whenthe electrical units of both the input and out-put variables are the same (e.g., voltage gain,current gain, power gain, etc.). In this case,a gain greater than one indicates an increasefrom input to output, while a value for gainless than one is indicative of a decrease (orattenuation). The overall gain of several cas-caded components is found by multiplyingthe individual gains of each component in thesystem. Gain is often expressed in decibelsto facilitate calculation of cascaded gains ina system.See alsodecibel.

(2) the ratio of the radiation intensity ofa particular antenna to that of an isotropicradiator, in the same direction and at the samedistance.

gain circles circles of constant gain plot-ted on the Smith chart that can be used tographically impedance match a device toachieve a desired gain. The circles are gener-ated by plotting on the Smith chart the solu-tion for the source reflection coefficient,0s ,or load reflection coefficient,0L, in the trans-ducer gain equation for a fixed value of

G =|S21|2

(1−|0S |2

)(1−|0L|2)

|(1−S110S)(1−S220L)−S12S210L0S |2


for a fixed value ofGt , whereS11, S22, S12,andS21 are the scattering parameters for thedevice.

gain coefficient factor multiplying dis-tance in a formula for the exponential am-plitude or intensity growth of a wave in anunsaturated amplifying medium.

gain compensation this deals with theassumption in motion estimation in inter-frame coding that illumination is spatiallyand temporally uniform. Under these as-sumptions, the monochrome intensitiesb(zt)andb(zt − τ) of two consecutive frames arerelated byb(zt) andb(z − D, t − τ)whereτis the time between two frames,D is the two-dimensional translation vector. Often this as-sumption about uniform illumination is notcorrect. In some situations, a multiplicativefactor calledgain is used to change the inten-sity. This is called gain compensation. Thisresults in estimatingD using gradient-typealgorithms to minimize the square of the pre-diction error.

gain focusing focusing or collimation ofan electromagnetic beam by the profile of thegain; gain guiding.

gain guided laser diode electricallypumped semiconductor laser in which themode fields are confined in the transverse di-rection by the profile of the gain.

gain medium medium for which an out-put electromagnetic wave has more powerthan the corresponding input, essential forlaser operation.

gain ripple difference between themaximum gain (Gmax = 10 log10 [poutmax/pincident ]) and the minimum gain across theband (Gmin = 10 log10 [poutmin/pincident ]),expressed in decibels. The band edges areusually defined as the highest and lowest fre-quencies within a contiguous band of inter-est at which the gain equals the minimum

required passband gain.

Gripple = Gmax −Gmin

= 10 log10

(poutmax

pincident

)− 10 log10

(poutmin

pincident

)

gain saturation reduction in gain that oc-curs when the intensity of a laser field de-pletes the population inversion.

gain switching rapid turn-on of the gain ina laser oscillator for the purpose of producinga large output pulse.

gain–bandwidth product for amplifiersbased on a voltage-feedback op-amp, in-creasing the closed-loop gain causes a pro-portional decrease to the closed-loop band-width; thus the product of the two is a con-stant. This constant is called the gain–bandwidth product, and is a useful figure-of-merit for the performance of the op-amp.

gain-guided laser laser in which the modefields are confined in the transverse directionby the profile of the gain.

gain-shape vector quantization (GSVQ)Seeshape-gain vector quantization.

gain-transfer measurement methodcommon antenna gain measurement schemein which the absolute gain of the antenna un-der test is determined by measuring its gainrelative to a gain standard (i.e., antenna withaccurately known gain).

GaInN/AlInN rapidly evolving semicon-ductor heterostructure system with ability toemit light in the green and blue regions of thespectrum for long-lifetime LEDs and lasers.

Galerkin’s method in an integral equa-tion technique used to solve a numerical elec-tromagnetic problem, the method in whichthe expansion and testing functions are thesame.


gallium arsenide (GaAs) a compositewidely used in the fabrication of active el-ements.

gallium phosphide (GaP) (1) a semicon-ductor for high-speed electronics and that ispart of the family of III-V compounds forsemiconductor lasers.

(2) an acousto-optic material with goodacousto-optic figure of merit and withlow acoustic attenuation for high-bandwidthacousto-optic cells.

galloping a low-frequency vibration ofelectric power lines caused by wind.

Galois field a finite field withq elementsdenoted as GF(q), important in the study ofcyclic codes.See alsofinite field.

Galvani, Luigi (1737–1798)Born: Bologna, Italy

Galvani is best known for the use ofhis name in the description of zinc plating,or “galvanizing” metals. Galvani was ananatomist who studied the flow of electric-ity in frog’s legs when metal contacts wereapplied. Many of Galvani’s ideas were laterproved incorrect by Alessandro Volta. How-ever, his pioneering work was honored bythe reference to Galvanic current as an earlyname for electric current. Andre Amperealso honored him by naming his current mea-surement apparatus a galvanometer.

gamma camera a device that uses a scin-tillation crystal to detect gamma photons forradionuclide imaging. Also known as a scin-tillation camera.

gamma correction inside a CRT, the lu-minance of a pixel is represented by a volt-ageV ; however, the luminance of the rayproduced by that voltage is not proportionalto V , but rather toV γ , whereγ is a device-dependent constant between 2 and 2.5. It isthus necessary to compensate for this effectby transforming the luminanceL into L1/γ ;this is called gamma correction. Some CCD

cameras implement gamma correction di-rectly on the signal that they generate. Somecomputer monitors allow modification of thefactor γ or gamma correction through soft-ware.

gamma rays electromagnetic radiation ofvery high energy (greater than 30 keV) emit-ted after nuclear reactions or by a radioactiveatom when its nucleus is left in an excitedstate after emission of alpha or beta particles.

gamma-ray laser laser producing its out-put in the gamma-ray region of the spectrum,often considered to be any wavelength belowabout 0.1 angstrom; graser, not yet demon-strated.

ganged operation of a multiple phase de-vice with all phases operated simultaneously.

GaP Seegallium phosphide.

gap filler a low-power transmitter thatboosts the strength of transmitted DAB RFsignals in areas that normally would be shad-owed due to terrain obstruction. Gap fillerscan operate on the same frequency as DABtransmissions or on alternate channels thatcan be located by DAB receivers using auto-matic switching.

gapless arrester a lightning arresterwhich is distinguished from a gapped arresterby having a continuous conductive path be-tween the conductor and ground.

gapped arrester a lightning arresterwhose conducting path contains a gas- or air-filled spark gap which must be broken downby the lightning impulse voltage.

garbage an object or a set of objects thatcan no longer be accessed, typically becauseall pointers that direct accesses to the objector set have been eliminated.

garbage collection the process by whichthe memory, a program no longer uses, is


automatically reclaimed and reused. Rely-ing on garbage collection to manage mem-ory simplifies the interfaces between pro-gram components. Two main methods usedto implement garbage collection are refer-ence counting and mark-and-sweep.

garbage collector a software run-timesystem component that periodically scansdynamically allocated storage and reclaimsallocated storage that is no longer in use(garbage).

gas cable an electric power cable whichis pressurized with an insulating gas, typi-cally sulfur hexaflouride, for its primary in-sulation.

gas capacitor a capacitor whose dielec-tric is composed of a high-pressure gas, of-ten nitrogen or an electronegative gas such assulfur hexaflouride.

gas circuit breaker a circuit breaker inwhich the arc between the contacts is extin-guished by immersion in or a blast of an elec-tronegative gas.

gas discharge excitation electron impactexcitation or pumping of a laser medium oc-curring as a result of collisions between thelasing atoms or molecules and the dischargeelectrons.

gas dynamic excitation excitation orpumping of a laser medium occurring as a re-sult of heating followed by rapid expansionand cooling of a gaseous laser medium.

gas insulated switchgear circuit break-ers and switches whose primary insulation iscompressed gas.

gas laser laser in which the amplifyingmedium is a gas.

gas metal arc welding (GMAW) a weld-ing process that produces coalescence of met-als by heating them with an arc between

a consumable filler metal electrode and theparts to be joined. The process is used withshielding gas and without the application ofpressure.

gas substation an electric power substa-tion in which the conductors are insulatedfrom each other and the ground by a high-pressure gas, generally sulfur hexaflouride.

gas tungsten arc welding (GTAW) awelding process that produces coalescence ofmetals by heating them with an arc betweena nonconsumable tungsten electrode and theparts to be joined. The process is used withshielding gas and without the application ofpressure. Filler may or may not be used.

gate (1) a logical or physical entitythat performs one logical operation, such asAND, NOT, or OR.

(2) the terminal of a FET to which controlsthe flow of electrons from source to drain. Itis usually considered to be the metal contactat the surface of the die. The gate is usuallyso thin and narrow that if any appreciablecurrent is allowed to flow, it will rapidly heatup and self-destruct due to I-R loss. Thissame resistance is a continuing problem inlow noise devices, and has resulted in the cre-ation of numerous methods to alter the gatestructure and reduce this effect.

gate array a semicustom integrated cir-cuit (IC) consisting of a regular arrangementof gates that are interconnected through oneor more layers of metal to provide customfunctions. Generally, gate arrays are pre-processed up to the first interconnect levelso they can be quickly processed with finalmetal to meet customer’s specified function.

gate length the length in microns of thatportion of the FET channel, as measured inthe direction of channel current flow, that canbe pinched off by application of the propercontrol bias. Gate lengths are usually on theorder of 0.1 to 2 microns. Thefmax of a FETis inversely proportional to the gate length.


If the length is very short (less than 0.05 mi-crons for GaAs), then quantum effects play arole in device operation.

gate turn off device (1) any power semi-conductor switching device which can turnedoff with a signal to its gate.

(2) a thyristor capable of gate-turn-off op-eration.

gate turnoff thyristor a thyristor that canbe turned off with a negative gate currentwhile conducting current in the forward di-rection.

gate width the width in microns of thatportion of a FET channel associated with asingle gate, as measured in the direction or-thogonal to channel current flow, that can bepinched off by application of the proper con-trol bias. A single FET structure may includemany subcells placed in parallel with eachother, resulting in a much higher gate periph-ery for more power (i.e., 4 individual FETs,each with a gate of 100 microns, may be par-alleled together to form a 4× 100 micronFET with a gate periphery of 400 microns).

gate-to-drain breakdown voltage thebreakdown voltage of the reverse biased gate-to-drain junction of an FET. Usually it isspecified at a predetermined value of currentper millimeter of gate periphery.

gate-to-drain capacitance the capaci-tance between the gate and drain terminalsof a FET. It is formed primarily by the ca-pacitance of the FET’s physical interconnectstructure (extrinsic fixed capacitance) and bythe depleted region capacitance (intrinsic ca-pacitance), which is a function of voltage andtemperature.

gate-to-source breakdown voltage thebreakdown voltage of the reverse-biasedgate-to-source junction of an FET. Usually itis specified at a predetermined value of cur-rent per millimeter of gate periphery.

gate-to-source capacitance the capaci-tance between the gate and source terminalsof a FET. It is formed primarily by the ca-pacitance of the FET’s physical interconnectstructure (extrinsic fixed capacitance) and bythe depleted region capacitance (intrinsic ca-pacitance), which is a function of voltage andtemperature.

gate-to-source voltage the potential dif-ference between the FET gate and sourceterminals, this voltage controls the channelfrom saturation to pinchoff. This voltage isnormally negative for an n-channel FET, andpositive for a p-channel FET. However, ei-ther of these devices can be operated in aslightly enhanced mode, allowing low-levelexcursions of the opposite polarity.

gateway Seerouter.

Gauss’ law fundamental law of electro-magnetic field that states that the total elec-tric/magnetic flux through a closed surfaceis equal to the total electric/magnetic chargeenclosed.

Gauss’ theorem Seedivergence theorem.

Gaussian aperture aperture in which thetransmission profile is a Gaussian function ofradius.

Gaussian approximation an approxima-tion used in statistics where the distribution ofa sum of random variables is approximatedby the Gaussian distribution. Such an ap-proximation is based on the central limit the-orem of probability.

Gaussian beam electromagnetic beamsolution of the paraxial wave equation inwhich the field has spherical phase fronts andis a Gaussian function of distance from thebeam axis.

Gaussian distribution a probability den-sity function characterized by a meanµ and


covariance6:

f (x) = (2π)−N/2|6|−1/2

exp(−(x − µ)T 6−1(x − µ)/2

),

where|6| represents the determinant of6andN represents the dimensionality ofx. Ifx is scalar then the above function simplifiesconsiderably to its more familiar form:

f (x) = 1√2πσ

exp

(−1

2

(x − µ)2

σ 2

).

The Gaussian distribution is tremendouslyimportant in modeling signals, images, andnoise, due to its convenient analytic prop-erties and due to the central limit theorem.See alsoprobability density function, mean,covariance. See alsoCauchy distribution,exponential distribution.

Gaussian elimination the standard directmethod for solving a set of linear equations.It is termed direct because it does not involveiterative solutions. Variations of this schemeare used in most circuit simulators.

Gaussian mirror mirror in which the re-flection profile is a Gaussian function of ra-dius.

Gaussian noise a noise process that has aGaussian distribution for the measured valueat any time instant.

Gaussian process (1) a random processwhere the joint distribution of a set of randomvariablesX1, X2, . . . , Xn determined as val-ues of the process at the pointst1, t2, . . . , tnis ann-variate Gaussian distribution for allsets of pointst1, t2, . . . , tn, and all values ofthe integern.

(2) a random (stochastic) processx(t) isGaussian if the random variablesx(t1), x(t2),. . . , x(tn) are jointly Gaussian for anyn.

Gaussian pulse pulse in which the fieldis a Gaussian function of time from the pulsemaximum.

Gaussian sphere unit sphere, with anassociated spherical coordinate system, nor-mally used to represent orientations.

gaussmeter an instrument to measure theflux density due to a magnetic field.

Ge Seegermanium.

geared robot an arbitrary robot equippedwith gears is called a geared robot.Seedirectdrive robot.

genco a contraction of "generating com-pany," which is a company which generateselectric power but does not engage in trans-mission or distribution activities.

gender an adapter presenting two male ortwo female connectors for reversing the typeof cable connector. Connectors can be of thesame type or not.

general response formula for 2-D Roessermodel the solution to the 2-D Roessermodel [

xhi+1,jxvi,j+1

]=

[A1 A2A3 A4

][xhijxvij

]+[B1B2

]uij (1a)

i, j ∈ Z+ (the set of nonnegative integers)

yij = [C1, C2]

[xhijxvij

]+Duij (1b)

with boundary conditionsxh0j , j ∈ Z+ andxvi0, i ∈ Z+ is given by[xhijxvij

]=

i∑p=1

Ti−p,j[

0xvp0

]

+j∑q=1

Ti,j−q[xh0q0

](2)

+i∑

p=0

j∑q=0

(Ti−p−1,j−q

[B10

]


+ Ti−p,j−q−1

[0B2

])upq

wherexhij ∈ Rn1 andxvij ∈ Rn2 are the hori-zontal and vertical state vectors,uij ∈ Rm isthe input vector,yij ∈ Rp is the output vec-tor, A1, A2, A3, A4, B1, B2, C1, C2, D aregiven real matrices, and the transition matrixTpq is defined by

Tpq =

In for p = q = 0[A1 A20 0

]for p = 1, q = 0[

0 0A3 A4

]for p = 0, q = 1 and

T10Ti−1,j + T01Ti,j−1 fori, j ∈ Z+ (i + j > 0)0 for p < 0 or/and q < 0

Substitution of (2) into (1b) yields the re-sponse formula.

general response formula for generalized2-D model the solutionxij to the gener-alized 2-D model


+ B0uij + B1ui+1,j

+ B2ui,j+1 (1a)


i, j ∈ Z+ (the set of nonnegative integers)with admissible boundary conditionsxi0, i ∈Z+ andx0j , j ∈ Z+ is given by

xij =i+n1∑p=1

j+n2∑q=1

(Ti−p−1,j−q−1B0

+ Ti−p,j−q−1B1

+ Ti−p−1,j−qB2)upq

+i+n1∑p=1

(Ti−p,j−1 [A1, B1]

+ Ti−p−1,j−1 [A0, B0]) [ xp0up0

]

+i+n2∑q=1

(Ti−1,j−q [A2, B2]

+ Ti−1,j−q−1 [A0, B0]) [ x0qu0q

]+ Ti−1,j−1 [A0, B0] +

[x00u00

](2)

wherexi,j ∈ Rn is the semistate local vector,ui,j ∈ Rm is the input vector,E, Ak, Bk(k = 0,1,2) C, D are given real matriceswith E possibly singular and the transitionmatricesTpq are defined by

ETpq =

A0T−1,−1 + A1T0,−1+A2T−1,0 + I for p = q = 0A0Tp−1,q−1 + A1Tp,q−1+A2Tp−1,qfor p 6= 0 and/or q 6= 0

and

[Ez1z2 − A0 − A1z1 − A2z2]−1

=∞∑

p=−n1

∞∑q=−n2

Tpqz−(p+1)1 z

−(q+1)2

pair (n1, n2) of positive integersn1, n2 suchthatTpq = 0 for p < −n1 and/orq < −n2is called the index of the model. Substitutionof (2) into (1b) yields the response formula.

general-purpose motor term often usedto describe National Electrical Manufactur-ers Association (NEMA) class B, and lessoften class B, induction motors. General-purpose motors are those typically used whenrelatively low starting currents, low slip,good speed regulation, moderate startingtorque, and high efficiency are the predomi-nant concerns.

general-purpose register a register thatis not assigned to a specific purpose, suchas holding condition codes or a stack pointer,but that may be used to hold any sort of value.General-purpose registers are typically notequipped with any dedicated logic to oper-ate on the data stored in the register.

general-use switch a manually operatedswitching device designed for general use. Itis designed to interrupt rated current at ratedvoltage.


generalization the process of inferringrules and taking decisions after a learningphase has taken place. The process is sup-posed to take place on data not used duringlearning.

generalized cone data structure for volu-metric representations, generated by sweep-ing an arbitrarily-shaped cross section alonga 3-D line called the “generalized cone axis.”

generalized delta rule the weight updaterule employed in the backpropagation algo-rithm.

generalized Lloyd algorithm (GLA) ageneralization of the Lloyd (or Lloyd–Max)algorithm for scalar quantizer design to op-timal design of vector quantizers.See alsoK-means algorithm.

generalized modus ponens generaliza-tion of the classicalmodus ponensbased onthe compositional rule of inference.

Let be a fuzzy ruleA −→ B, that canbe interpreted as a fuzzy relationR, and afuzzy setA′, then the compositional rule ofinference maps a fuzzy set

B ′ = A′ ◦ R = A′ ◦ (A −→ B)

that can be interpreted as

premise1: x isA′(fact)premise2 : ifx isA then y isB(fuzzy rule)consequence : y isB ′(conclusion)

For example, if we have the fuzzy rule“If the tomato is red then it is ripe,”and weknow“The tomato is more or less red”(fact),the generalized modus ponens can infer“Thetomato is more or less ripe.”

See alsocompositional rule of inference,fuzzy inference, fuzzy relation, linguisticvariable, modifier.

generator in electrical systems, any ofa variety of electromechanical devices that

convert mechanical power into electricalpower, typically via Faraday induction ef-fects between moving and stationary current-carrying coils and/or magnets. Electrostaticgenerators use mechanical motion to physi-cally separate stationary charges to producea large electrostatic potential between twoelectrodes.

generator coherency a group of gener-ators where the rotor angles swing in syn-chronism with one another following a dis-turbance. Usually, generators in close elec-trical proximity and at some distance fromthe fault tend to be coherent.

generator differential relay a genera-tor differential relay is a differential relayspecifically designed for protection of elec-tric power generators. Variations include al-lowances for split-phase winding machines.

generator inertia constant a term pro-portional to the combined moment of inertiaof the turbine-generator mass.

generator matrix a matrix used to de-scribe the mapping from source word to codeword in a linear forward error control code.The mapping is described through multipli-cation of the source word by this matrix usingelement-wise finite field arithmetic. A linearcode is completely specified by its generatormatrix.

generator polynomial uniquely specifiesa cyclic code and has degree equal to the num-ber of the parity bits in the code. For an(n, k)cyclic code, it is the only code word polyno-mial of minimum degree(n− k).

genetic algorithm an optimization tech-nique that searches for parameter values bymimicking natural selection and the lawsof genetics. A genetic algorithm takes aset of solutions to a problem and measuresthe “goodness” of those solutions. It thendiscards the “bad” solutions and keeps the“good” solutions. Next, one or more genetic


operators, such as mutation and crossover,are applied to the set of solutions. The “good-ness” metric is applied again and the algo-rithm iterates until all solutions meet certaincriteria or a specific number of iterations hasbeen completed.See alsooptimization.

genlock a shortened term for “generatorlock,” meaning that one sync generator sys-tem is locked to another. When video sys-tems are genlocked, sync and burst informa-tion is the same for both systems.

geodesic lens a lens composed of cir-cularity symmetric spherical depression (ordimples) in the surface of a thin-film waveg-uide. In this type of lens, light waves are con-fined in the waveguide and follow the longercurved path through the lens region. Wavespropagating near the center of the lens travela longer path than waves near the edge, mod-ifying the wave front so that focusing canoccur.

geometric distribution a discrete proba-bility density function of a random variablex that has the form

p{x = k} = p(1 − p)kk = 0,1,2, . . . .

This is the probability thatk independent tri-als, each with probability of successp andfailure 1− p, fail before one succeeds.Seealsoprobability density function.

geometric Jacobian the Jacobian (ormore precisely the Jacobian matrix, butroboticists usually shorten it to simply Jaco-bian) is a mapping from velocities in jointspace (generalized velocities) to velocitiesin Cartesian space. This mapping is writ-ten in the formv = J (q)q where v is asix-dimensional vector of linear and angu-lar velocities of the end-effector andq isa vector of generalized velocitiesq(t) =[q1(t), q2(t), . . . , qn(t)]T . The Jacobian hasdimensions 6× n. All points in which theJacobian is not invertible are called singular-ities of the mechanism or singularities. The

Jacobian is defined for an arbitrary manipula-tor and depends on its geometric parameters(more precisely, Denavit–Hartenberg param-eters).See alsoanalytical Jacobian.

geometric radian center frequency thelogarithmic radian center frequency, it is thelogarithmic mid-point between the higher(ωH ) and lower (ωH ) band edges, expressedin radians/second. The band edges are usu-ally defined as the highest and lowest fre-quencies within a contiguous band of inter-est at which the loss equalsLAmax , the max-imum attenuation loss across the band.

ωOG = √ωH · ωL

or

log10(ωOG) = log10(ωH )+ log10(ωL)

2

geometric theory of diffraction a cor-rection to geometrical optics that includesdiffracted fields due to corners or edgesand accounts for energy diffracted into theshadow region.

geometric transformation transformsthe pixel co-ordinates of an image to effecta change in the spatial relationships of ele-ments in the image. The change often takesthe form of a stretching or warping of theimage.

geometrical optics a high-frequencytechnique using ray tracing for tubes of raysto determine incident, reflected, and refractedfields. Most useful in real media when thewave amplitude varies slowly compared tothe wavelength.

geosynchronous orbit an orbit 22,753miles above the earth in which an object willorbit the earth once every 24 hours above theequator and will appear to be stationary fromthe earth’s surface.

geothermal energy thermal energy in theform of hot water and steam in the earth’scrust.


germanium (Ge) an optical material forconstruction of components and systems atinfrared wavelengths, especially in the 8 to 14micron region. Also, an acousto-optic mate-rial for infrared wavelengths.

Gershgorin circle measures the relativesize of the off-diagonal elements in a transferfunction matrix at a given frequency. Whenevaluated over a range of frequencies, thesecircles sweep out a band, centered on thediagonal elements, that is used in the pre-diction of closed loop stability for multi-input–multi-output systems. This theory isbased on Gerhsgorin’s theory on the boundsof matrix eigenvalues. See alsodiagonaldominanceandinverse Nyquist array.

GFCI Seeground-fault circuit interrupter.

GFI Seeground fault interrupter.

GFLOP Seegigaflop.

ghosting the formation of an image inwhich a ghost image (i.e., a similar, fainter,slightly displaced image) appears superim-posed on the intended image; ghosts resultfrom a form of crosstalk, and typically ariseby radio transmission via an alternative pathoccurring as a result of random reflection.

giant magnetoresistance effect hugechange in electrical resistance with relativelysmall change in magnetic field exhibited bymultilayers of many combinations of metal-lic materials. Used in the next generation ofhard disk memory.

giant pulse large output pulse from alaser oscillator that results when the cavitylosses are quickly raised from a high valueholding a pumped laser below threshold tolow value bringing the laser above threshold(Q-switching or loss-switching) or when thegain occurs in the form of a short pulse (gainswitching).

Gibbs phenomenon the rippling phe-nomenon in the reconstruction of a functionaround a discontinuity based on the Fourierseries of the function. Specifically, given adiscontinuous periodic functionf which issquare-integrable on its period, the Fourierseries off converges tof in the square meansense. When the periodic function corre-sponding to then first terms of that Fourierseries is built, then ripples appear at the vicin-ity of each discontinuity off . Whenn tendsto infinity, these ripples tend to zero in thesquare mean sense, but their maximum am-plitude does not tend to zero; this producesan overshoot by a factor of 1.1789797 andan undershoot by a factor of 0.9028233. Asimilar effect arises for a square-integrablenonperiodic function reconstructed from itsFourier transform. The phenomenon causesringing around edges in low bit rate imagecoding. SeeFourier integral, Fourier series,norm.

Gibbs random field a class of randomfields described by normalized exponentialsof potential functions over cliques. LetS =s1, s2, . . .be points on a lattice, and letω ∈ �be a sample of a random field on this lattice.Points on the lattice interact with other latticepoints in a local, predefined manner; that is,the energy of the lattice is given by the sumof the energies of locally-interacting sets ofpoints in the lattice: each such set of pointsis known as a cliqueC ∈ C. Specifically,the energy of cliqueC in sampleω isVC(ω),whereV is a potential function dependingonly upon the points inC. Then the totalenergy of a sampleω is given by

E(ω) =∑C∈C

VC(ω).

With this definition, a probability measureπis a Gibbs random field if

π(omega) = 1

Ze−E(ω)/T

whereZ = ∑ω e

−E(ω)/T is a normalizationconstant.


Gibson mix an analysis of computer ma-chine language instructions that concludedthat approximately 1/4 of the instructions ac-counted for 3/4 of the instructions executedon a computer.

GIF See Graphics Interchange Format.

gigaflop (GFLOP) 1000 million floatingpoint operations per second.

Gilbert cell a four-transistor configura-tion combining the differential pair and cur-rent mirror concepts. With appropriate sig-nal conditioning at the input and output ter-minals, the cell can be used for many analogsignal processing applications such as analogmultiplication.

Gilbert, William (1544–1603) Born:Colchester, Essex, England

Gilbert is best remembered as an earlyinvestigator into electric charge and mag-netism. He is also considered by manyto be the inventor of the modern scientificmethod. This is due to his rigorous experi-mental methodology, and the detailed recordshe kept on his investigations. Isaac New-ton and Francis Bacon both acknowledgedhis contributions in this regard. Gilbert was,by training, a physician and held the post ofroyal physician in the courts of Elizabeth Iand James I.

Givens transformation a transformation,proposed by Givens, that transforms a gen-eral matrix to a triangular form. The GivenstransformationsGijθ are functions of threeparameters. Transforming a vectorb, a =Gijθb; thenan = bn for all n 6= i, n 6= j ,and the two-vector[aiaj ] is equal to the ro-tation of vector[bibj ] by an angleθ in theplane. Givens rotations can be used to suc-cessively set elements of a matrix to zero byan appropriate selection ofijθ .

GKS Seegraphical kernel system.

GLA Seegeneralized Lloyd algorithm.

glass laser laser in which the host mediumfor doping with laser atoms is a glass.

glitch (1) an incorrect state of a signal thatlasts a short time compared to the clock pe-riod of the circuit. The use of “glitch” in de-scribing power systems is generally avoided.See alsohazard.

(2) slang for a transient that causes equip-ment crashes, loss of data, or data errors.

global alignment a method of alignmentwhere the mask is aligned globally to thewhole wafer (as opposed to field-by-fieldalignment).

global interconnection interconnectionin which every source is connected to alldetectors and every detector is connected toall sources. Global interconnection is easilyimplementable using optics because, unlikeelectrons, photons do not interact with eachother, and an optical system is inherently aparallel processor.

global memory in a multiprocessor sys-tem, memory that is accessible to all pro-cessors.See alsolocal memory, distributedmemory.

global minimum a point at which a func-tion attains its lowest value over the domainof its arguments.

global observability of generalized 2-Dmodel the generalized 2-D model

Exi+1,j+1 = A1xi+1,j + A2xi,j+1

+ B1ui+1,j + B2ui,j+1

yij = Cxij

is called globally observable if any of itsglobal semistates

X(q) := {xi−q,−i , i = . . . ,−1,0,1, . . . ;

q = 0,1, . . .}can be calculated using future outputs andinputs of the model, i.e.,


U(k) := {ui+k,−i , i = . . . ,−1,0,1, . . . ;

k = 0,1, . . .}Y (k) := {

yi+k,−i , i = . . . ,−1,0,1, . . . ;k = 0,1, . . .} for k ≥ q

global positioning system (GPS) systemof 18 primary satellites in medium earth orbit,distributed so that at least four are simultane-ously visible from each point on the globe;typically used in timing and positioning ap-plications.

global reconstructibility of generalized 2-D model the generalized 2-D model

Exi+1,j+1 = A1xi+1,j + A2xi,j+1

+ B1ui+1,j + B2ui,j+1

yij = Cxij

is called globally reconstructible if any of itsglobal semistates

X(q) := {xi−q,−i , i = . . . ,−1,0,1, . . . ;

q = 0,1, . . .}may be calculated using past outputs and in-puts of the model, i.e.,

U(k) := {ui+k,−i , i = . . . ,−1,0,1, . . . ;

k = 0,1, . . .}Y (k) := {

yi+k,−i , i = . . . ,−1,0,1, . . . ;k = 0,1, . . .} for k ≤ q

Global System for Mobile Communica-tions (GSM) a set of systems specifica-tions that describe the Pan European digi-tal mobile cellular radio system. This set of13 recommendations describe service, syn-chronization, hardware, operating and main-tenance characteristics for the system.

globally asymptotically stable equilibriuman asymptotically stable equilibrium (see thedefinition) with a region of attraction (see thedefinition) equal to<n (n-dimensional realEuclidean space).

globally asymptotically stable state Seeasymptotically stable in the large.

glove in power line work, this refers toconductors which are energized with volt-ages low enough to be safely contacted byworkers wearing suitable rubber gloves, tothe range of voltages at which the practice isallowed, and to the gloves themselves.

GMAW Seegas metal arc welding.

GMSK See minimum-shift keyingGaussian.

goat head the top of a tower.

Gold book SeeIEEE Color Books.

Gold sequences a set of spreading se-quences developed by the coding theorist R.Gold in 1966 that are typically used in a mul-tiple access system utilizing direct sequencespread spectrum. The set of Gold sequencesis the solution of a problem in sequence de-sign, where the criteria is the minimization ofthe maximum cross-correlation between anytwo sequences in the set under all possiblecyclic shifts of one sequence relative to theother.See alsocross-correlation.

Gompertz dynamics the simplest nonlin-ear, with sigmoidal growth function, modelof population dynamics. The unperturbedgrowth of population is described by the first-order differential equation

x = gxln(xmax

x

)wherex is an average number of individualsin the population with maximal value (theso-called plateau population)xmax andg isa positive growth parameter. The Gompertzequation may be simplified using the dimen-sionless scaling

y = ln(xmax

x

)which leads to a linear equation of the form

y = −gy


The simplest way to introduce perturbationsto the dynamics represented by affine con-trol actions is to add a bilinear term into theGompertz equation

x = gln(xmax

x

)− kux

whereu is a control variable andk representsa perturbed population loss (or growth) pa-rameter. By applying the dimensionless scal-ing, the equation could be once more trans-formed into the linear one with respect tovariabley. The unperturbed Gompertz dy-namics could also be represented by two first-order differential linear equations

x = q(t)x; q = −gq; x(0) = x0; q(0) = q0

with x0e− q0

g = xmax.

Gordon–Haus limit expresses the max-imum performance of a soliton-based com-munication system as a bitrate-length prod-uct.

GOS Seegrade of service.

governor a device connected to a rotat-ing machine by which the speed-regulatingsystem is automatically adjusted to maintainconstant speed under various load conditions.

governor power flow the inherent re-sponse of prime movers or governors to achange in the operating condition of thepower system in an attempt to balance thepower equation. All generators participate atsome level in the change, with larger gener-ators picking up relatively larger amounts.

GPS Seeglobal positioning system.

grade of service (GOS) a probabilisticmeasure of service or equipment availabilityin a telecommunications network, expressedas the probability of a particular service (e.g.,a completed telephone connection) being de-nied at request time.

graded index medium in which the indexof refraction varies as a function of position;

usually refers to variations transverse to thedirection of propagation, and quadratic trans-verse variations are especially important.

graded index fiber an optical fiber wherethe core refractive index varies from a maxi-mum at its center to a minimum value at thecore-cladding interface. The profile is typ-ically designed so that all modes of a mul-timode fiber propagate at approximately thesame velocity.See alsostep index fiber.

graded index lens cylindrical optical el-ements with a refractive-index profile acrossthe cross section of the element to provide thetransfer; lens in which the transmitted wavesare focused or defocused by transverse vari-ations of the index of refraction rather thanby transverse variations of the thickness.

graded index optical fiber an opticalfiber with a refractive index in the corethat decreases monotonically from the fiberaxis to the interface between the core andcladding. The index usually goes froma higher value in the center of the coreto a lower value of refractive index at thecore/cladding interface. See alsogradedindex profile.

graded index optics optical elements thatuse a refractive-index profile across the cross-section of the element, typically a cylinder,to provide particular transfer functions as forlenses; the profile is generally produced byion implantation. Also called GRIN optics.

graded index profile an index of refrac-tion in an optical fiber core that decreaseswith distance from the core axis out to thecore/cladding boundary. The variation of theindex of refraction with the radial distancefrom the fiber axis can be given approxi-mately as a power law.

gradient a vector function denoted by∇for grad f , wheref is a continuous, differ-entiable scalar function. For a 2-D function


f (x, y), the gradient is

∇f = ∂f∂x

∂f∂y

The magnitude of this gradient is

|∇f | =√(

∂f

∂x

)2

+(∂f

∂y

)2

.

In image processing, the term gradient oftenrefers to the magnitude of the gradient.SeealsoSobel operator.

gradient descent a method for findingthe minimum of a multidimensional functionf (x). The technique starts at some point andadvances towards the minimum by iterativelymoving in a direction opposite to that of thegradient:

xi+1 = xi − α∂f

∂x

whereα is some scalar, usually set empiri-cally. See alsogradient, optimization.

gradient edge detector an edge detec-tor that defines an edge to be present at apixel only if the magnitude of the gradient atthat pixel is greater than some threshold.Seealso edge, edge detection, gradient, Sobeloperator.

gradient index optics optical compo-nents, e.g., optical fibers, within which therefractive index changes gradually betweentwo extremes.

gradient space a 2-D representation of theorientation of a 3-D surface in space, in whichthe two components of a pointP ≡ (p, q)arethe first-order partial derivatives of a surfaceof the formz = f (x, y).

Gram-Schmidt orthogonalization a re-cursive procedure for whitening (decorrelat-ing) a sequence of random vectors.See alsowhitening filter.

grammar an ordered 4-tuple(G =T , V, P, S), whereT is said to be the set ofterminals,V is the set of variables or non-terminals,P a set of production rules andS ∈ V is called the start symbol. Terminalsare the symbols with which the strings of thelanguage are made. For example,T may bethe English alphabet, the ASCII character set,or the set{0,1}. Non-terminals are symbolsthat are replaced by a string of zero or moreterminals and non-terminals. The productionrules specify which strings can be used to re-place non-terminals. The symbolS is the firstsymbol with which every production starts.All the strings that can be generated fromSusing rules inP are said to be inL(G), thelanguage ofG.example:LetGbe the grammar(T , V, P, S)such that

T = {a, b}V = {S}

whereS is the start symbol and

P = {S → aSb, S → ε}

whereε is the empty string. The languageL(G) is the set of all strings of the formanbn. The → symbol signifies thatS canbe replaced with whatever follows the arrow.

grant signal a control signal on a bus thatgives permission to control the bus to anothermodule.

graph a coupleG = (E, V ) whereV isa set of nodes andE ⊆ V × V is a set ofedges. Graphs are widely used in modelingnetworks, circuits, and software.

graph search an optimization techniqueused to find the minimum cost path from astarting point to a goal point, through a graphof interconnected nodes. Each link betweennodes has an associated path cost, which mustbe selected based on the problem of interest.See alsooptimization.


graphic adapter an adapter for inter-facing the computer toward a monitor.SeeMDA, CGA, EGA, VGA.

graphic controller Seegraphic adapter.

graphical kernel system (GKS) a stan-dard for computer graphics recognized byboth ANSI and ISO. GKS defines the ma-nipulation of graphic objects, including theirvisualization, print, etc. All manipulation isperformed by regarding the graphic adoptionas a independent device driver. In this way,GKS applications can be executed on differ-ent kinds of graphics adapters.

graphics the discipline dealing with thegeneration of artificial images by a computer.Its two main aspects are geometric modeling,whose subject is the computational represen-tation of the geometry and topology of ob-jects and scenes, and rendering, which stud-ies the generation of images from the interac-tion of light and objects. Graphics has beengeneralized to the synthesis of animated im-age sequences, in which case one speaks ofcomputer animation. Also called computergraphics, image synthesis.

Graphics Interchange Format (GIF) apopular image-file format that compressesthe image with LZW coding. See alsofile format, image compression, Lempel-Ziv-Welch (LZW) coding.

The Graphics Interchange Format(c) is theCopyright property of CompuServe Incorpo-rated. GIF(sm) is a Service Mark property ofCompuServe Incorporated.

grating Seediffraction grating.

grating lobe a lobe in an antenna radiationpattern whose peak amplitude is equal to thatof the main lobe.

grating spectrometer an instrument thatprovides a spectral decomposition of a opti-cal source using a diffraction grating to givespatial dispersion of wavelengths.

gravitational torque torque that dependsupon the position of the robot in the gravita-tional field.

Gray book SeeIEEE Color Books.

Gray code a code in which each of asequence of code words differs by one bitfrom the preceding one, and the assignedvalue of each code word is one more thanthat of the preceding one. Such a codeavoids glitches (i.e., sharp momentary un-wanted spikes) when, in an electromechani-cal system, the sensors giving the code wordsare imperfectly aligned. For example, onepossible three-bit Gray code is:

000 001 011 010 110 111 101 100.

gray level the individual numerical valuecorresponding to a particular degree ofbrightness in a digital image, often on an 8-bit gray scale consisting of 256 gray levelsstretching from pure black to pure white.

gray level co-occurrence a means ofmeasuring texture and other brightness varia-tions in digital images by generating matriceswhich tabulate the frequencies with whichdifferent gray levels co-exist at different dis-tance vectors from each other.

gray level saturation the restriction ofimage gray scale so that intensities above acertain level become fixed at the white levelcorresponding to the highest numerical valueavailable within the current storage capacity,typically one byte, of each pixel.

gray scale (1) an optical pattern in discretesteps between light and dark.

(2) intrinsic modulation property of themarking technology that enables dots of ei-ther different size or intensity to be printed.

(3) in digital images, a method used torepresent varying levels of brightness. Typi-cally, gray images use 8-bit gray level for thegray scale so that for each picture element, 0represents pure black and 255 represent pure


Gray code

decimal digit 0 1 2 3 4 5Gray code 000 001 011 010 110 111

white with other gray levels in between theseextremes.

grayscale Seegray scale.

Green book SeeIEEE Color Books.

Green’s function the function that satis-fies a given differential equation having assource term a Dirac delta function.Seedeltafunction.

Green’s theorems consider a closed, reg-ular, surfaceS bounding a volumeV wherethe two scalar functionsφ andψ , continuoustogether with their first and second deriva-tives throughoutV and on the surfaceS, aredefined. Green’s first identity states∫

V

∇ψ · ∇φ dV +∫V

ψ∇2φ dV =∫S

ψ∂φ

∂ndS

while Green’s second identity takes the form∫V

(ψ∇2φ − φ∇2ψ

)dV

=∫S

ψ∂φ

∂ndS −

∫S

φ∂ψ

∂ndS

which is frequently referred to as Green’s the-orem.

grey level Seegray level.

grey scale Seegray scale.

grid refers to the regular array of verticaland horizontal wires used for interconnectingthe chip.

grid array a technique for combining theoutput of amplifiers or oscillators in spaceby using a two-dimensional spatial array ofelements placed on a uniformly spaced grid.

GRIN optics Seegraded index optics.

grip a twisted wire tie which secures awire to an insulator or other fixture.

grip teeth a set of jaws which secures awire to a hoist or come-a-longcfso that it canbe pulled into place.

gripple Seegain ripple.

grooved media on an optical disk, theembossment of the disk surface with groovessuch that the disk tracks are either the groovesthemselves or the regions between the groves.

ground (1) an earth-connected electricalconducting connection that may be designedor non-intentionally created.

(2) the electrical “zero” state, used as thereference voltage in computer systems.

ground bounce a transient variation in thepotential of the ground terminal of a logic de-vice caused by variations in the supply cur-rent acting on the ground impedance of thecircuit as seen by the device. Usually causedby simultaneous turnon of the pullup andpulldown sections of totem-pole outputs.

ground bounce noise ground bounce oc-curs when a large number of semiconduc-tor circuit components are mounted on acommon semiconductor chip substrate, sothat they are imperfectly insulated from eachother. In normal operation the substrateshould act as an insulator; however, duringcertain unusual fluctuations in signal levels,the systems power and ground connectionscan experience fluctuations, which affect theperformance of each component in a ran-dom way that has the characteristics of noise,much like capacitive coupling.


ground current the current that flows ina power system in a loop involving earth and(in some usages) other paths apart from thethree phases.

ground fault interrupter a protectivedevice used in commercial and residentialwiring which monitors equipment connectedto an electrical outlet and shuts off the powerwhen a ground fault in the equipment is de-tected.

ground fault neutralizer an inductorconnected between the neutral of Y windingsof a generator or transformer and ground. Itis tuned to the machine’s capacitance so as tominimize ground fault current.

ground lamp indicator lamp on electricaldistribution switchboards that darkens whena ground condition exists on one (or more) ofthe busses.

ground loop an undesired conductive pathbetween two conductive bodies in a radialgrounding system that are connected to acommon ground.

ground plane a perfectly or highly con-ducting half space beneath an antenna. Also,an unetched layer of metal on a printed circuitboard over which microstriplines and printedantennas are formed.

ground rod a metallic, rod-type electrodedesigned to be driven into the earth. It servesas an earth connection for grounding pur-poses. Other types of earth electrodes in-clude buried plates, rings, and grids. Forbuildings, its primary function is to keep theentire grounding system at earth potential.

ground wave a vertically polarized TEMwave propagating close to the ground. Itis one of the three modes of propagation(ground, sky, and space waves).

ground-fault circuit interrupter (GFCI)a device designed to detect ground-fault cur-

rent above a threshold value (several mil-liamperes) and then interrupt the source ofelectrical power by opening a circuit breakeror a set of contacts. GFCIs are designed forpersonnel protection and are generally avail-able in the form of circuit breakers and re-ceptacles.

ground-signal-ground (GSG) (1) config-uration of the contact tips of a coplanar mi-crowave probe. A single signal contact is po-sitioned in the center of two parallel groundplane contacts.

(2) a network of wires buried in the eartharound a electric power transmission linetower to reduce footing impedance.

grounded system an electrical distribu-tion system in which one of the normalcurrent-carrying conductors, often the neu-tral, is intentionally grounded.

grounding Seeground.

grounding transformer a transformerconnected to an otherwise ungrounded three-phase system for the purpose of providing apath for ground current flow. Zig-zag trans-formers and grounded wye-delta transform-ers can be used as grounding transformers.

group code a recording method used for6250 bpi (bits per inch) magnetic tapes.

group decision support system a specialdecision support system used in support ofgroups of decision makers that might be ge-ographically dispersed.

group delay the derivative of the radianphase with respect to frequency.

group detection a special strategy formultiuser detection in a multiple access sys-tem. The users get divided into a numberof detection groups according to an appro-priate criterion. The users in the designatedfirst group are detected using a predetermineddetection technique; the corresponding mul-


tiple access interference generated by thisgroup of users is estimated, and then sub-tracted from the received signal. This de-tection and cancellation is continued succes-sively until all users in all groups are detected.

group velocity (1) a quantity proved toindicate (in most dispersive media) the speedof energy propagation of wavepackets.

(2) in an optical fiber, the speed at which asignal superimposed on an optical wave prop-agates.

Groupe Special Mobile (GSM) a digitalcellular communications network standarddeveloped in Europe in the 1980s and nowimplemented in many countries throughoutthe world. The acronym originally denotedGroupe Special Mobile, after the French-based group that developed the system. Nowthe acronym denotes Global System for Mo-bile. GSM is one of a group of systems thatare usually referred to as second-generationcellular systems. The system uses 200 kHzchannels, GMSK modulation and frequencyhopping. GSM transmits digitally encodedspeech at a rate of approximately 13 kbpsand makes extensive use of channel error-correcting codes.

GSG Seeground-signal-ground.

GSM See Global System for Mobile Com-munications.

GSVQ gain-shape vector quantization.Seeshape-gain vector quantization.

GTAW Seegas tungsten arc welding.

GTO Seegate turnoff thyristor.

guaranteed cost control a methodologyof robust controller design whose objectiveis to ensure that the value of the performanceindex for the real system represented by thefamily of models with uncertainty chang-ing within a given set is guaranteed to benot greater than a given value. Guaranteed

cost control techniques are usually applied tonominally linear systems and quadratic costfunctionals.

guard band a design technique for a colorCRT intended to improve the purity perfor-mance of the device by making the lightedarea of the screen smaller than the theoreticaltangency condition of the device geometry.

guard digits those extra digits of the sig-nificand in a floating-point operand that mustbe retained in order to allow correct normal-ization and rounding of the result’s signifi-cand.

guarded machine classification describ-ing an electrical machine constructed with anopen frame in which rotating components orlive electrical parts are guarded by screens.Screen openings are restricted so that a rodwith a diameter of 0.75 inches cannot contactany part within 4 inches of the guarded open-ing, and a rod with a diameter of 0.5 inchescannot contact any part more than 4 inchesinside the guarded opening.

guide wavelength the distance over whichthe fields of propagating modes repeat them-selves in a waveguide.

guided scrambling an extension of theself-synchronizing scrambling line code pro-cedure in which the source bit stream is aug-mented with appropriately valued bits priorto scrambling in order to ensure that thescrambled sequence contains good line codeproperties.

guided wave interconnect a means forconnecting electrical systems such as com-puters and computer circuit boards optically,using optical fibers and optical circuits ofchannel waveguides.

guy a wire which extends at an angle froma utility pole to the ground in order to bracethe pole against toppling due to unbalancedforces from the utility lines it supports.


guy anchor one of several appliances usedto hold a guy under tension in the earth.

guy guard a metal or plastic plate or tube

mounted on a guy wire at ground level toincrease the visibility of the wire and de-crease its ability to cut into a person or objectpressed against it.


HH Seehorizontal.

H∞ design a group of robust controllerdesign methods based on the methodologyof the Hardy spaceH∞ consisting of allcomplex-valued functions of complex vari-able that are analytic and bounded in theopen right half-plane. The least bound thatmay be imposed on this function is itsH∞norm. Since the open right half-plane maybe replaced by the imaginary axisjω, H∞methods provide a direct generalization ofthe classical frequency domain approach tocontrol system design.

A standard problem is to design a con-troller that ensures the internal stability of theclosed-loop system and minimizes theH∞norm of the transfer function between the in-puts (reference signals, disturbances) and er-rors. Since this transfer function is equal tothe sensitivity function, such design resultsin optimal sensitivity. The standard problemis then transformed into an equivalent model-matching problem with a fixed, possibly un-stable transfer function derived from the plantand a “free parameter” stable compensatorto be chosen. The compensator is found byminimization of the supremum over all fre-quencies of the modeling error. Finally, anoptimal (or suboptimal) controller is synthe-sized based on the found optimal (or subop-timal) solution to the model-matching prob-lem. To meet specific dynamic objectives thetransfer functions are modified by pre- andpostfilters in the form of frequency dependentweighting functions. Although the primaryproblem is formulated in frequency domain,it may be solved both by input–output andstate space techniques. In the former case thealgorithms are based on spectral and inner–outer factorizations and approximation the-orems for complex functions. In the latter

case the problem may be attacked by linear-quadratic game theoretic approach resultingin a set of Riccati equations.H∞ (H infin-ity) methods may be used in robust stabiliza-tion, robust performance design, disturbanceattenuation, optimal tracking, model follow-ing, optimal sensitivity design, etc.

H infinity design SeeH∞ design.

H modes the wave solutions with zeroelectric field component in the direction ofpropagation. Also known as transverse elec-tric (TE) modes.

H parameters characterizes a microwavenetwork with an arbitrary number of ports byrelating the total voltages and currents at theports.

H-D curve SeeHurter–Driffield curve.

H-mode Seetransverse electric wave.

H-plane in measuring an antenna’s radi-ation pattern, the plane that is perpendicularto the current in the element and thereforecontains the magnetic field intensity vectorfield. This plane is perpendicular to the elec-tric field (E) plane cut.

H-plane sectoral horn a horn antennawhere the aperture is formed by flaring thewalls in the direction of the H-plane. Theelectric field (E) plane dimension is left un-changed from that of the waveguide feed.

H-tree a popular clock distribution treetopologically that resembles the H shape. Itintroduces the least amount of clock skewcompared to other distribution topologies.

Haar transform unitary transform map-pingN samples g(n) toN coefficientsG(k)in a way that corresponds to repeated two-point averaging and two-point differencing.The 2× 2 Haar transform is

H2 = 1√2

(1 11 −1

)


The scaling factor allows the same matrix tobe used for the inverse transform. The 4× 4Haar transform can be interpreted as follows:first apply the 2× 2 transform to two in-dependent pairs of samples; then apply the2×2 transform to the two average coefficientsjust computed. Larger Haar transforms areconstructed by continuing this process recur-sively.

The Haar transform yields coefficientsequal to the subband values generated bydyadic decomposition with the Haar wavelet.This transform has achieved rather less usethan the other transforms in this family,such as the Discrete Cosine, Fourier andHadamard transforms.

Haar wavelet the orthonormal waveletpair ( 1√

2, 1√

2), ( 1√

2, −1√

2). Analysis and syn-

thesis pairs are identical. This is the mostcompact wavelet pair. Dyadic subband de-composition with the Haar wavelets yieldscoefficients equal to those from the Haartransform.

Haas effect states that the first soundheard will mask subsequent short delay ar-riving sounds, the combination appearing asa louder source. Also called law of the firstwavefront.

hacker a person who explores computerand communication systems, usually for in-tellectual challenge, commonly applied tothose who try to circumvent security barri-ers (crackers).

Hadamard matrix ann×nmatrixH withelements±1 is a Hadamard matrix of ordernif HHT = nI , i.e., the rows are all mutuallyorthogonal, as are the columns. Hadamardmatrices can only exist forn = 1,2 or n aninteger multiple of 4. Hadamard matrices oforder 2i can be constructed by the recursion

H1 = (1)

H2n =(Hn HnHn −Hn

)

These are known as the Sylvester matrices.Because of their properties, Hadamard matri-ces find application in the theory of error con-trol codes, and code division multiple access.Named after Jacques Salomon Hadamard(1865–1963). See alsoorthogonal, Walshcover.

Hadamard transform a unitary trans-form mappingN samplesg(n) to N coeffi-cientsG(k)according to the transform matrixHN , where

H2 = 1√2

(1 11 −1

)and larger arrays are formed by the recursivedefinition

HN = 1√2

(HN

2HN

2

HN2

−HN2

).

The inverse transform is identical. TheHadamard transform was formerly used fordata compression because its entries are all1 or−1, allowing computation without mul-tiplications. In this context it is now super-seded by the discrete cosine transform.SeealsoWalsh–Hadamard transform.

half adder a logic circuit that produces thesum and carry outputs for two input signals.A half adder has no carry input.

half bridge amplifier a class-D amplifierbased on a half-bridge inverter configuration.Not suitable for amplification of DC to low-frequency signals because the capacitor legcannot provide unidirectional current.

half subtracter a logic circuit that pro-vides the difference and borrow outputs fortwo input signals. A half subtracter has noborrow input.

half-band filter a filter whose even-indexed coefficients are all zeros except theone in the filter center.

half-height point a point at which themembership grade is equal to 0.5.


half-life the average time needed for halfthe nuclei of a radioactive element to decay.

half-power bandwidth the frequency atwhich a lowpass filter amplitude responsefalls to 1√

2of the DC response is known as the

half-power point. The two-sided half-powerbandwidth is twice the half-power point.

half-wave rectifier a device that passespositive polarity portions of a signal andblocks negative polarity portions of an ACsignal. Ideally, for a sinusoidal inputvi(t) =Vm cos(ωt), the output equals the input whilethe input is positive and is zero while the in-put is negative.

half-wave symmetry a periodic functionthat satisfiesx(t) = −x(t − T/2).

half-wave voltage the voltage required toproduce an amount of refractive index changein a medium that will retard the phase of atraversing optical wave.

halftone technique of simulating continu-ous tones by varying the amount of area cov-ered by the colorant. Typically accomplishedby varying the size of the printed dots in re-lation to the desired intensity.

Hall effect the phenomenon wherebycharge carriers are displaced perpendicularlyto their drift velocity when current flows inthe presence of a magnetic field. The re-sulting shift in carriers inside the conductoror semiconductor produces a transverse Hallvoltage that is proportional to the strength ofthe magnetic field (for constant current).

Hall measurement a semiconductor char-acterization that uses a crossed electric andmagnetic field to yield information on theconductivity of a test sample.

Hall probe a device with a Hall effecttransducer to sense a magnetic field.

halogen one of the halide atoms knownfor being highly reactive.

halt instruction an instruction (typicallyprivileged) that causes a microprocessor tostop execution.

Hamming code an encoding of binarynumbers that permit error detection and cor-rection first discovered by Richard Hammingat Bell Laboratories.

A Hamming code is a perfect codes withcode word lengthn = 2m−1 and source wordlengthk = n−m for anym > 2. A Hammingcode can correct an error involving a singlebit in the binary number. It can also detectan error involving two bits.See alsoparity.

Hamming distance the number of digitpositions in which the corresponding digitsof two binary words of the same length dif-fer. The minimum distance of a code is thesmallest Hamming distance between any pairof code words. For example, if the sequencesare 1010110 and 1001010, then the Ham-ming distance is 3.

Hamming net a pattern recognition net-work that has a set of prototype patternsstored in its weights. A given input patternis identified with the prototype whose Ham-ming distance from the input pattern is least.

Hamming weight the number of nonzerosymbols in a given sequence of symbols.

Hamming window a tapering functionused to truncate functions spectrally or in thetime domain (i.e., in frequency it is a specifi-cally formed low pass filter). It is defined asw(x) = 0.54+0.46∗cos(2∗π ∗x/xo)|x| <xo/2. The Hamming window sacrifices fil-ter sharpness at the cutoff frequency for asmoother stop band behavior.

hand Seeend-effector.

hand-held computer a small lightweightcomputer that performs functions such as


electronic mail, handwriting recognition fortaking notes, and holding addresses and ap-pointments. Also called a “palm-top.”

hand-off in a cellular system, the pro-cess by which the mobile terminal switchesfrom communicating with one base stationto a communication with neighboring basestation as the mobile travels through the dif-ferent radio cells.

handline a rope used to pull tools andequipment from an assistant on the ground toa worker atop a utility pole.

handover Seehand-off.

handshaking I/O protocol in which a de-vice wishing to initiate a transfer first tests thereadiness of the other device, which then re-sponds accordingly. The transfer takes placeonly when both devices are ready.

handwritten character recognition theprocess of recognizing handwritten charac-ters that are clearly separated.

Hankel transform the 2-D Fourier trans-form of a function with circular symmetryand arises in the analysis of optical systems.The (zero-order) Hankel transformF(ρ) ofa functionf (r) for r ≥ 0 is

F(ρ) = 2π∫ ∞

0rf (r)J0(2πρr)dr

and the (zero-order) inverse Hankel trans-form is

f (r) = 2π∫ ∞

0ρF(ρ)J0(2πρr)dρ,

whereJ0(r) is the zero-order Bessel functionof the first kind, i.e.,

J0(r) = 1

π

∫ π

0cos(r sinθ)dθ.

See alsoFourier transform.

Hanning window a raised cosine windowfunction whose impulse response is

wn ={

0.5 + 0.5 cos(nπ/N) |n| < N

0 |n| ≥ N.

hard bake the process of heating thewafer after development of the resist in orderto harden the resist patterns in preparationfor subsequent pattern transfer. Also calledpostbake.

hard bug a name for a crimped copperwire connector.

hard contaminant a contaminant or for-eign object which is at least partly opaque toX-radiation: typical hard contaminants arepieces of metal, glass or stone.

hard decision demodulation that outputsa q-ary value for each demodulated symbolin a sequence ofq-ary symbols.See alsosoftdecision.

hard disk a rigid magnetic disk used forstoring data. A typically nonremovable col-lection of one or more metallic disks cov-ered by a magnetic material that allows therecording of computer data. The hard diskspins about its spindle while an electromag-netic head on a movable arm stays close tothe disk’s surface to read from or write to thedisk. Each disk is read and written on bothabove and below.N disks are read/written byusing 2N heads. The information is stored bycylinders, circular segments of the collectionof the disks. Cylinders are divided in sectorsas a pie. The mean time to access data istypically close to 10 msec.

Generally, hard disks are the backingmemory in a hierarchical memory.See alsofloppy disk.

hard fault Seepermanent fault.

hard ferrite Seeceramic ferrite.


hard magnetic material a ferromagneticmaterial that retains its magnetization whenthe magnetizing field is removed; a magneticmaterial with significant coercivity.

hard real-time Seehard real-time system.

hard real-time system a real-time systemin which missing even one deadline results insystem failure.Compare withsoft real-time,firm real-time.

hard X-ray having sufficient photon en-ergy to penetrate “hard” material; usuallymore than about 15 keV.

hard-decision decoding decoding of en-coded information given symbol decisionsof the individually coded message symbols.Compare withsoft-decision decoding.

hardware computer constructs that havea physical manifestation.

hardware accelerator a piece of hard-ware dedicated to performing a particu-lar function (such as image convolution ormatrix-vector products) which would other-wise be performed in software. Althoughmuch less flexible, dedicated hardware im-plementations can give significant speed im-provements over software, and are especiallyuseful for real-time applications.

hardware interrupt an interrupt gener-ated by a hardware device, for example, key-board, the DMA, PIC, the serial adapter, theprinter adapter, etc. Other hardware inter-rupts can be generated by the control unit orby the ALU, for example, for the presence ofa division per zero, for attempting to executean unknown instruction. This last class ofhardware interrupts is called internal excep-tion.

hardware noise radio frequency emis-sions due to arcing of utility lines at defectiveconnectors.

harmonic (1) the name associated with anumber used to denote the frequency com-ponents that exist in a certain fourier seriesrepresentation for a certain function of timef (t). The Fourier series representation isgiven

f (t) =∞∑

n=−∞Fn∗[cos(nωot)+ j sin(nωot)]

wheren = 0,1,2,3, . . ., ωo = 2π/T , j =√−1, T is the period of the functionf (t),Fn is the coefficient of the Fourier series fora certain value ofn:

1st harmonic hasFn = F1 and

[cos(1ω0t)+ j sin(1ω0t)]

2nd harmonic hasFn = F2 and


3rd harmonic hasFn = F3 and


etc.(2) sinusoidal component of a periodic

waveform that has a frequency equal to aninteger multiple of the basic frequency (orfundamental frequency). Thus the third har-monic of a power system voltage in the U.S.has a frequency of 3×60, or 180 Hz. For elec-tric systems powered by sinusoidal sources,harmonics are introduced by nonlinear de-vices such as saturated iron cores and powerelectronic devices.

harmonic amplifier a type of amplifierthat utilizes various forms of harmonic andmixing actions. These amplifiers may pumpup the fundamental by increasing the switch-ing efficiency of the active device. Othersmay actually be used as frequency multipliersor frequency converters (mixers). All classF, G, and H amplifiers fit into this general


group. Parameters such as device character-istics, quiescent bias point, RF load line, sig-nificant harmonic and/or mixing frequencies,and amplitude and waveform of the appliedsignal(s) should be included with the classdefinition, thus defining the major contribu-tors to the physical actions taking place inone of these amplifiers.

harmonic analysis the branch of math-ematics dealing with the decomposition ofsignal functions as a linear combination ofbasis functions which represent “waves” ofvarious frequencies. When the basis func-tions are sines and cosines each with a fre-quency that is an integer multiple of thesignal’s frequency, we have trigonometricharmonic analysis, in other words classicalFourier analysis, which provides the ampli-tudes and phases of the constituent sinusoids.( SeeFourier transform. ) With other basisfunctions, for example wavelets, we havenon-trigonometric harmonic analysis( Seewavelet, wavelet transform). Abstract har-monic analysis studies the generalization ofFourier analysis to abstract spaces.

harmonic balance technique one of sev-eral techniques for analyzing nonlinear cir-cuits. The nonlinear circuit is divided intotwo portions of linear and nonlinear ele-ments, and a portion of linear elements is cal-culated in a frequency domain and a portionof nonlinear elements is calculated in a timedomain, respectively. The calculated volt-ages or currents at connecting nodes of theseportions are balanced by using Fourier trans-forming or inverse Fourier transforming.

harmonic component a Fourier compo-nent of order greater than one of a periodicwaveform.

harmonic content the internally gener-ated, harmonically related spectral outputfrom a device or circuit. Harmonic energyis that energy that is at exact multiples ofthe fundamental frequency, generated by the

nonlinearities within the device or circuit act-ing on the fundamental frequency.

harmonic converter found in a mi-crowave receiver, this component uses thetechnique of harmonic mixing to convert theRF signal to a lower IF frequency for fur-ther processing. Harmonic converters can beused as part of a vector network analyzer.

harmonic distortion caused by the non-linear transfer characteristics of a device orcircuit. When a sinusoidal signal of a singlefrequency (the fundamental frequency) is ap-plied at the input of a nonlinear circuit, theoutput contains frequency components thatare integer multiples of the fundamental fre-quency (harmonics). The resulting distortionis called harmonic distortion.

harmonic frequency integral multiplesof fundamental frequency. For example, fora 60-Hz supply, the harmonic frequencies are120, 180, 240, 300,. . ..

harmonic generation in nonlinear op-tics, the process in which a laser beam in-teracts with a material system to producenew frequency components at integer multi-ples of the frequency of the incident beam.Under carefully controlled circumstances,the lower-order harmonics (e.g., second andthird) can be generated with high (> 50%) ef-ficiency. Under different circumstances, har-monics as high as the 30th can be generated.

harmonic load-pull measurement ameasurement method where transfer char-acteristics of a device at the fundamentalfrequency can be measured by electricallychanging the load impedance at harmonicfrequencies.

harmonic orthogonal set the set of func-tions ejωt . It is called harmonic becauseeach basis function is a harmonic of a cer-tain frequency and because the inner productbetween any two functions is zero:∫ +∞

−∞ ejω1t ejω2t dt = 0, ω1 6= ω2


∫ +∞−∞ ejωt1ejωt2 dω = 0, t1 6= t2.

harmonic tuning the process of tuning anamplifier circuit to a frequency that is an in-tegral multiple of the fundamental frequencyat which the circuit would normally operate.

harmonically pumped mixer mixerwhere the intermediate frequency (IF) sig-nal is at a frequency which is the sum ordifference of the RF and an integer multi-ple (usually two) of the LO (local oscillator)frequency.

Hartley oscillator a particular case of LC-oscillators whenX1 +X2 + 2Xm is realizedas a single tapped coil, andX3 is a capacitor.Well suited for variable-frequency operationby varying a single capacitor.

Hartley oscillators are usually not used atVHFs of higher frequencies. Similarly, thecircuit is avoided at very low audio frequen-cies. It is important to distinguish the Hartleyoscillator from the Armstrong topology. Inthe Armstrong oscillator, no ohmic connec-tion exists between the two inductors. In-stead, coupling is entirely magnetic.

Harvard architecture a computer designfeature where there are two separate mem-ory units: one for instructions and the otherfor data only. The name originates from anearly computer development project at Har-vard University in the late 1940s.Comparewith Princeton architecture.

hash table a table storing a mapping func-tion whose domain is sparsely used and that isaccessed by indices that are computed fromthe search field (“key”) using a many–onemapping (called a hash function). Hash ta-bles are used for many memory and namemapping functions, such as symbol tables inassemblers and compilers.

hashed page table a page table wherethe translation of each virtual page numberis stored in a position determined by a hashfunction applied to the virtual page number.

This technique is used to reduce the size ofpage tables.

hashing the act of translating a search keyinto a table index using a many–one mapping.See alsohash table.

Hausdorff distance an important distancemeasure, used in fractal geometry, amongother places. Given a distance functionddefined on a Euclidean spaceE, one derivesfrom it the Hausdorff distanceHd on the fam-ily of all compact (i.e., bounded and topo-logically closed) subsets ofE; for any twocompact subsetsK,L of E,Hd(K,L) is theleastr ≥ 0 such that each one ofK,L is con-tained in the other’s dilation by a closed ballof radiusr, that is:

K ⊆⋃p∈L

Br(p) and L ⊆⋃p∈K

Br(p),

where

Br(p) = {q ∈ E | d(p, q) ≤ r}.

Hayes-compatible modem refers to amodem when it is capable of responding atthe commands of modems made by HayesMicrocomputer Products. The Hayes set ofcommands represents a sort of standard formodems.

haystack response bandpass frequencyresponse characterized by flat midband re-sponse with sloping sides.

hazard a momentary output error that oc-curs in a logic circuit because of input signalpropagation along different delay paths in thecircuit.

hazardous location a classification sys-tem used to define locations that are suscep-tible to fire and explosion hazards associatedwith normal electrical arcing. A class I haz-ardous location contains a flammable con-centration of flammable vapors. A class IIhazardous location contains a combustible


concentration of combustible dusts. A classIII hazardous location contains an ignitableconcentration of ignitable fibers.

hazardous outage an outage which hasbeen assessed to have potential life threat-ening consequences. The criteria for a haz-ardous outage varies considerably from oneutility to another. In some cases arcing linesis sufficient for a hazardous outage. Otherutilities would consider at least fire or explo-sion to qualify an outage as hazardous.

haze tails when self-converging deflec-tion yokes are used on a CRT, overfocusingoccurs on the vertical center line of the beamspot. Around the periphery of the screen, thisproduces a hazy area above and below thespot, referred to as haze tails. This degradespicture contrast.

HBMT Seehybrid bipolar MOS thyristor.

HBT Seeheterojunction bipolar transistor.

HDA Seehead disk assembly.

HDSL Seehigh-speed digital subscriberline.

HDTV Seehigh definition television.

head an electromagnet that producesswitchable magnetic fields to read and recordbit streams on a platter’s track.

head disk assembly (HDA) collection ofplatters, heads, arms, and acutators, plus theair-tight enclosing of a magnetic disk system.

head-medium gap the distance betweenthe read–write head and the disk in magneticor optical disk memory devices.

header (1) a data structure containing con-trol information that is placed at the headof a datagram; when placed at the end of a

datagram this is referred to as a trailer. Ad-dressing information and checksum are ex-amples of the control information containedin a header.

(2) a section of an image file, usually offixed size and occurring at the start of the file,that contains information about the image,such as the number of rows and columns andthe size of each pixel.

heap data storage structure that acceptsitems of various sizes and is not ordered.Contrast with stack.

heat sink aluminum mountings of vari-ous shapes and sizes used for cooling powersemiconductor devices. Heat sinks can becooled by either natural convection or a fan,and heat dissipation can be improved by acoating of black oxide or if the heat sink ismade with fins.

heater Seeoverload heater.

Heaviside characteristic an activationfunction according to which the neuron out-put takes on the value unity when its weightedinput exceeds (or equals) the neural thresholdvalue and zero otherwise.

Heaviside, Oliver (1850–1925) Born:London, England

Heaviside is best known for his theoret-ical work in electrical engineering. Muchof his work is contained in his three-volumework calledElectromagnetic Theory.The fi-nal volume was published in 1912. Heavisideextended and improved the works of Hamil-ton and Maxwell and deduced the conceptsof capacitance, impedance, and inductance.Heaviside was self-taught and irascible. Atfirst, much of his work was dismissed as un-orthodox or too theoretical to be of practicalvalue. Heaviside is best known in physics forhis correct prediction of the existence of theionosphere.


heavy water water in which a heavy iso-tope of hydrogen substitutes for the hydrogenatoms.

Hebb’s principle to update the weights ofthe simple feedforward neural network (e.g.,perceptron) a very simple idea, called Hebb’sprinciple, can be used. The principle statesthe following: Apply a given pattern to theinputs and clamp the outputs to the desiredresponse, then increase the weights betweennodes that are simultaneously excited. Thisprinciple was formulated by Hebb in 1949, inan attempt to model neural networks. Math-ematically it can be expressed as follows:

wij (t + 1) = wij + γ ui(t)(y0j (t)− yj (t)

)wherey0

j is the desired response andyj isthe response of the network. By regardingthe weights as parameters, one may treat theabove formula as a gradient method for pa-rameters estimation.

Hebbian algorithm in general, a methodof updating the synaptic weight of a neuronwi using the product of the value of theithinput neuron,xi , with the output value of theneurony. A simple example is:

wi(n+ 1) = wi(n)+ αy(n)xi(n)

wheren represents thenth iteration andα isa learning-rate parameter.

Hebbian learning a method of modifyingsynaptic weights such that the strength of asynapic weight is increased if both the presy-naptic neuron and postsynapic neuron are ac-tive simultaneously (synchronously) and de-creased if the activity is asynchronous. In ar-tificial neural networks, the synaptic weightis modified by a function of the correlationbetween the two activities.

hedge a special linguistic term by whichother linguistic terms are modified. Exam-ples are very, fairly, more or less, and quite.

height defuzzification See centroiddefuzzification.

heirarchical coding coding where im-age data are encoded to take care of differentresolutions and scales of the image. Addi-tional data is transmitted from the coder torefine the image search.See alsoprogressivetransmission.

heirarchical interpolation a techniqueof forming image pyramids. In this method,pixels are interpolated to higher levels in thehierarchy. Thus, only the interpolated pixelsneed a residual in order to be reconstructedexactly, since the subsampled pixels are al-ready correct.

helical antenna a wire antenna that is he-lical in shape. Typically, the helical antennais fed by a coaxial cable, the outer conduc-tor of which is connected to a ground planeand the inner conductor of which protrudesthrough the ground plane and is shaped in theform of a helix. Since a helical antenna emitsan elliptically polarized wave when operat-ing in axial mode, it is often used in satellitecommunication applications.

helical beam tube a backward wave am-plifier based on interaction between a for-ward helical electron beam launched intoa waveguide and backward traveling mi-crowave in the waveguide.

helical scan tape a magnetic tape in whichrecording is carried out on a diagonal to thetape, by a head that spins faster than the tapemovement. Popular for VCRs, camcorders,etc., as it improves recording density and tapespeed.

helium neon laser neon laser in which ex-citation of the neon atoms occurs primarilyby collisions with electron-impact-excitedhelium atoms; the first gas laser, and espe-cially important for its CW visible oscillationlines.

Helmholtz coils a pair of coaxial coils ofthe same diameter that are spaced one radiusapart. They are used to generate a field of


uniform strength or to measure the magneticmoment of a magnet.

Helmholtz equation a partial differentialequation mathematically described by

(∇2 + k2

)φ = 3Df ,

where∇2 is the Laplacian,k is the wavenum-ber, f is the forcing function, andφ is theequation’s solution.

HEMT See high electron mobilitytransistor.

Henry, Joseph Henry is best known asthe first Director (1846) of the SmithsonianInstitution, and President of the NationalAcademy of the Sciences. Henry was largelyself-taught, but his early experiments gar-nered him sufficient recognition to become aProfessor of Natural Philosophy at New Jer-sey College (now Princeton). Henry’s earlyexperiments resulted in the development of apractical electric motor and a relay later quiteimportant in telegraphy.

hermetic seal a seal that is such that theobject is gas-tight (usually a rate of less than1 × 10−6 cc/s of helium).

Hermite form of 2-D polynomial matrixdenote byFm×n(z1) [z2] (Fm×n [z1] [z2])

the set ofm × n polynomial matrices inz2with coefficients in the fieldF(z1) (poly-nomial coefficients inz1). 2-D polyno-mial matrix A(z1, z2) ∈ Fm×n [z1, z2] offull rank has Hermite form with respect to

Fm×n [z1] [z2] if

AH(z1, z2) =

a11 a12 ... a1n0 a22 ... a2n... ... ... ...

0 0 ... ann0 0 ... 0... ... ... ...

0 0 ... 0

if m > n

a11 a12 ... a1n0 a22 ... a2n... ... ... ...

0 0 ... ann

if m = n

a11 a12 ... a1m ... a1n0 a22 ... a2m ... a2n... ... ... ... ... ...

0 0 ... amm ... amn

if m < n

where degz2 aii > degz2 aki for k 6= i (degz2denotes the degree with respect toz2). In asimilar way, the Hirmite form ofA(z1, z2)

with respect toF [z1] [z2] can be defined.A(z1, z2) can be reduced to its Hermite formAH(z1, z2) by the use of elementary rowoperations or equivalently by premultiplica-tion by suitable unimodular matrixU(z1, z2)

(detU(z1, z2) ∈ F(z1)), i.e.,AH(z1, z2) =U(z1, z2)A(z1, z2). See for example, T.Kaczorek,Two-Dimensional Linear Systems,Springer-Verlag, Berlin, 1985.

Hermite Gaussian beam electromag-netic beam solution of the paraxial waveequation in which the field is a product ofa Hermite-polynomial and a Gaussian func-tion of distance from the beam axis.

Hermitian matrix a square matrix thatequals its conjugate transpose.

hertz a measure of frequency in whichthe number of hertz measures the number ofoccurrences (of whatever is being measured)per second.

Hertz dipole a straight, infinitesimallyshort and infinitesimally thin conducting fil-ament with uniform current distribution. The


current amplitude is usually assumed to varysinusoidally in time.

Hertz, Heinrich Rudolf (1857-1894)Born: Hamburg, Germany

Hertz is best known as the person who dis-covered radio waves. Hertz had several pro-fessorships including posts at Berlin, Kiel,Karlsruhe, and Bonn. Hertz’s work was boththeoretical and experimental. He succeededin proving important elements of Maxwell’sequations, and was able to demonstrate thatradio waves could be generated. He diedearly from a degenerative bone disease. Wehonor his name by referring to the hertz as aunit of frequency.

Hertzian dipole SeeHertz dipole.

Hertzian potential potentials used for thecomputation of electromagnetic field. By us-ing the Hertzian vector potentials of electricand magnetic type5e,5h, respectively, thefields are expressed as

E = k25e + ∇∇ ·5e − jωµ∇ ×5h

H = jωε∇ ×5e + k25h + ∇∇ ·5hwhereε is the dielectric permittivity,µ themagnetic permeability, andk = ω

√εµ, with

ω denoting the angular frequency.

heterodyne detection method of measur-ing the frequency content of an optical sig-nal in which that signal is combined with aknown CW reference signal from a local os-cillator by means of a square-law detector;sometimes provides a more direct and sensi-tive representation of the frequency contentof a signal than homodyne detection.

heterodyne receiver a receiver where thelow-level RF signal is mixed with a localoscillator signal to produce an intermediatefrequency (IF). The IF frequency is usuallybetween 10 to 100 MHz. It can be ampli-fied with a low noise amplifier. Heterodynereceivers have much better sensitivity andnoise characteristics than direct detection re-ceivers.

heterogeneous having dissimilar compo-nents in a system; in the context of comput-ers, having different types or classes of ma-chines in a multiprocessor or multicomputersystem.

heterojunction a junction between twocrystals of different bulk composition, typ-ically referring to semiconductor–semicon-ductor interfaces. The prototype example isAlAs/GaAs.

heterojunction bipolar transistor (HBT)proposed by Shockley (U.S. Patent 2569347).

heterojunction field effect transistor(HFET) a field effect transistor that usesa heterojunction in the channel parallel tothe current flow direction to improve thecarrier transport by separating the dopantfrom conduction region. Other names forthis device include MODFET, (MOdulationDoped Field Effect Transistor and HEMT,High Electron Mobility Transistor.

heterojunction laser Seediode laser.

heterostructure refers to a compositestructure in which two dissimilar materialsare chemically joined at a common interface.Examples are GaAs-AlGaAs, metal on semi-conductor, oxide on semiconductor, etc.

Heuristic special problem-specific knowl-edge that can be used to help direct a searchefficiently towards an appropriate, thoughpossibly suboptimal, solution.Seeheuristicsearch.

heuristic search a search that uses prob-lem-specific knowledge in the form of heuris-tics in an effort to reduce the size of the searchspace or speed convergence towards a solu-tion. Heuristic searches sacrifice the guaran-tee of an optimal solution for reduced searchtime.


hexadecimal notation expressing num-bers in base-16 format.See alsohexadecimalnumber system.

hexadecimal number system the num-bering system that uses base-16, commonlyused in computer systems. The digits 10–15are generally represented by the letters A–F.

hexagonal pixel Seepixel.

HFET See heterojunction field effecttransistor.

HgCdTe semiconductor alloy importantas the active photoconductive or photovoltaicelement in mid- to long-wavelength infrareddetectors.

HgTe/CdTe heterostructure combinationwith properties similar to the HgCdTe alloy,but with added tailorability from the abilityto modify properties such as band gap layerby layer.

hi-bi fiber optical fiber with high lin-ear birefringence. Light launched on one ofthe fiber’s polarization axes will maintain itspolarization state as it propagates along thelength of the fiber.

hidden layer a layer of neurons in a multi-layer perceptron network that is intermediatebetween the output layer and the input layer.

hidden Markov model (HMM) adiscrete-time, discrete-space dynamical sys-tem governed by a Markov chain that emits asequence of observable outputs, usually oneoutput (observation) for each state in a trajec-tory of such states. More formally, a HMMis a five-tuple (�X,�O,A,B, π), where�Xis a finite set of possible states,�O is a finiteset of possible observations,A is a set of tran-sition probabilities,B is a set of observationprobabilities, andπ is an initial state distri-bution.

hidden station problem problem relatedto multiple access protocols in mobile envi-ronment where two mobile stations may notbe able to detect each other’s transmission,leading to the possibility of a transmissioncollision.

hidden unit a neural unit in a network thathas no direct connection to either the networkinputs or to the network outputs.

hierarchical clustering a clustering tech-nique that generates a nested category struc-ture in which pairs or sets of items or cat-egories are successively linked until everyitem in the data set is connected. Hierar-chical clustering can be achieved agglom-eratively or divisively. The cluster struc-ture generated by a hierarchical agglomer-ative clustering method is often a dendro-gram, but it can be also as complex asany acyclic directed graph. All conceptformation systems are hierarchical cluster-ing systems, including self-generating neu-ral networks. See alsoclustering, conceptformation, self-generating neural network.

hierarchical coding coding a signal atseveral resolutions and in order of increasingresolution. In an hierarchical image coderan image is coded at several different sizesand in order of increasing size. Typically,the smaller sized images are used to encodethe larger size images to obtain better com-pression.

hierarchical control operation or struc-ture of the control system, where the con-troller, or the given control layer, is com-posed of several local decision units coordi-nated (Seecoordination) by a supremal unit(coordinator unit); hierarchical control in theabove narrow sense is tantamount to the mul-tilevel control, yet the term hierarchical con-trol is also used in broader meaning — cover-ing both multilevel and multilayer control aswell as integrated multilayer–multilevel con-trol structures occurring in large industrialand environmental applications.


hierarchical feature map a hybrid learn-ing network structure also called counter-propagation network that consists of a un-supervised (hidden) layer and a supervisedlayer. The first layer uses the typical (in-star) competitive learning, the output will besent to the second layer, and an outstar super-vised learning is performed to produce the re-sult. See alsocounter-propagation learning,outstar training.

Kohonen calls a feature map that learns hi-erarchical relations as a “hierarchical featuremap.”

hierarchical memory also known asmulti-level memory. The organization ofmemory in several levels of varying speedand capacity such that the faster memories lieclose to the processor and slower memorieslie further away from the processor. Fastermemories are expensive, and therefore aresmall. The memories that lie close the pro-cessor store the current instruction and dataset of the processor. When an object is notfound in the memories close to the proces-sor, it is fetched from the lower levels of thememory hierarchy. The top levels in the hi-erarchy are registers and caches. And thelowest level in the hierarchy is the backingmemory, which is usually a disk.

high byte the most significant byte of amultibyte numeric representation.

high definition television (HDTV) a setof standards for broadcast television thatincorporate all-digital transmission, greatlyimproved audio and video quality, audio andvideo compression, and integration of datawith television audio and video.

high electron mobility transistor (HEMT)a depletion-mode field-effect transistor in

which the channel is a two-dimensional elec-tron gas at a heterointerface. Other names:MODFET, TEGFET, and 2DEGFET.

high frequency emphasis filter a band-limited filter whose frequency response in-

creases with frequency. Typically such a fil-ter would have unity gain or less at low fre-quencies.

high order interleaving in memory inter-leaving, using the most significant addressbits to select the memory module and leastsignificant address bits to select the locationwithin the memory module.

high pass filter filter exhibiting fre-quency selective characteristic that allowshigh-frequency components of an input sig-nal to pass from filter input to output unatten-uated; all lower frequency components areattenuated.

high phase order (HPO) polyphase sys-tems that contain 6, 9, or more phases —rather than the standard three-phase system.HPO systems may be used to provide a meansof transmitting more electrical power downexisting right-of-ways than three-phase sys-tems, without an increase in transmissionvoltage, and without an increase in EMF lev-els at the edge of the right-of-way.

high rupturing capacity (HRC) a termused to denote fuses having a high interrupt-ing rating. Most low-voltage HRC-type fuseshave an interrupting rating of 200 kA RMSsymmetrical.

high side pertains to the portion of a cir-cuit connected to the higher voltage windingof a power transformer.

high state a logic signal level that has ahigher electrical potential (voltage) than theother logic state. For example, the high stateof TTL is defined as being greater than orequal to 2.0 V.

high-impedance state the third state oftri-state logic, where the gate does not driveor load the output line.Seetri-state circuit.

high-level transmitter a transmitter inwhich the modulation process takes place at


a point where the power level is at or near theoutput power.

high-level vision the highest stage of vi-sion, leading to the full extraction of the 3-Dinformation and to its exploitation for the de-scription and understanding of the scene.

high-loss resonator a resonator having ahigh value of loss (usually diffraction loss)per round trip; unstable resonator.

high-pass filter a filter that has a trans-fer function, or frequency response, whosevalues are small for frequencies lower thansome intermediate frequency. A filter whoseimpulse response is a high-pass signal.

high-pass signal a signal whose Fouriertransform, or power spectral density, is asmall value for all frequency components thatare less than some intermediate frequencyvalue. As a formal definition, ifX(ω) isthe Fourier transform of the signalx(t), thenX(ω) = 0 for |ω| < B, for someB > 0.Compare withband-pass signal.

high-performance metal oxide semicon-ductor (HMOS) technology a variationof the MOS process that scales down devicesto achieve higher speed and thus lower powerconsumption in integrated circuits.

high-rate quantization theory a theorydeveloped by Bennet, Schutzenberger, Zadorand others. The theory analyzes and predictsthe performance of (infinitely) fine quantiza-tion (high-rate quantization).

high-resistance grounded system anelectrical distribution system in which theneutral is intentionally grounded througha high resistance. The high-resistancegrounded wye system is an alternative tosolidly grounded and ungrounded systems.High-resistance grounding will limit groundfault current to a few amperes, thus remov-ing the potential for arcing damage inherentin solidly grounded systems.

high-speed carry in an arithmetic logicunit, a carry signal that is generated sepa-rately from the generation of the result and istherefore faster; a lookahead carry.

high-speed digital subscriber line (HDSL)a digital subscriber line (DSL) in which twotwisted-pairs provide a rate of 1.544 Mbps(T1 rate) in both directions (full-duplex).Each twisted-pair provides a full-duplex con-nection of 784 kbps. (This is larger than1.544/2 Mbps due to the duplication of a syn-chronization bit).

high-speed metal oxide semiconductor(HMOS) technology Seehigh-perfor-mance metal oxide semiconductor technology.

high-stability diode gun Plumbicon adiode gun Plumbicon tube with electrostaticfocus and magnetic deflection, which uses ahigh-stability electrode structure evaporatedand bonded to the tube envelope.

high-voltage DC (HVDC) transmissiontransmission of electric power (at typically500–1500 kV) using DC rather than AC. Thiscan be desirable for several reasons:

1. For economic reasons when a largeamount of power is to be transmitted overa long distance, i.e., 300–400 miles, or viaunderwater cables;

2. For the connection of asynchronous ACsystems; and

3. Improved transient stability and dy-namic damping of the electrical system os-cillations.

higher-order mode mode that spatiallyvaries faster than the fundamental in a cavity,or waveguide.

higher-order unit a neural unit whoseinput connections provide not the usualweighted linear sum of the input variables butrather a weighted polynomial sum of thosevariables.


highway Seebus.

Hilbert space an inner product space withthe additional property that is a completemetric space. An inner product space is alinear space on which an inner product is de-fined, while completeness means that thereare no “missing” vectors arbitrarily close tobut not included in the space.

Hilbert transform a transform that re-lates the real and imaginary parts of a com-plex quantity, such as the index of refraction.For f , a function of one real variable, thetransformH(f ) defined as follows:

H(f )(x) = limε→0

1

π

∫|t |≥ε

f (x − t)

tdt.

An alternate formula, which coincides withthe previous one, except possibly on a zero-measure set of points of discontinuity off ,is:

H(f )(x) = limε→0

1

π

∫ +∞

−∞f (x−t) t

t2 + ε2dt.

When f is square-integrable, the FouriertransformsF(f ) and F(H(f )) of f andH(f ) are related by:

F(H(f ))(u) = −i sign(u)F(f )(u)almost everywhere; thus for positive frequen-cies, the Fourier phase is shifted by−π/2;in particular,f andH(f ) have the sameL2

norm. SeeFourier amplitude, Fourier phase,Fourier transform.

histogram (1) a plot of the frequency ofoccurrence of the gray levels in an image.

(2) the frequency distribution of a set ofnumbers. In image processing, the distribu-tion of the gray levels in an image, typicallyplotted as the number or percentage of pix-els at a certain gray level vs. the gray levels.If the ordinate is the ratio of the number ofpixels at a gray level to the total number ofpixels, the histogram is an approximation ofthe probability density function of the graylevels.See alsoprobability density function.

histogram equalization a technique forcomputing an image gray level transforma-tion to redistribute the pixel intensity valuesand “flatten” the pixel intensity histogram.Histogram equalization can be used to en-hance the contrast of the image. Also calledhistogram leveling, histogram flattening.Seealsocontrast enhancement, histogram.

histogram modeling (1) making the his-togram of an image have another shape, forexample, the flat line produced by histogramequalization. This procedure is usually notautomatic, i.e., the shape is usually specifiedby a person. Also called histogram modifi-cation or histogram specification.See alsohistogram equalization.

(2) the fitting of a function to a histogramin order to obtain an analytical expressionthat approximates the histogram.

histogram sliding the addition of a con-stant to all pixels in an image to brighten ordarken the image. A positive constant makesthe histogram slide (translate) to the right; anegative constant makes it slide to the left.See alsohistogram.

histogram stretching expansion or con-traction of the gray-level histogram of an im-age in order to enhance contrast. Usually per-formed by multiplying or dividing all pixelsby some constant value.See alsocontrastenhancement, histogram.

hit the notion of searching for an addressin a level of the memory hierarchy and findingthe address mapped (and the data present) inthat level of the hierarchy. Often applies tocache memory hierarchies.

hit rate the percentage of references toa cache that find the data word requested inthe cache. Also the probability of finding thedata and given by

number of times required word found in cachetotal number references

Also known as hit ratio.

hit ratio Seehit rate.


hit–miss ratio in cache memory, the ratioof memory access requests that are success-fully fulfilled within the cache to those thatrequire access to standard memory or to anauxiliary cache.

hit–miss transform a class of transformsin image processing that locate objects largerthan some specific size,S1, and smaller thansome sizeS2.

HL7 a data communications protocol forinterfacing components of a hospital infor-mation system.

HMIC Seehybrid microwave integratedcircuit.

HMM Seehidden Markov model.

HMOS See high-performance metaloxide semiconductor technology.

hold signal an input signal to a processorindicating that a DMA device is requestingthe use of the bus. Also called bus request.

hold time the time required for the datainput(s) to be held stable prior to (or after)the control input changes to latch, capture, orstore the value indicated by the data inputs.

hold-up time the time duration that theoutput voltage is maintained, within speci-fication, during an interruption of the inputpower. Typical hold-up time is one to threeline cycles.

hole fictitious positive charge representingthe motion of electrons in the valence band ofa semiconductor; the number of holes equalsthe number of unoccupied quantum states inthe valence band.

hole burning localized reduction in gainof a laser amplifier due to saturation by anintense signal; localization may be in spaceor frequency.

hole mobility Seespatial hole burning,spectral hole burning.

Hollerith card a punched card used fordata storage. Developed by Herman Hol-lerith for use in the 1890 U.S. census, thesecards were widely used for early computerdata storage. Now obsolete.

Hollerith code an encoding of datawhere the data is represented by a pattern ofpunched holes in a card.

Hollerith, Herman (1860–1929) Born:Buffalo, New York, U.S.A.

Hollerith is best known for his develop-ment of a tabulating machine using punchedcards for entering data. This machine wasused to tabulate the 1890 U.S. census. Thesuccess of this machine spurred a number ofother countries to adopt this system. Thecompany Hollerith formed to develop andmarket his ideas became one of several com-panies that merged to form IBM, Interna-tional Business Machines. Hollerith’s initialinterest in automatic data processing prob-ably stems from his work as an assistant toseveral census preparers at Columbia Univer-sity. The success of Hollerith’s work giveshim claim to the title of the father of auto-mated data processing.

hollow beam an electron beam with hol-low core.

hollow cathode a negative electrodeformed using a section of hollow tube; usedin low pressure discharges.

hologram medium that, when illuminatedoptically, provides a three-dimensional im-age of stored information, sometimes calledholograph.

holograph Seehologram.

holographic data storage a techniqueused to store multiple images, or 2-D ar-rays of digital information, as multiplexed


holograms in an optically sensitive mate-rial. Multiplexing techniques include an-gle, frequency, and spatial position. Angleand frequency multiplexing techniques arebased on Bragg selectivity where angle mul-tiplexing favors transmission geometries andfrequency multiplexing favors reflection ge-ometries.

holographic interconnect free-space in-terconnect that uses holograms to control op-tical paths from sources to detectors. Thehologram provides the interconnection byforming the intentionally distorted image ofthe source array on the detector array. Ahologram can be viewed as a combination ofgratings. Each grating directs a light beampassing through it to a new direction. Thehologram can be displayed by a spatial lightmodulator to achieve dynamic reconfigura-tion.

holography the science of making andreading holograms.

homodyne detection method of measur-ing the frequency content of an optical signalin which that signal is combined with itselfby means of a square-law detector; simplerbut sometimes less informative and sensitiverepresentation of the frequency content of asignal than heterodyne detection.

homogeneous having all nodes or ma-chines in a multiprocessor or multicomputerbe of the same type or class.

homogeneous broadening spectral broad-ening of a transition in a laser medium dueto irreversible dephasing processes like spon-taneous emission, collisions, photon interac-tions, or spin exchange.

homogeneous coordinate representationthe representation of anN–component po-

sition vector by an(N + 1)–component vec-tor is called homogeneous coordinate rep-resentation. For example, a vectorp =[px, py, pz

]T is represented by an aug-

mented vector[wpx,wpy,wpz,w

]in the

homogeneous coordinate representation. Inthe homogeneous transformation matrix it isassumed thatw = 1.

homogeneous coordinates a 2-D point(x, y)with a third coordinate point(x, y,W),whereW is a constant typically 1. Pointsrepresented in homogeneous coordinates al-low for translation, rotation and scaling to beapplied as multiplications of transformationmatrices and row vectors. Without homo-geneous coordinates, translations are treatedas additions while scaling and rotation aretreated as multiplications.

homogeneous linear estimator an esti-mator that is a homogeneous linear functionof the data.

homogeneous medium optical mediumin which none of the properties (e.g., gain,index of refraction, etc.) are functions of po-sition.

homogeneous solution a system of lin-ear constant-coefficient differential equa-tions has a complete solution that consistsof the sum of a particular solution and a ho-mogeneous solution. The homogeneous so-lution satisfies the original differential equa-tion with the input set to zero. Analogousdefinitions exist for difference equations.

homogeneous transformation matrix a4 × 4 matrix that maps a position vector ex-pressed in homogeneous coordinates fromone coordinate system to another coordi-nate system (reference coordinate frame). Ahomogeneous transformation matrix has theform

T =[R3×3 p3×101×3 1

]=

nx ox ax pxny oy ay pynz oz az pz0 0 0 1

whereR3×3 represents the rotation matrix,which consists of three unit vectorsn (nor-mal), o (orientation), a (approach). The


orientation matrix is an orthonormal matrix.p3×1 represents the position vector of the ori-gin of the rotated coordinate system with re-spect to the reference system. 01×3 is 1× 3zero vector.

homojunction a junction between regionsof the same bulk material that differs in theconcentration of dopants. The typical exam-ple is the n-p diode.

homomorphic filter an image enhance-ment technique based upon the illumination-reflectance model. The homomorphic fil-ter assumes the image functionf (x, y) isthe combination of two functionsi(x, y) andr(x, y). By taking the natural log of the im-ages the components are separated and canbe processed separately; the reflectance of animage usually contains high frequency com-ponents while the illumination componenttends to vary slowly (the low frequencies);thus, the filter applied to the logarithm im-age should affect the low and high frequencycomponents differently. The most commonuse for the filter is to enhance images by com-pressing the brightness range while increas-ing contrast; the filter applied in this case isa circularly symmetric high pass filter withthe stop band magnitude< 1 and pass bandmagnitude> 1.

homomorphic signal processing the pro-cessing of signals from a nonadditive modelunder a tranformation which renders themodel additive. For example, a logarithmictransformation may be used to transform amultiplicative noise model into an additivenoise model.

homopolar generator an electromag-netic generator in which the magnetic fluxpasses in the same direction from one mag-netic member to the other over the whole ofa single air gap area. Such generators havebeen built to supply very large pulsed cur-rents.

homopolar machine See homopolargenerator.

homopolar magnetic bearing a magneticbearing in which the rotating member alwaysexperiences a magnetic field of the same po-larity.

homothetic a copy of a signal that retainsthe original shape, but is scaled in size. Fora signalx(t), the homotheticxλ(t) = x(λt)

retains the shape ofx(t) but is scaled alongthe t axis as a function of the parameterλ.Typically used in association with structuringelements in mathematical morphology.

homotopy method a technique for solv-ing nonlinear algebraic equationsF(x) = 0based on higher-dimensional function em-bedding and curve tracing. The idea is toconstruct a parameterized function such thatat one parameter value, sayλ = λ0, the sys-tem of equations is easy to solve or has oneor more known solutions, and at another pa-rameter value, sayλ = λf , the system ofequations is identical to that of the systemof interest,F(x) = 0. A homotopy methodmay then be interpreted as geometric curvefollowing through solution space from theknown solutions of the “easy” problem to theunknown solutions ofF(x) = 0.

Hopfield memory SeeHopfield model.

Hopfield model a neural algorithm capa-ble of recognizing an incomplete input. Alsoknown as Hopfield memory.

Hopfield suggested that an incomplete in-put can be recognized in an iterative process,in which the input is gradually recognizedin every cycle of the iteration. The iterationis completed when the input finally matcheswith a stored memory. The Hopfield modelis a sort of associative memory. A holo-gram can also be directly used as associativememory. The main difference of the non-neural holographic associative memory andthe Hopfield model is as follows. The di-rect holographic associative memory is one


step and its signal-to-noise ratio depends onthe incompleteness of the input that cannot beimproved. The Hopfield model is an iterativeprocess involving a nonlinear operation, suchas thresholding in which the signal-to-noiseratio of the input can be improved graduallyduring the iterative process. A large num-ber of optical systems have been proposedto implement the Hopfield model, includingthe first optical neural networks. Those op-tical implementations are primarily based onoptical matrix-vector or tensor-matrix multi-plication.

Hopfield network a recurrent, associa-tive neural network withn processing ele-ments. Each of the processing elements re-ceives inputs from all the others. The inputthat a processing element receives from it-self is ignored. All of the processing ele-ments output signals are bipolar. The net-work has an energy function associated withit; whenever a processing element changesstate, this energy function always decreases.Starting at some initial position, the sys-tem’s state vector simply moves downhill onthe network’s energy surface until it reachesa local minimum of the energy function.This convergence process is guaranteed to becompleted in a fixed number of steps.Seealso continuous Hopfield network, discreteHopfield network.

Hopper, Grace Murray (1906–1992)Born: New York, New York, U.S.A.

Hopper is best known as the author ofthe first compiler. Hopper began her careeras a mathematics professor at Vassar Col-lege. During WW II she volunteered for ser-vice and was assigned to work at Harvardwith Howard Aiken on the Mark I computer.She later joined J. Presper Eckert and JohnMauchly working on the UNIVAC computer.It was at this time she wrote the first compiler.Her compiler and her views on computer pro-gramming significantly influenced the devel-opment of the first “English-like” businesscomputer language, COBOL.

hopping sequence Seefrequency hopping.

horizontal (H) in television signals,Hmay refer to any of the following: the hori-zontal period or rate, horizontal line of videoinformation, or horizontal sync pulse.

horizontal microinstruction a microin-struction made up of all the possible micro-commands available in a given CPU. In prac-tice, some encoding is provided to reduce thelength of the instruction.

horizontal polarization a term used toidentify the position of the electric field vec-tor of a linearly polarized antenna or propa-gating EM wave relative to a local reference,usually the ground or horizon. A horizon-tally polarized EM wave is one with its elec-tric field vector aligned parallel to the localhorizontal.

horizontal rate the rate at which a lineof video is drawn; for NTSC television, thehorizontal frequency is 15,734.264 Hz for arate of 63.5µs/line (the same as line rate).

horizontal scanning in radar systems, therotation of the antenna parallel to the horizon.In video display systems, the process of ex-amining or displaying an image with multiplehorizontal lines.

horizontal sync pulse a blacker-than-black signal level contained within the hor-izontal blanking interval of the compositeNTSC signal. The horizontal sync pulseduration is 7.5± 0.5 percent of the hori-zontal line time and an amplitude of 25%of the peak-to-peak video signal. The hor-izontal sync pulse frequency is 15750 Hzmonochrome and 15734 Hz for color NTSCtelevision formats. The NTSC modified themonochrome horizontal frequency slightly tointerleave the color subcarrier and to providean integer relationship of 286 between thehorizontal rate and the 4.5 MHz sound car-rier.


During the equalizing pulse interval, thehorizontal sync pulse rate is doubled and thepulse width is halved. Accurate horizontalsynchronizing the video display is preservedthroughout the NTSC frame by measuringthe horizontal line time between the blankinglevel to the sync level transitions.

horn antenna an aperture antenna formedby a waveguide that has been flared out onone end.

horn gap a V-shaped spark gap whichprovides a method of extinguishing a power-follow arc by allowing the arc to climb thesides of the V until it is too long to be main-tained.

horsepower-rated switch a manually op-erated switching device designed for motorcircuit applications. It is designed to inter-rupt the rated overload current at rated volt-age of a motor with a horsepower rating thatis less than or equal to the horsepower ratingof the switch.

hose Seeline hose.

host a computer that is the one responsi-ble for performing a certain computation orfunction.

hot an energized conductor.

hot electron an electron in the conductionband of a semiconductor having a superther-mal kinetic energy.

hot electron bolometer a superconduct-ing resistor structure that uses rapid heatingand cooling of electrons by an RF field toproduce a time varying resistance, useful asa mixer element at submillimeter and THzfrequencies.

hot electron transistor a transistor, usu-ally fabricated from heterostructure materi-als, that uses electron transport over an en-

ergy step to obtain high electron velocitiesand hopefully high speed operation.

hot line work work performed on ener-gized electric power lines.Seeglove, hotstick, bare-hand.

hot reserve the state of an idle thermalgenerating plant whose boilers and turbinesare hot and can thus be quickly brought intoservice.

hot restart reassumption, without loss, ofall operations in the system from the point ofdetected fault.

hot standby a technique for achievingfault tolerance that requires having a backupcomputer running in parallel with the primarycomputer so that the backup may take over,if the primary should fail.

hot stick an insulated pole used by lineworkers to make connections to and other-wise manipulare energized overhead conduc-tors.

hot tap a clamp, applied with a hot stickthat connects a branch circuit to an exist-ing conductor and typically applied while thesystem is energized.

hot wire an energized conductor, par-ticularly as opposed to a neutral or groundwire, both of which are typically maintainedat ground potential

hot-carrier diode SeeSchottky barrierdiode.

Hotelling transform a transformationwhose basis vectors are the eigenvectors ofthe covariance matrix of the random vectors.Seealso Karhunen–Loeve transform.

Hough transform (1) template-matchingmethod for detecting straight lines and curvesin gray level images: given a family of


curves, the output of the transform is the setof such curves that are present in the image.

(2) a transform that transforms image fea-tures and presents them in a suitable form asvotes in a parameter space, which may thenbe analyzed to locate peaks and thereby inferthe presence of desired arrangements of fea-tures in the original image space: typically,Hough transforms are used to locate specifictypes of object or shape in the original im-age. Hough transform detection schemes areespecially robust.

Householder transformation a matrixQthat maps each vector to its reflection througha defined hyperplane; specifically,

Q = I − 2uuT

uT u

reflects through the plane having normal vec-tor u.

HPA acronym for high-power amplifier.

HPO Seehigh phase order.

HRC Seehigh rupturing capacity.

HSI/HSV System a system whereby coloris represented by hue, saturation, and inten-sity or value, hence the acronym. This systemtends to be more intuitive for users than theother two since it is similar to an artist’s tint,shade and tone. The HSV hexcone is createdin a cylindrical coordinate system, where hueis the angle around the vertical axis, value isthe height along the vertical axis and satu-ration is the perpendicular distance from thevertical axis.See alsosaturation.

HTC Seehydro-thermal coordination.

HTGR acronym for high temperature gas-cooled reactor.

hue one of the characteristics that distin-guishes one color from another. Hue definescolor on the basis of its position in the spec-trum (red, blue, green, yellow, etc.). Hue is

one of the three characteristics of televisioncolor. Hue is often referred to as tint. InNTSC and PAL video signals, the hue infor-mation at any particular point in the pictureis conveyed by the corresponding instanta-neous phase of the active video subcarrier.

hue saturation intensity (HSI) a colormodel based on the specification of hue (H),saturation (S), and intensity (I). A useful andconvenient property of this model is the factthat intensity is separate from the color com-ponents. The hue and saturation components(together referred to as chromaticity) relateclosely to color perception in humans, how-ever the HSI intensities are linear and do notcorrespond to those observed by the eye. Theconversion from RGB to HSI goes as

H = arccos(0.5((R −G)

+ (R − B))/√((R −G)2

+ ((G− B) ∗ (R − B))

S = 1 − (3/(R +G+ B))

∗ (min{R,G,B})I = (R +G+ B)/3

See alsoRGB, color space.

Huffman coding (1) a method for loss-less source coding that constructs the code ofminimum average length for a random vari-able. The method is optimal in the sense thatno other method can give a higher compres-sion rate. It is capable of achieving the boundon compression given by the source codingtheorem. It is due to D. A. Huffman (1952).

(2) a variable length coding scheme whosecodewords are generated from the probabil-ity distribution of the source. Decoding aHuffman codeword corresponds to travers-ing an unbalanced binary tree according tothe value (0 or 1) of each bit in the word; theleaves of the tree are the source symbols, withthe most probable ones being the shortest dis-tance from the root of the tree. Huffman cod-ing achieves an average code rate equal to thesource entropy if and only if all the probabil-ities are negative powers of 2. In general


it achieves less compression than arithmeticcoding but is easier to implement.

Huffman encoding SeeHuffman coding.

hum bars horizontal black and white barsthat extend over the entire TV picture andusually drift slowly through it. Hum bars arecaused by an interfering power line frequencyor one of its harmonics.

human visual system (HVS) the collec-tion of mechanisms in humans which processand interpret the visual world. These mech-anisms include the eye, the retina, the opticnerve, the visual cortex, and other parts of thebrain.

hunting a mechanical oscillation in thespeed of a synchronous machine due tochanges in the load. Damper windings areused to reduce the hunting by providing atorque that opposes the change in speed.

Hurter–Driffield (H-D) curve the stan-dard form of the H-D or contrast curve isa plot of the relative thickness of resist re-maining after exposure and development of alarge clear area as a function of log-exposureenergy. The theoretical H-D curve is a plotof log-development rate versus log-exposureenergy. (Hurter and Driffield are the two sci-entists who first used a related curve in 1890.)

Hurwitz matrix a square matrix with realelements and whose all eigenvalues have neg-ative real parts.

Hurwitz polynomial a polynomialp(s) =ans

n+an−1sn−1+an−2s

n−2+· · ·+a0, whereall of theai ’s are either real or complex num-bers, is said to be Hurwitz if all of its rootshave negative real part.

Huygen’s principle a principle statingthat each point of a wave front can be consid-ered to be a source of a secondary sphericalwave that combines with the other secondaryspherical waves to form a new wave front.

Huygen’s principle is often used as an ap-proximate technique for solving diffractionand antenna radiation problems.See alsoequivalence theorem, equivalent source.

HVC color coding technique whereby theHVS contrast sensitivity functions for differ-ent wavelengths are taken into account forcompressing the color planes independentlywith different sets of parameters.

HVDC transmission Seehigh-voltageDC transmission.

hybrid bipolar MOS thyristor (HBMT)like its transistor cousin, it is an advanced de-vice designed to take advantage of the uniquecharacteristics of both bipolar devices (highcurrent capability with low forward volt-age drop) and MOS device (high switchingspeeds with low gate drive requirements).

hybrid circuit a circuit based on at leasttwo different technologies. For instance, acircuit built by using solid state circuits andtubes.

hybrid computer a computer based onat least two different technologies. For in-stance, a computer presenting both digitaland analog circuits and signals.

hybrid coupler generally a four port cir-cuit, which has an electrical response suchthat a signal applied to port one is dividedequally in amplitude between ports three andfour. The phase of the signals at ports threeand four are either 0, 90, or 180 degrees apart,dependent on the type of hybrid coupler. Porttwo is completely isolated, and no signal ap-pears at this port.

hybrid magnetic bearing an active mag-netic bearing that also incorporates perma-nent magnets, thus reducing energy con-sumption and improving performance.

hybrid microwave integrated circuit(HMIC) a planar assembly that com-


The hybrid position/force controller.

bines different circuit functions formed bystrip or microstrip transmission lines printedonto a dielectric substrate and incorporatingdiscrete semiconductor solid state devicesand passive distributed or lumped circuit el-ements, interconnected with wire bonds.

hybrid mode a solution to Maxwell’sequations for a particular structure that is nei-ther TE nor TM.

hybrid numerical method a method thatmakes use of two or more different tech-niques (e.g., mode matching and finite dif-ferences) in order to efficiently solve a givenboundary value problem. See alsofinitedifferences, mode matching.

hybrid parameters circuit parametersused especially in transistor models; whenreferring to a two-port, they relate the volt-age at port 1 with its current, and the voltageat port 2 with the current of port 2 with itsvoltage, and the current at port 1 accordingto the equation

V1 = h11I1 + h12V2

I2 = h21I1 + h22V2

hybrid position/force controller a gen-eral hybrid position/force controller is pre-sented in the figure. The hybrid controllerconsists of two subcontrollers. One is a posi-tion controller and the second is a force con-troller. Trajectory generator inputs a Carte-sian path in terms of position, velocity, andacceleration.Fd is a desired external forceassociated with the task in Cartesian space.Matrix S is a selection matrix with ones andzeros on the diagonal. Where a one is present

in S, a zero is present inS′, and position con-trol is in effect. It is assumed that joint posi-tions, velocities, and accelerations are mea-sured. Next they are transformed via kine-matical computations to Cartesian positions,velocities, and accelerations. Desired andactual parameters are compared in positioncontroller. Desired and measured forces arecompared in force controller. Finally, manip-ulator is controlled at the joint force and/ortorque level.

hybrid redundancy a technique combin-ing active and passive redundancy. Passiveredundancy prevents generation of error re-sults, and active redundancy improves faulttolerance capabilities by replacing faultyhardware with spare elements.

hybrid stepper motor a stepper motorthat combines the rotor design characteristicsof variable-reluctance and permanent magnetstepper motors. Hybrid stepper motor rotorsconsist of an axially magnetized cylindricalpermanent magnet capped on each pole bytoothed, soft iron caps. Teeth on the caps aredisplaced with respect to each other to pro-vide stepping control. Hybrid stepper motorscombine the higher torque capability of per-manent magnet motors with the higher stepresolution of variable-reluctance motors.

hydroelectric generator large, three-phase synchronous alternator powered by awater-driven turbine.See alsogenerator.

hydro-thermal coordination the practiceof manipulating water levels in the reservoirsof a power system’s hydroelectric plants withrespect to the generation levels of the thermalplants in the system with the objective of min-imizing generation costs, as well as satisfyingwaterway considerations such as water con-servation, flood control capacity, recreation,and environmental requirements.

hydrophone receiving sensors that con-vert sound energy into electrical or opti-


cal energy (analogous to underwater micro-phones).

hydropower conversion of potential en-ergy of water into electricity using generatorscoupled to impulse or reaction water turbines.

hyperpolarizability a measure of thenonlinear optical response of an atom ormolecule. The hyperpolarizability of ordernγ (n) is defined by the relationp(n) =Cγ (n)En, where E is the electric fieldstrength of the applied laser field andp(n) isthenth order contribution to the dipole mo-ment per molecule.C is a coefficient of orderunity whose definition is different among dif-ferent workers.

hysteresis (1) the phenomenon that themagnetic state of a substance is dependentupon its magnetic history, so that its magne-tization for an increasing magnetizing forcediffers from that for a decreasing magnetiz-ing force.

(2) the characteristic of magnetic materi-als that causes the trajectory of the flux den-sity vs. field intensity curve as the intensity isincreased to be different from that when theintensity is decreased, giving rise to a loss,which is proportional to the area enclosed bythe two trajectories.

hysteresis brake a braking device utiliz-ing hysteresis to provide a constant brakingtorque irrespective of slip speed.

hysteresis control a time-optimal feed-back control method in which the controlvariable reaches a reference value in theshortest possible time and then stays withina prescribed hysteresis band around the setpoint through manipulation of the systemstate between two configurations. The ac-tual variable is compared with the referencevalue, and if the error exceeds the hysteresis

band, then the control input is changed suchthat the control variable is forced to decrease.On the other hand, if the actual variable fallsbelow the hysteresis band then the controlinput is changed such that the control variableincreases in magnitude.

hysteresis curve a graph describing therelationship between the magnetic flux den-sity and the magnetic field intensity in a (usu-ally ferromagnetic) material.

hysteresis drive Seehysteresis torquecoupling.

hysteresis loss the energy loss due to hys-teresis in a magnetic material subjected to avarying magnetic field.

hysteresis motor any of a variety ofsingle-phase AC motors that use the hystere-sis properties of hard magnetic materials todevelop torque. Stator windings of a hystere-sis motor can be of any design that producesa rotating flux within the machine. Motion ofthe rotating flux over the rotor magnetizes thehard magnetic material on the rotor; however,the hysteresis characteristics of the materialcause the alignment of magnet flux to lagthe rotating stator flux. This misalignmentproduces rotor torque. Because of the na-ture of the torque production, hysteresis mo-tors operate at synchronous speed and havea constant torque characteristic, which per-mits them to synchronize any load that theycan accelerate.

hysteresis torque coupling a magneticdrive in which the magnetizing stator magnetdrives a rotor of hysteresis material throughthe complete hysteresis cycle once per rota-tion, resulting in a constant torque character-istic irrespective of relative speed.

Hz hertz.Seefrequency.


II and Q the I and Q signals are used invideo transmission to generate a chroma pha-sor or vector. By varying the amplitude of theI and Q signals, all colors in the visible spec-trum can be generated and transmitted. The Iand Q signals are derived from the RGB sig-nals where I = .6R− 0.28G− 0.32B, Q=.21R− .52G+ .31B.

IDC DC current in amperes.

I INp peak input current in amperes.

IOUTp peak output current in amperes.

IDS FET channel current.

I& P Seeinterconnectionsandpackaging.

I-V characteristics the charts describingthe current through a diode as a function ofits voltage bias.

I/O input/output. Operations or devicesthat provide data to or accept data from acomputer.

I/O bandwidth the data transfer rate intoor out of a computer system. Measured inbits or bytes per second. The rate dependson the medium used to transfer the data aswell as the architecture of the system. Insome instances the bandwidth average rateis given and in others the maximum rate isgiven. See alsoI/O throughput.

I/O buffer a temporary storage area whereinput and output are held. Having I/O buffersfrees a processor to perform other tasks whilethe I/O is being done. Data transfer ratesof the processor and an I/O device are, ingeneral, different. The I/O buffer makes this

difference transparent to both ends.

I/O bus a data path connecting input oroutput devices to the computer.

I/O card a device used to connect a pe-ripheral device to the main computer; some-times called an I/O controller, I/O interface,or peripheral controller.

I/O channel Seeinput/output channel,subchannel I/O.

I/O command one of the elementary typesof commands that a computer can execute.I/O commands are typically used to initi-ate I/O operations, sense completion of com-mands, and transfer data. Not all comput-ers have I/O commands.See alsomemorymapped I/O.

I/O controller (1) a device used to connecta peripheral device to the main computer;sometimes called an I/O card.

(2) the software subroutine used to com-municate with an I/O device.See alsoI/Oroutine.

I/O device a physical mechanism attachedto a computer that can be used to store out-put from the computer, provide input to thecomputer, or do both.

I/O interface a device used to connecta peripheral device to the main computer;sometimes called an I/O card or I/O con-troller.

I/O interrupt a signal sent from an inputor output device to a processor that the statusof the I/O device has changed. An interruptusually causes the computer to transfer pro-gram control to the software subroutine thatis responsible for controlling the device.Seealso interrupt-driven I/O.

I/O port (1) a register designed specifi-cally for data input–output purposes.


(2) the place from which input and out-put occurs in a computer. Examples includeprinter port, serial port, and SCSI port.

I/O processor an input/output processorwith a unique instruction set, dedicated toperforming I/O operations exclusively, thusalleviating the burden off the CPU. It usuallyhas a separate local bus for I/O operationsdata traffic, thus permitting the CPU to ac-cess memory on the main system bus with-out interruption. For example, the Intel 8086CPU has an 8089 I/O processor associatedwith it. Both can operate simultaneously, inparallel.

I/O register a special storage locationused specifically for communicating with in-put/output devices.

I/O routine a function responsible forhandling I/O and transferring data betweenthe memory and an I/O device.See alsoI/Ocontroller.

I/O system the entire set of input/outputconstructs, including the I/O devices, devicedrivers, and the I/O bus.

I/O throughput the rate of data trans-fer between a computer system and I/O de-vices. Mainly determined by the speed of theI/O bus or channel. Throughput is typicallymeasured in bits/second or bytes/second. Insome instances, the throughput average rateis given and in others the maximum rate isgiven. See alsoI/O bandwidth.

I/O trunk SeeI/O bus.

I/O unit the equipment and controls nec-essary for a computer to interact with a hu-man operator or to access mass storage de-vices or to communicate with other computersystems over communication facilities.

I-line a line of the mercury spectrum cor-responding to a wavelength of about 365 nm.

IA Seeindex assignment.

IAEA International Atomic Energy Agency,an organization which monitors nuclear ma-terials and energy.

IC Seeintegrated circuit.

ICE Seein-circuit emulator.

ideal filter (1) a system that completelyrejects sinusoidal inputs of the formx(t)4 =A cosωt , −∞ < t < ∞, for ω within a cer-tain frequency range, and does not attenuatesinusoidal inputs whose frequencies are out-side this range. There are four basic types ofideal filters: low-pass, high-pass, band-pass,and bandstop.

(2) a filter that has zero gain in the stop-band, a constant, usually unity, gain in thepassband, and a zero transition band width. A1-D ideal lowpass filter has unity gain for fre-quencies less than or equal to the cutoff fre-quency, and zero gain for frequencies greaterthan the cutoff frequency. An ideal filter isunrealizable in practice.

ideal operating amplifier an op amp hav-ing infinite gain from input to output, withinfinite input resistance and zero output re-sistance and insensitive to the frequency ofthe signal. An ideal op amp is useful in first-order analysis of circuits.

ideal transformer a transformer with zerowinding resistance and a lossless, infinitepermeability core resulting in a transformerefficiency of 100 percent. Infinite permeabil-ity would result in zero exciting current andno leakage flux. For an ideal transformer, theratio of the voltages on the primary and sec-ondary sides would be exactly the same asthe ratio of turns in the windings, while theratio of currents would be the inverse of theturns ratio.

ideality factor the factor determining thedeviation from the ideal diode characteristicm = 1. At small and large currentsm ≈ 2.


idempotent an operator is idempotent ifapplying it twice gives the same result as ap-plying it only once; mathematically speak-ing, if ψ is the operator andX the ob-ject to which it is applied, this means thatψ(ψ(X)) = ψ(X).

idler wave the additional, often unwanted,output wave produced by an optical paramet-ric oscillator designed to generate a signalwave.

IEC SeeInternational Electro-technicalCommission.

IEEE See Institute of Electrical andElectronics Engineers.

IEEE Color Books a series of sevenbooks related to industrial and commer-cial power systems containing recommendedIEEE/ANSI standards and practices. Thebooks are color-coded as to subject as fol-lows:

Gray: Power systems in commercialbuildings

Green: Grounding industrial and com-mercial power systems

Brown: Power system analysisGold: Design of reliable industrial and

commercial power systemsOrange: Emergency and standby power

for industrial and commercial power systemsRed: Electric distribution practice in in-

dustrial plantsBuff: Protection and coordination of in-

dustrial and commercial power systems.

IEEE float encoding SeeNaN.

IEEE representation an encoding con-vention for the representation of floating-point values within computing systems.

IEEE Standard 754 the specification thatdefines a standard set of formats for the rep-resentation of floating-point numbers.

IF Seeintermediate frequency.

IF amplifier an amplifier having highand controllable gain with a sharp band-passcharacteristic centered around the intermedi-ate frequency (IF).

IGBT Seeinsulated gate bipolar junctiontransistor.

ignitron a high-voltage mercury switch.The device is found in modulators used todump the capacitor bank voltage in the eventof a PA crowbar. An ignitron passes electricalcurrent to a pool of liquid mercury at groundpotential.

II-VI semiconductor binary semicon-ductor made from elements in the periodictable two columns to the “left” and “right” ofsilicon, e.g., CdTe, ZnS, etc.

IID See independent and identicallydistributed.

III-V semiconductor a binary semicon-ductor made from elements from columns IIIand V in the periodic table.

illumination (1) in vision science, theamount of light per unit area delivered to asurface and accounting for the spectral sen-sitivity of the human eye.

(2) informally, the amount of light fallingon an object. Also called illuminance.Seealso luminance.

(3) the effect of a visible radiation fluxreceived on a given surface. Illumination ismeasured by the illuminance, which is the lu-minous flux received by surface unit, usuallyexpressed in lux. One lux equals 1 lumen/m2.

illumination system the light source andoptical system designed to illuminate themask for the purpose of forming an imageon the wafer.

ill-posed problem a problem whose so-lution may not exist, may not be unique, ormay depend discontinuously on the data. (Aproblem is well-posed if it can be shown that


its solution exists, is unique, and varies con-tinuously with perturbation of the data. Ifany of the three conditions does not hold, theproblem is ill-posed.)

image information stored digitally in acomputer, often representing a natural sceneor other physical phenomenon, and presentedas a picture. The brightness at each spot inthe picture is represented by a number. Thenumbers may be stored explicitly as pixels orimplicitly in the form of equations.See alsobitmapped image, image acquisition, imageprocessing, imaging modalities, pixel, vectorimage.

image acquisition the conversion ofinformation into an image. Acquisitionis the first stage of an image processingsystem and involves converting the inputsignal into a more amenable form (suchas an electrical signal), and sampling andquantizing this signal to produce the pix-els in the image. Hardware, such aslenses, sensors and transducers are partic-ularly important in image acquisition.Seealsoanalog-to-digital converter, digitization,image, pixel, quantization, sample.

image analysis the extraction of infor-mation from an image by a computer. Theresults are usually numerical rather than pic-torial and the information is often complex,typically including the recognition and inter-pretation of objects in or features of the im-age.

image classification the division by acomputer of objects in an image into classesor groups. Each class contains objects withsimilar features. Also, the division of theimages themselves into groups according tocontent.

image coding the usual reasons for codingmessages lie in the areas of data compres-sion and data security. However, the mainmotivation for recoding images is to achievethe high rates of data compression required

to cope with the storage and transmission ofprodigious amounts of image data.Seestillimage codingandvideo codingfor methods.

image coding for videoconferencing thisis also called coding at primary rates forvideoconferencing. In North America andJapan, one of the most plentiful is the DSI rateof 1.544 Mb/s. In Europe, the correspondingrate is 2.044 Mb/s. The main difference isthat in videoconferencing the video camerais fixed and picture data is produced only ifthere are moving objects in the scene. Typ-ically, the amount of moving data is muchsmaller than the stationary area.

image compression representation of stilland moving images using fewer bits thanthe original representation.Seestill imagecodingandvideo codingfor methods.

image compression or coding the pro-cess of reducing the number of binary digitsor bits required to represent the image.

image correlation the calculation of thecorrelation function where the two signal in-puts are images.

image degradation blurring of an imagewavefront due to scattering as it passes aninhomogeneous medium, e.g., astronomicalimages due to atmospheric scattering.

image dissector a nonstorage photoemis-sion device used as a light detector in earlytelevision camera tubes. Electron streams,emitted by light striking a photocathodefilm, are focused on a plane containing asmall aperture into a photomultiplier struc-ture. Scanning is accomplished by magneti-cally deflecting the electron stream across theaperture.

image enhanced mixer mixer that op-erates on the principle of terminating theimage frequency component at an optimalimpedance level that improves the conversion


loss of the mixer. The optimal impedance ter-mination allows the image product to remixwith the LO frequency component such thatthe down-converted signal is in phase withthe IF component and thus increases its mag-nitude. The image spectral component is at afrequency that is either the sum or differenceof the LO and IF components.

image enhancement processing an imageto improve its appearance, or to make it moresuited to human or machine analysis. En-hancement is based not on a model of degra-dation (for which,Seerestoration) but onqualitative or subjective goals such as remov-ing noise and increasing sharpness. Contraststretching, edge emphasis, and smoothing areexamples of enhancement techniques.

image feature an attribute of a block ofimage pixels.

image frequency rejection ratio ratio ofoutput for the image frequency below thatfor the desired signal when both image anddesired signal levels at the input are the same.

image frequency response output gener-ated in a superheterodyne receiver due to anundesired signal with a frequency spaced bytwice the IF from desired frequency.

image fusion the merger of images takenwith different imaging modalities or with dif-ferent types of the same modality, for ex-ample, different spectral bands.See alsomultispectral image.

image guide a type of dielectric waveg-uide based on a dielectric rod waveguide witha conducting plate placed at the plane of sym-metry.

image hierarchy a way to organize theimage data in order of importance. Eachlevel corresponds to a reconstructed imageat a particular resolution or level of quality.Generally, they are classified into fixed andvariable resolution hierarchies.

image impedance the input impedanceinto a passive 2-port device when that deviceis terminated in an infinite chain of identi-cal devices connected to both ports such thateach port is connected to the same port (i.e.,port 1 connects to port 1 and port 2 connectsto port 2 throughout the infinite chain, giv-ing a perfect match). The impedance lookinginto port 1 isZI11, and into port two isZI22.For a uniform transmission line, the imageimpedances are equal to each other and alsoequal to the characteristic impedance of theline.

image processing the manipulation ofdigital images on a computer, usually doneautomatically (solely by the computer) butalso semi-automatically (the computer withthe aid of a person). The goal of the pro-cessing may include compressing the image,making it look better, extracting informationor measurements from it or getting rid ofdegradations. See alsoimage acquisition,image analysis, image classification, imagecompression, image enhancement, imagerecognition, image restoration.

image recognition the identification andinterpretation of objects in an image. Im-age recognition is a higher-level process thatrequires more intelligence than other partsof image processing such as image enhance-ment. Artificial intelligence, neural nets orfuzzy logic are often used.See alsoartificialintelligence, artificial neural network, fuzzylogic, classifier, image classification, imageunderstanding, pattern recognition.

image reconstruction (1) the process ofobtaining an image from nonimage data thatcharacterizes that image.

(2) the conversion of a digital image to acontinuous one suitable for display.See alsoanalog-to-digital converter, interpolation.

(3) in tomography, the reconstruction of aspatial image from its projections.See alsocomputed tomography, Radon transform.


image registration the spatial alignmentof a pair of images, which may involve cor-recting for differences in translation, rotation,scale, deformation or perspective. Especiallyimportant in binocular vision.

image rejection filter a filter usuallyplaced before a mixer to preselect the fre-quency, which when “mixed” with the localoscillator frequency, will down-convert to thedesired intermediate frequency. In the ab-sence of an image rejection filter, there aretwo RF frequencies, which when mixed withthe local oscillator frequencyflo, will down-convert to the intermediate frequencyfif .Specifically, these are given byflo+fif andflo−fif . For example, if the local oscillatorfrequency is 10 GHz and the local oscilla-tor frequency is 1 GHz, then signals at both11 GHz and 9 GHz will produce an interme-diate frequency signal of 1 GHz without animage rejection filter.

image resolution a measure of the amountof detail that can be present in an image. Spa-tial resolution refers to the amount a scene’sarea that one pixel represents, with a smallerarea being a higher resolution. Brightnessresolution refers to the amount of luminanceor intensity that each gray level represents,with a smaller amount being a higher resolu-tion.

image restoration (1) recovery of the am-plitude and/or the phase of an image wave-front, degraded on scattering, by either digi-tal or optical procedures, e.g., deconvolution,inverse scattering, or speckle interferometrytechniques.

(2) the removal of degradations in an im-age in order to recover the original image.Some typical image degradations and possi-ble causes are: blurring (intervening atmo-sphere or media, optical abberrations, mo-tion), random noise (photodetectors, elec-tronics, film grain), periodic noise (elec-tronics, vibrations) and geometric distor-tions (optical abberrations, angle of image

capture). See alsodegradation, motioncompensation, noise, noise suppression.

image reversal a chemical process bywhich a positive photoresist is made to be-have like a negative photoresist.

image segmentation the division of animage into distinct regions, which are laterclassified and recognized. The regions areusually separated by edges; pixels within aregion often have similar gray levels.Seealso edge detection, image classification,image recognition.

image smoothing the reduction of abruptchanges in the gray levels of an image. Usu-ally performed by replacing a pixel by theaverage of pixels in some region aroundit. Also called image averaging.See alsosmoothing.

image theory a solution technique for de-termining the fields radiated by sources inthe presence of material boundaries; typi-cally the boundary is between a dielectricand a perfect electric conductor. The equiva-lence theorem is used to generate an equiva-lent problem in a homogeneous medium withimage sources that provide the same bound-ary conditions as the original problem.

image transform the conversion ofthe image from the spatial domain to an-other domain, such as the Fourier do-main. It is often easier and more effec-tive to carry out image processing suchas image enhancement or image compres-sion in the transformed domain.See alsoblock transform, cosine transform, discreteFourier transform (DFT), discrete Hadamardtransform, discrete Hadamard transform,discrete cosine transform (DCT), discretesine transform (DST), discrete wavelettransform (DWT),distance transform,Fouriertransform,Haartransform,Hadamardtransform,Hankel transform,Hilbert transform,Hotellingtransform,Hough transform,Karhunen–Loevetransform (KLT), medial axis transform


(MAT), Radon transform, separable imagetransform, sine transform, slope transform,top hat transform, transform,Walsh transform,wavelet transform.

image understanding the interpretationby a computer of the contents of an image.The process seeks to emulate people’s abil-ity to intelligently extract information fromor make conclusions about the scene in animage. Also called image interpretation.

imaginary power Seereactive power.

imaging modalities the general phys-ical quantity that the pixels in an imagerepresent; the type of energy an imageprocessing system converts during imageacquisition. The most common modal-ity is visible light, but other modali-ties include invisible light (infrared or x-ray), sound (ultrasound) and magnetism(nuclear magnetic resonance).See alsocomputed tomography (CT), fluoroscopy,image acquisition,magnetic resonance imaging(MRI),multispectral image,nuclear magneticresonance, positron emission tomography(PET), synthetic aperture radar (SAR),tomography, ultrasound, X-ray image.

IMC Seeinternal model control.

immediate addressing an addressingmode where the operand is specified in theinstruction itself. The address field in the in-struction holds the data required for the op-eration.

immediate operand a data item containedas a literal within an instruction.

immersed flow a flow of electrons emittedfrom an electron gun exposed to the focusingmagnetic fields.

immittance a response function for whichone variable is a voltage and the other a cur-rent. Immittance is a general term for both

impedance and admittance and is generallyused where the distinction is irrelevant.

immunity to a disturbance an equipmentor systems capability to operate if an electro-magnetic disturbance occurs.

impact excitation excitation of an atomor molecule resulting from collision by an-other particle such as an electron, proton, orneutron.

IMPATT diode acronym for impactavalanche and transit time diode. Negativeresistance device used at high frequenciesused to generate microwave power. Typicallyused in microwave cavity oscillators.

impedance (1) electrical property of anetwork that measures its ability to conductelectrical AC current for a given AC volt-age. Impedance is defined as the ratio ofthe AC voltage divided by the AC currentat a given point in the network. In general,impedance has two parts: a real (resistive)part and an imaginary (inductive or capac-itive “reactive”) part. Unless the circuit ispurely resistive (made up of resistors only),the value of impedance will change with fre-quency.

(2) in an antenna, usually defined at theinput to an antenna, the impedance is the ra-tio of the applied (or induced) voltage to thecurrent flowing into (or out of) the antennainput. More generally, it is defined as the ra-tio of the electric field to the magnetic field.

impedance inverter circuit whose inputimpedance, when terminated with a load, isinversely proportional to said load impedance— typically implemented with a quarter-wave transmission line.

impedance matching one of the main de-sign activities in microwave circuit design.An impedance matching network is made upof a combination of lumped elements (re-sistors, capacitors, and inductors), or dis-tributed elements (transmission lines of vary-


ing characteristic impedance and length).Impedance matching networks transformnetwork impedance from one value to an-other. For example, on the input to a lownoise transistor, the impedance of an incom-ing 75 ohm transmission line would be trans-formed by the input matching network tothe impedance Zopt, required to achieve theminimum noise figure of the transistor. TheSmith chart is a tool commonly used by mi-crowave engineers to aid with impedancematching.

impedance matching network a com-bination of reactive elements connected be-tween a circuit and a load that transformsthe load impedance into another impedancevalue to achieve improved performance suchas maximum power transfer, reduced reflec-tions, or optimum noise performance.

impedance matrix the matrix formed inthe method of moments by expanding theelectric current in a series of basis functions,expressing the electric field in terms of thecurrent, and taking the inner product with aset of testing functions. The resulting matrixequation is of the form

[Vm] = [Zmn] [In]

where[Vm] represents the known excitation,[Zmn] is the impedance matrix, and[In] isthe unknown current.

impedance measurement measurementof the impedance (ratio of voltage to current)at a circuit port.

impedance parameters circuit parame-ters relating the voltages across the ports ofa multiport to the port currents.

impedance relay a protective relay thatsenses the operational impedance at a loca-tion, i.e., the ratio of voltage to current atany given time. During fault conditions onthe protected line, the impedance relay willsense the impedance (distance in ohms) be-tween the location of the relay and the fault.

Typical impedance relay characteristics aremho and reactance. Impedance relays arewidely used in sensing phase faults on trans-mission lines. Ground impedance relays areavailable that measure the distance to a sin-gle phase to ground fault using a modifiedtechnique. Also known as distance relay.

impedance scaling modifying a filter cir-cuit to change from a normalized set of ele-ment values to other usually more practicalelement values by multiplying all impedanceby a constant equal to the ratio of the de-sired (scaled) impedance to the normalizedimpedance.

impedance standard substrate (ISS) asubstrate, usually made of alumina, contain-ing thin film gold calibration standards usedin the calibration of vector network analyzersused in conjunction with microwave coplanarprobes.

impedance transformation Seematching.

implicant a first-order implicant is a pairof logically adjacent minterms. A second-order implicant is a set of logically adjacentfirst-order implicants, and so on.

implication a rule of reasoning whichstates that the truth of a proposition or condi-tion implies the truth of another. A commonimplication rule is of the form “If A then B.”

implied addressing a form of address-ing where the register or memory address isnot specified within the instruction but is as-sumed.

imprecise interrupt an implementationof the interrupt mechanism in which instruc-tions that have started may not have com-pleted before the interrupt takes place, andinsufficient information is stored to allow theprocessor to restart after the interrupt in ex-actly the same state. This can cause prob-lems, especially if the source of the interrupt


is an arithmetic exception.See alsopreciseinterrupt.

imprecision a sense of vagueness wherethe actual value of a parameter can assumethe specified value to within a finite tolerancelimit.

impressed current a current generatedfrom an independent source. Often used torepresent antennas.

improper modes in open waveguidesthe eigenfunctions relative to the continuousspectrum, which are defined over an infiniteinterval, are often referred to as impropermodes.

impulse a unit pulse.See alsoimpulsivetransient, surge.

impulse breakdown a test of electrical in-sulation in which lightning or switching im-pulses are applied.

impulse function a functionI (n) whosevalue is non-zero at sample positionn = 0,and zero elsewhere. The unit impulse func-tion δ(n) is unity at sample positionn = 0and zero elsewhere. A shifted impulse is alsotermed an impulse.See alsoDirac’s deltafunction.

impulse generator (1) an electronicdevice delivering single pulses of variousshapes, preferably square.

(2) a high-voltage trigger generator.

impulse noise refers to a noise processwith infrequent, but very large, noise spikes;it is also known as shot noise or Salt and Pep-per noise. The phrase “Impulsive noise” isfrequently used to characterize a noise pro-cess as being fundamentally different fromGaussian white noise, in being derived from aprobability density function with very heavy(long) tails. Applying impulsive noise to animage leaves most pixels unaffected, but withsome pixels very bright or dark.

impulse response the output of a lin-ear time-invariant system when the input isa pulse of short time duration. The systemcan be entirely characterized by the impulseresponse.

In the case of a continuous time systemwith input f (t), the impulse signalδ(t) isdefined as

(i) δ(t) = 0, t 6= 0

(ii)

∫ ε

−εδ(t)dt = 1, for any ε > 0,

and the impulse response is the zero state sys-tem response to an inputf (t) = δ(t). Inthe case of a discrete time system with inputf [k], the impulse signal is defined as

(i) δ[k] = 1, k = 0

(ii) δ[k] = 0, k 6= 0 ,

and the impulse response is the zero state sys-tem response to inputf [k] = δ[k].

impulse train a signal that consists of aninfinitely long series of Dirac delta functionswith periodT :

∑∞n=−∞ δ(t−nT ). This sig-

nal can be used to sample a continuous-timewaveform.

impulsive transient a rapid frequencyvariation of voltage or current during steady-state operation in which the polarity is mostlyunidirectional.

in-circuit emulator (ICE) a device thatreplaces the processor and provides the func-tions of the processor plus testing and debug-ging functions.

in-line gun a CRT electron gun structurethat has the red, green, and blue electron guncomponents aligned in a horizontal plane.The in-line gun structure requires color regis-tration (color convergence) correction in thevertical CRT face plate axis only.

in-order issue the situation in which in-structions are sent to be executed in the orderthat the instructions appear in the program.


in-phase signal in quadrature modula-tion, the signal component that multipliescos 2πfct , wherefc is the carrier frequency.

INA Seeinverse Nyquist array.

incandescent lamp a lamp made by heat-ing a metal filament in vacuum; not a burningcandle.

incident power power in an electromag-netic wave that is traveling in an incident di-rection.

incoherent illumination a type of illu-mination resulting from an infinitely largesource of light that illuminates the mask withlight from all possible directions. This ismore correctly called spatially incoherent il-lumination.

incoherent light light in which detailedknowledge of the fields at one point in spacedoes not permit prediction of the fields at an-other point.

increment to add a constant value (usu-ally 1) to a variable or a register. Pointers tomemory are usually incremented by the sizeof the data item pointed to.

incremental encoder similar to an abso-lute encoder, except there are several radiallines drawn around the disc, so that pulses oflight are produced on the light detector as thedisc rotates. Thus incremental informationis obtained relative to the starting position.A larger number of lines allows higher res-olution, with one light detector. A once perrev counter may be added with a second lightdetector.See alsoencoder.

incremental gain a systemH : Xe → Xeis said to be Lipschitz continuous, or simplycontinuous, if there exists a constant0(H) <

∞, called incremental gain, such that

0(H) = sup‖(Hx1)T − (Hx2)T ‖

‖(x1)T − (x2)T ‖

where the supremum is taken over allx1, x2 ∈ Xe and all real numbersT for whichx1T 6= x2T . See alsoextended space, gain.

incremental model small-sign differ-ential (incremental) semiconductor diodeequivalent RC circuit of a diode, biased ina DC operating point.

incremental passivity a systemH :Xe → Xe is said to be incrementally pas-sive if

〈Hx1 −Hx2, x1 − x2〉T ≥ 0

∀x1, x2 ∈ XeSee alsoextended space, inner product space,and passivity.

incremental resistance the slopedV/dIof the I-V characteristic of a circuit element.

incremental strict passivity a systemH :Xe → Xe is said to be incrementally strictlypassive if there existδ such that

〈Hx1 −Hx2, x1 − x2〉T ≥δ 〈x1 − x2, x1 − x2〉T

∀x1, x2 ∈ XeSee alsoextended space, inner product space,passivity, strict passivity, and incrementalpassivity.

incrementally linear system a systemthat has a linear response to changes in theinput, i.e., the difference in the outputs is alinear (additive and homogeneous) functionof the difference in the inputs.See alsolinearsystem.

independence a complete absence ofany dependence between statistical quanti-ties. In terms of probability density functions(PDFs), a set of random quantities are inde-pendent if their joint PDF equals the productof their marginal PDFs:

px1,x2,...(x1, x2, . . .) = px1(x1)·px2(x2)·. . . .


Independence implies uncorrelatedness.Seealso correlation. See alsoprobabilitydensity function.

independent and identically distributed(IID) a term to describe a number of ran-dom variables, each of which exhibits iden-tical statistical characteristics, but acts com-pletely independently, such that the state oroutput of one random variable has no influ-ence upon the state or output of any other.

independent event event with the prop-erty that it gives no information about theoccurrence of the other events.

independent identically distributed pro-cess a random processx[i], wherex[i]and x[j ] are independent fori 6= j , andwhere the probability distributionp(x[i]) foreach element of the process is not a func-tion of i. See alsoindependence, probabilitydensity function, random process.

independent increments process a ran-dom processx(t), where the process incre-ments over non-overlapping periods are in-dependent. That is, forti < si , si ≤ ti+1,then

(x(s1)− x(t1)) , (x(s2)− x(t2)) , . . .

are all independent.See alsoindependence,random process.

independent joint control control of asingle joint of a robot while all the other jointsare fixed.

index that part of memory address used toaccess the locations in the cache, generallythe next most significant bits of the addressafter the tag.

index assignment (IA) each reproductionvector in a (vector) quantizer is representedby an index that is transmitted or stored. Theindex assignment problem is the problem ofassigning indices to reproduction vectors, insuch a way that errors in the transmission (or

storage) influence the reproduction fidelity aslittle as possible.

index grating through the photorefractiveeffect. This volume index grating is referredto as a photorefractive grating. The photo-induced refractive index grating can also beerased by the illumination of another beamof light. See alsoindex-guided laser diode.

index of fuzziness an index to measure thedegree of ambiguity of a fuzzy set by com-puting the distance between the fuzzy set andits nearest ordinary set.

index of refraction (1) a parameter of amedium equal to the ratio of the velocity ofpropagation in free space to the velocity ofpropagation in the medium. It is numericallyequal to the square root of the product of therelative permittivity and relative permeabilityof the medium.

(2) in optics, a complex quantity describ-ing the optical transmission properties of amedium; the real part corresponds to the ra-tio of speed of light in vacuum to the speed oflight in the medium, while the imaginary partcorresponds to attenuation of the light by themedium.

index register register used to index a datastructure in memory; the starting address ofthe structure will be stored in a base register.Used in indexed addressing.

index-guided laser electrically pumpedsemiconductor laser in which the mode fieldsare confined in the transverse direction by theprofile of the index of refraction.

index-guided laser diode See index-guided. laser.

indexed addressing an addressing modein which an index value is added to a baseaddress to determine the location (effectiveaddress) of an operand or instruction in mem-ory. Typically, the base address designatesthe beginning location of a data structure in


memory such as a table or array and the in-dex value indicates a particular location inthe structure.

indexing Seeindexed addressing.

indicator Seeflag.

indirect addressing an addressing modein which an index value is added to a baseaddress to determine the location (effectiveaddress) of an operand or instruction in mem-ory. Typically, the base address designatesthe beginning location of a data structure inmemory such as a table or array and the in-dex value indicates a particular location in thestructure. Also known as deferred address-ing.

indirect weld parameters (IWP) a col-lection of parameters that establish the weld-ing equipment setpoint values. Examplesinclude voltage, current, travel, speed, elec-trode feed rate, travel angle, electrode geom-etry, focused spot size, and beam power.

indiscernibility relation in Pawlak’s in-formation systemS = (U,A) whose uni-verseU hasn members denotedx, y, z, . . .and the setA consisting ofmattributesaj , theindiscernibility relation onU with respect tothe setB ⊆ A defines the partitionU/B ofUinto equivalence classes such thatx, y ∈ U

belong to the same class ofU/B if for everyattributea ∈ B,

a(x) = a(y).

In terms of negotiations, the partitionU/Bmay be interpreted as the partition of the in-verseU into blocks of negotiators that havethe same opinion on all of the issues ofB.

induced emission enhanced emission ofelectromagnetic radiation due to the presenceof radiation at the same frequency; also calledstimulated emission.

induced voltage voltage produced by atime-varying magnetic flux linkage.See alsoFaraday’s law.

inductance a parameter that describes theability of a device to store magnetic flux. Theunits are henrys per meter.

induction furnace a method of smeltingor heating metals with eddy currents inducedby a high-frequency coil surrounding the cru-cible.

induction generator an induction ma-chine operated as a generator. If the machineis connected to an AC system and is drivenat greater than synchronous speed, the ma-chine can convert mechanical energy to elec-trical form. The induction generator requiresa source of reactive power.

induction machine classification of anyof a variety of electrical machines in which anAC current in the stator coils is used to pro-duce a rotating magnetic flux that, by Faradayaction, induces an AC voltage in a set of coils(the induction coils) on the machine’s rotor.The rotor coils are shorted to cause a secondAC current to flow in the rotor coils, whichproduces, in turn, a second rotating flux. Theinteraction of the rotor- and stator-producedfluxes creates torque.

induction motor Seeinduction machine.

induction regulator See inductionvoltage transformer.

induction theorem states that if the in-cident tangential magnetic and electric fieldsare known everywhere on some closed sur-face, then these fields may be replaced withequivalent electric and magnetic surface cur-rents.

induction voltage transformer speciallyconstructed transformer with a rotating pri-mary coil that is used to provide voltage regu-lation on individual power circuits. The sec-


ondary of an induction regulator is mountedon the stationary shell of cylindrical core, andthe primary is mounted on a movable, centerrotor. In the neutral position, the magneticaxes of the primary and secondary coils areoriented 90 degrees to each other, reducingthe magnetic coupling to zero. In this posi-tion, energizing the primary does not inducevoltage in the secondary; however, rotatingthe primary coil in either direction from theneutral position creates mutual flux linkageand causes a secondary voltage to appear.Rotation in one direction causes secondaryvoltage to be in phase with the primary; ro-tation in the opposite direction causes sec-ondary voltage to be out of phase with theprimary. Voltage regulation is provided byconnecting the primary coil across the line tobe regulated and connecting the secondarycoil in series with the load. By positioningthe primary coil based on load demand inthe line, secondary voltage can be used toadjust line voltage either up or down. In-duction regulators are also equipped with ashort-circuited coil mounted on the primaryin spatial quadrature with the primary coil. Inthe neutral position, this coil has maximumcoupling with the secondary coil, which min-imizes the inductive reactance in the load linedue to the secondary coil.

inductive coupling a means of transfer-ring electrical energy from one part (area)of a circuit to another part without requir-ing any ohmic (wire) connection. Instead,magnetic flux linkages couple two inductors(coils). The coils must be in close proximityin order to establish sufficient mutual induc-tance.

inductive discontinuity a type of discon-tinuity that exhibits an inductive, or quasi-inductive, behavior. As an example, mag-netic plane iris in metallic rectangular waveg-uides produces this type of response.

inductor a two-terminal electrical ele-ment that satisfies a prescribed algebraic re-lationship in the flux-current(φ − I ) plane.

industrial inspection the use of computervision in manufacturing, for example, to lookfor defects, measure distances and sizes, andevaluate quality.See alsocomputer vision,inspection system.

industrial robot a mechanical structureor manipulator that consists of a sequence ofrigid body (links) connected by means of ar-ticulation joints. A manipulator is character-ized by an arm, a wrist, and an end-effector.Motion of the manipulator is imposed by ac-tuators through actuation of the joints. Con-trol and supervision of manipulator motion isperformed by a control system. Manipulatoris equipped with sensors that measure the sta-tus of the manipulator and the environment.

inelastic scattering situation where thereis energy exchange between the impingingwave and the medium. Inelastic scatteringalso produces frequency shifts.

inertia constant the energy stored inan electric machine running at synchronousspeed and given in megajoules per mega-voltampere of machine rating.

inertial confinement a method of initiat-ing and containing the plasma of a fusion re-action by illuminating a pellet of solid heavyhydrogen with powerful laser beams from alldirections simultaneously.

inference engine a reasoning mechanismfor manipulating the encoded informationand rules in a knowledge base to form in-ferences and draw conclusions.

infinite bus an electrical supply with suchlarge capacity that its voltage (and frequency,if AC) may be assumed constant, indepen-dent of load conditions. If a machine’s ca-pacity is small relative to the electric supplysystem to which it is connected, it may beassumed to be operating on an infinite bus.

infinite impulse response (IIR) filterany filter having an impulse response of infi-


nite length, therefore having a frequency re-sponse with at least one pole. In direct con-trast with finite impulse response (FIR) fil-ters, IIR filters employ feedback, which leadsto an impulse response of infinite extent; con-sequently a causal IIR filter will depend onall previous inputs. As a simple example, ev-ery autoregressive process can be expressedas the output of an IIR filter driven by whitenoise.See alsofinite impulse response filter,autoregressive .

infinite symmetric exponential filter(ISEF) one type of optimal edge detecter.In two dimensions, the filter is

f (x, y) = a × e−p(|x|+|y|)

The resulting edge detector has excellent lo-calization and signal-to-noise ratio.Com-pare withCanny edge detector.

infinite-dimensional dynamical systema linear stationary system described by thefollowing abstract ordinary differential stateequation:

x′(t) = Ax(t)+ Bu(t)

wherex(t) is the state vector that belongs toinfinite-dimensional Banach spaceX, u(t)is the input vector that belongs to infinite-dimensional Banach spaceU , A is a lineargenerally unbounded operator that is a gen-erator of a strongly continuous semigroup oflinear bounded operatorsS(t) : X → X, fort > 0,B : u → x is a linear bounded oper-ator. The solution of the state equation hasthe form

x(t, x(0), u) = S(t)x(0)

+∫ t

0S(t − s)Bu(s)ds

information a mathematical model of theamount of surprise contained in a message.For a discrete random source with a finitenumber of possible symbols or messages theinformation associated with the symbolxk is

Ik = − log2(pk)

wherepk is the probability of the symbolxk. The expected value of the informationof the symbols is the first order entropy ofthe source.

information gain for an attribute in a setof objects to be classified, a measure of theimportance of the attribute to the classifica-tion. The information gainGi of the ith at-tributeAi of a set ofn objectsS in the clas-sification is defined as

Gi = I (S)− Ei

whereI (S) is the expected information (orentropy) for the classification andEi is theexpected information required for the valueof Ai to be known.I (S) is defined as

I (S) = −Nc∑c=1

nc

nlog2

nc

n

whereNc is the total number of classes in theclassification, andnc is the number of objectsin thecth classCc. Ei is defined as

Ei =Ni∑k=1

nik

nI (Sik)

whereSik is the subset ofS in whichAi of allobjects takes itskth value,Ni is the numberof valuesAi can take,nik is the number ofobjects inSik, and the information requiredin Sik is

I (Sik) = −Nc∑c=1

nikc

niklog2

nikc

nik

wherenikc is the number of objects inSikbelonging to classCc.

information hiding a program designprinciple that makes available to a functiononly the data it needs.

information theory theory relating theinformation content of a message to its repre-sentation for transmission through electronicmedia. This subject includes the theory of


coding, and also topics such as entropy, mod-ulation, rate distortion theory, redundancy,sampling.

infrared detector semiconductor or othermaterial with a measurable change, usually,but not always, in electrical conductivity,when exposed to infrared light.

infrared (IR) invisible electromagneticradiation having wavelengths longer thanthose of red light; often considered to rangefrom about 0.7 micrometers to 100 microm-eters.

infrared laser laser producing its outputin the infrared region of the spectrum.

InGaAs/InAlAs/InP lattice matchedsemiconductor heterostructure system usedfor

1. optical fiber communications sourcesand detectors, and

2. high-speed electronic circuits.

inheritance in object orientation, the pos-sibility for different data types to share thesame code.

inhibit to prevent an action from takingplace; a signal that prevents some action fromoccurring. For example, the READY signalon a memory bus for a read operation may in-hibit further processing until the data item hasbecome available, at which time the READYsignal is released and the processor contin-ues.

inhibit sense multiple access (ISMA) pro-tocol multiple access protocol attempt-ing to solve the hidden station problem. InISMA, the base station transmits an “idle”signal at the beginning of a free slot. Eachuser has to monitor this signal before it at-tempts to access the base station. ISMAis sometimes called busy tone multiple ac-cess (BTMA).See alsobusy tone multipleaccess.

inhomogeneous broadening spectralbroadening of a transition in a laser mediumdue to intrinsic differences between the con-stituent atoms or molecules.

inhomogeneous medium medium whoseconstitutive parameters are functions of po-sition. For example, in optical media, theproperties of gain, index of refraction, etc.,might be functions of position.

initial rest a system satisfies an initialrest assumption if the output of the systemis zero until the input of the system becomesnonzero (x(t) = 0, t ≤ t0 ⇒ y(t) = 0, t ≤t0). Systems described by linear constant-coefficient differential or difference equa-tions are causal, linear, and time invariant ifat initial rest.

initialize (1) to place a hardware systemin a known state, for example, at power up.

(2) to store the correct beginning data in adata item, for example, filling an array withzero values before it is used.

injection electroluminescence refers tothe case where the electroluminescence theinjection of carriers across a p-n junction.

injection laser diode semiconductor laserin which amplification is achieved by injec-tion of holes or electrons into the active re-gion.

injection locking two oscillators can beinjection-locked together by providing smallcoupling between them. In that case, thephase relation between the two oscillators isgoing to be constant over time.

inner product space over the fieldF , His an inner product space if it is a linear spacetogether with a function〈, 〉 : H × H → F ,such that

1. 〈x, y〉 = 〈y, x〉, ∀x, y ∈ H.2. 〈λx + µy, z〉 = λ〈x, z〉 + µ〈y, z〉,

∀x, y, z ∈ H,∀λ,µ ∈ F .3. 〈x, x〉 ≥ 0.


4. 〈x, x〉 = 0 if and only ifx = 0.

The fieldF could be either the real or thecomplex numbers.

input data conveyed to a computer systemfrom the outside world. The data is conveyedthrough some input device.

input admittance the inverse of inputimpedance, measured in mhos.

input backoff Seebackoff.

input buffer a temporary storage areawhere input is held. An input buffer is nec-essary when the data transfer rate from an in-put device is different from the rate at whicha computer system can accept data. Havingan input buffer frees the system to performother tasks while waiting for input.

input current shaping a technique toforce the line current to assume the sameshape as the line voltage of a rectifier in orderto achieve low harmonic distortion and highpower factor correction.

input device a peripheral unit connectedto a computer system, and used for trans-ferring data to the system from the outsideworld. Examples of input devices includekeyboard, mouse, and light pen.

input impedance (1) when a voltage isapplied to a conducting material, a currentwill flow through the material. The ratio ofthe voltage to the current is known as the in-put impedance and is a complex number withmagnitude measured in units of ohms.SeealsoOhm’s Law.

(2) impedance seen when looking into theinput terminals of an antenna.

input layer a layer of neurons in a networkthat receives inputs from outside the network.In feedforward networks, the set of weightsconnected directly to the input neurons is of-ten also referred to as the input layer.

input neuron a neuron in a network thatreceives inputs from outside the network. Infeedforward networks, the set of weights con-nected directly to the input neurons is oftenalso referred to as the input layer.

input register a register to buffer trans-fers to the memory from the I/O bus. Thetransfer of information from an input devicetakes three steps:

1. from the device to its interface logic,2. from the interface logic to the input reg-

ister via the I/O bus, and3. from the input register to the memory.

input return loss negative ratio of the re-flected power to the incident power at theinput port of a device, referenced to thesource or system impedance, expressed indecibels. The negative sign results from theterm “loss.” Thus, an input return loss of20dB results when 1/100th of the incidentpower is reflected.

input routine (1) a function responsiblefor handling input and transferring the inputto memory.

(2) the software subroutine used to com-municate with an input device.See alsoI/OcontrollerandI/O routine.

input stability circle circular regionshown graphically on a Smith chart that de-fines the range of input impedance termina-tions for which an active device or circuit re-mains unconditionally stable, and those re-gions for which it is potentially unstable, ata particular frequency.

input vector a vector formed by the inputvariables to a network.

input-output curve a plot of the rate offuel consumption vs. the power output for athermal generating plant.

input–output equivalence of generalized2-D system matrices the 2-D polynomial


matrix [P −QC D

]P := Ez1z2 − A0 − A1z1 − A2z2

Q := B0 + B1z1 + B2z2

is called the system matrix of the generalized2-D model


+ A2xi,j+1 + B0uij

+ B1ui+1,j + B2ui,j+1

yij = Cxij +Duij

i, j ∈ Z+ (the set of nonnegative integers)wherexij ∈ Rn is the semistate local vector,uij ∈ Rm is the input vector,yij ∈ Rp is theoutput vector andE,Ak,Bk (k = 0,1,2),C,D are real matrices withE possibly singularor rectangular. Two system matricesSi :=[Pi −QiCi Di

]for i = 1,2 are called input–output

equivalent ifC1P−11 Q1+D1 = C2P

−12 Q2+

D2.

input-output stability system conditionwhere bounded outputs are obtained as a re-sponse to bounded inputs.See alsoboundedfunction.

input/output channel (I/O channel) in-put/output subsystem capable of executing aprogram, relieving the CPU of input/output-related tasks. The channel has the abil-ity to execute read and write instructionsas well as simple arithmetic and sequenc-ing instructions that give it complete con-trol over input/output operations. (IBM ter-minology) See alsodirect memory access,selector channel, multiplexer channel.

input/output port a form of register de-signed specifically for data input/output pur-poses in a bus-oriented system.

INR Seeinterference to noise ratio.

inrush current the transient currentdrawn by an electrical apparatus when it is

suddenly connected to a power source. Theinrush current may be larger in magnitudethan the steady-state full-load current. Thetransient response is short in time and theelectrical equipment generally supports theinrush current, provided it does not happenfrequently. For a single transient, the ther-mal limit of the equipment is not reached, butif it is switched on and switched off severaltimes within a short period, the temperaturecan rise very quickly. In case of transform-ers, the inrush current is not sinusoidal evenif the voltage is due to the hysteresis of theferromagnetic core.

insertion loss (1) worst-case loss of thedevice across the stated frequency range. Theloss due to the insertion of the unit in serieswith a signal path.

(2) transmission loss of an RF or mi-crowave component or system, typicallymeasured in decibels.

insolation incident solar radiation.

inspection system an image processingsystem, usually automatic or semi-automatic,that performs industrial inspection.

instantaneous the range of 0.5 to 30 cy-cles of the supply frequency.

instantaneous contact the contacts of acontactor or relay that open or close with notime delay.

instantaneous frequency (1) the time rateof change of the angle of an angle-modulatedwave.

(2) the frequency of radiation at some cho-sen instant of time; the rate of change of phasein radians per second, divided by 2π .

instantaneous overcurrent relay anovercurrent relay that operates with no in-tentional delay following sensing of a powerfrequency overcurrent, i.e., a current aboveits set point.


instantaneous power power in an AC ormodulated signal at a given instant in time.

instantaneous-trip circuit breaker es-sentially an inverse-time circuit breaker withthe thermal element removed. It will onlytrip magnetically in response to short-circuitcurrents. Thus, it is often referred to by othernames, such as magnetic circuit breaker andmagnetic-only circuit breaker.

instar Seeinstar configuration.

instar configuration a term used for aneuron fed by a set of inputs through synapticweights. An instar neuron fires whenever aspecific input pattern is applied. Therefore,an instar performs pattern recognition.Seealso instar training. Compare withoutstarconfiguration.

instar training an instar is trained to re-spond to a specific input vector and to noother. The weights are updated according to

wi(t + 1) = wi(t)+ µ(xi − wi(t))

wherexi is the ith input,wi(t) the weightfrom inputxi , µ is the training rate startingfrom about 0.1 and gradually reduced duringthe training. See alsoinstar configuration.Compare withoutstar training.

Institute of Electrical and Electronics En-gineers (IEEE) a professional organiza-tion of electrical engineers and computer sci-entists. The world’s largest professional or-ganization.

instruction specification of a collection ofoperations that may be treated as an atomicentity with a guarantee of no dependenciesbetween these operations.

instruction access fault a fault, signaledin the processor, related to abnormal instruc-tion fetches.

instruction cache Seecode cache.

instruction counter the memory registerwithin a computer that maintains the locationof the next instruction that will be fetchedfor execution. This register is also called aprogram counter.

instruction decoder the part of a proces-sor that takes instructions as input and pro-duces control signals to the processor reg-isters as output. All processors must per-form this function; some perform it in severalsteps, with part of the decoding performedbefore instruction issue and part after.

instruction decoder unit in modern CPUimplementations, the module that receives aninstruction from the instruction fetch unit,identifies the type of instruction from theopcode, assembles the complete instructionwith its operands, and sends the instructionto the appropriate execution unit, or to an in-struction pool to await execution.

instruction fetch unit in modern CPUimplementations, the module that fetches in-structions from memory, usually in conjunc-tion with a bus interface unit, and preparesthem for subsequent decoding and executionby one or more execution units.

instruction field a portion of an instruc-tion word that contains a specified value,such as a register address, a 16-bit immediatevalue, or an operand code.

instruction format the specification ofthe number and size of all possible instructionfields in an instruction-set architecture.

instruction issue the sending of an in-struction to functional units for execution.

instruction pipeline a structure thatbreaks the execution of instruction up intomultiple phases, and executes separate in-structions in each phase simultaneously.

instruction pointer another name for pro-gram counter, the processor register holding


the address of the next instruction to be exe-cuted.

instruction pool in modern CPU imple-mentations, a holding area in which instruc-tions that have been fetched by an instructionfetch unit await access to an execution unit.

instruction prefix a field within a pro-gram instruction word used for some spe-cial purpose. Found only rarely. The IntelX86 architecture occasionally uses instruc-tion prefixes to override certain CPU address-ing conventions.

instruction reordering a technique inwhich the CPU executes instructions in anorder different from that specified by the pro-gram, with the purpose of increasing the over-all execution speed of the CPU.

instruction repertoire See instructionset.

instruction scheduling the relocationof independent instructions, which is doneto maximize instruction-level parallelism(and/or minimize instruction stalls).

instruction set the instruction set of a pro-cessor is the collection of all the machine-language instructions available to the pro-grammer. Also known as instruction reper-toire.

instruction window for an out-of-orderissue mechanism, a buffer holding a groupof instructions being considered for issue toexecution units. Instructions are issued fromthe instruction window when dependencieshave been resolved.

instruction-level parallelism the conceptof executing two or more instructions in par-allel (generally, instructions taken from a se-quential, not parallel, stream of instructions).

instrument transformer a two-windingtransformer designed and optimized for me-

tering applications. The essential featuresare accurate input to output ratios for themeasured parameter and minimal burden (orload) on the circuit being tested.See alsocurrent transformer, voltage transformer.

insulated gate bipolar junction transistor(IGBT) a hybrid electronic switch thathas the high input impedance and high speedcharacteristic of a MOSFET with the conduc-tivity characteristic (low saturation voltage)of a bipolar transistor.

insulation class specification of insulat-ing material. Primarily determined by themaximum temperature that the material canwithstand. The class of insulating systemused in a transformer or electrical machineis a key factor in determining its rated load.

insulation coordination the practice ofselecting the insulation of a power sys-tem such that breakdowns from overvoltages(such as from a lightning strike) will occur atpoints where the least harm will result.

insulation resistance measured value ofelectrical resistance of the insulation of aproduct or device.

insulator a device designed to separateand prevent the flow of current between con-ductors. Properties of the dielectric (insu-lating) material and geometry of the insula-tor determine maximum voltage and temper-ature ratings.

insulator string a chain of two or morestrain insulators that are coupled together toincrease the total insulation level of the as-sembly.

integer a fixed-point, whole number valuethat is usually represented by the word sizeof a given machine.

integer unit in modern CPU implemen-tations, a type of execution unit designed


specifically for the execution of integer-typeinstructions.

integral control control scheme wherebythe signal that drives the actuator equals thetime integral of the difference between theinput/desired output and the measured actualoutput.

integral equation a type of equationwhere the unknown function appears underan integral. When this is the only place thatthe unknown appears, the integral equation iscommonly called a first-kind equation, whileif the unknown also appears outside the inte-gral, it is a second-kind integral equation.

One-dimensional examples are

y(s) = λ

∫ b

a

K(s, t)x(t)dt

and

x(s) = y(s)+ λ

∫ b

a

K(s, t)x(t)dt

whereλ is a possibly complex scalar param-eter andy(s) is known (often called the driv-ing term);x(t) is the unknown function andK(s, t) is the known kernel. The above arelinear Fredholm integral equations of the firstand second kind, respectively.

integral horsepower motor a motor builtin a frame as large as or larger than that ofa motor of open construction having a con-tinuous rating of one horsepower at 1700–1800 rpm.

integrated AOTF acousto-optical tunablefilter device using integrated wave guide op-tics and surface acoustic waves (SAWs) foracousto-optical interaction.

integrated circuit (IC) (1) an assemblyof miniature electronic components simulta-neously produced in batch processing, on orwithin a single substrate, that performs anelectronic circuit function.

(2) many transistors, resistors, capacitors,etc., fabricated and connected together to

make a circuit on one monolithic slab ofsemiconductor material.

integrated optics also known as guided-wave optics; generally, describes the devicesthat couple to fiber optics, but does not in-clude the fibers.

integrated project support environmenta software engineering environment support-ing all the stages of the software processfrom initial feasibility studies to operationand maintenance.

integrated services digital network (ISDN)a network that provides end-to-end digital

connectivity to support a wide range of ser-vices, including voice and nonvoice services,to which users have access by a limited setof standard multipurpose user–network inter-faces.

integrating A/D device that takes an ana-log input signal and integrates it over timeto produce a digital signal that represents thearea under the curve, or the integral.

integrator for an input ofx(t) and an out-put of

∫ t−∞ x(τ)dτ , is one of the basic signal

processing building blocks. Integrators canbe implemented using operational amplifiers.

integrity (of data) (1) a belief in the truthof the information represented by a set ofdata.

(2) a condition stating that the informationin a set of data does satisfy a set of logicalconstraints (the integrity constraints).

intelligence the aggregated and processedinformation about the environment, includ-ing potential adversaries, available to com-manders and their staff.

intelligent control a sensory-interactivecontrol structure incorporating cognitivecharacteristics that can include artificial intel-ligence techniques and contain knowledge-based constructs to emulate learning behav-


ior with an overall capacity for performanceand/or parameter adaptation. Intelligent con-trol techniques include adaptation, learning,fuzzy logic, neural networks, genetic algo-rithm, as well as their various combinations.

intelligent material Seesmart material.

intelligibility objective quantitative mea-sure of speech perception.

intensifying transformer one of a classof hedges — for example very, extremely— which intensifies the characteristics of afuzzy set.

intensity a measure of the strength of alight field. Two different definitions of in-tensity are commonly encountered. Espe-cially in the discipline of radiometry, inten-sity is usually defined to be power per unitsolid angle. Alternatively, especially in laserphysics, define intensity to be power per unitarea. The intended definition can usually bededuced from context or from dimensionalanalysis.

intensity average spatially or temporallyaveraged value of the intensity.

intensity modulation alteration of thepower level of an optical beam, usually toimpose information onto the beam.

intensity optical same as optical powerper unit area; time-average value of thePoynting vectorE×H measured in watts permeter squared, sometimes also representableas the energy density times the speed of light.

intensity-dependent refractive index theproperty of many materials that the measuredvalue of the refractive index depends on theintensity of the light field used to performthe measurement. Typically, the refractiveindex changes by an amount describable byδn = n2I , wheren2 is the coefficient of thenonlinear refractive index andI is the inten-

sity of the field in units of power per unitarea.

inter-area mode oscillation this mode ofoscillation is characterized by the swingingof many machines in one part of the systemagainst machines in other parts. These oscil-lations are caused by the presence of weakties between two or more areas.

inter-cell interference the interferencecaused by the transmitters in a cellular com-munications system (base station or mobileterminal) on a receiver of interest where theinterfering transmitters are located on cellsother than the receiver cell.Compare withintra-cell interference.

inter-feature distance the distance be-tween a pair of feature points in an image,often used to help identify the object, or (inconjunction with other measures of this type)to infer the presence of the object.

inter-record gap the space between tworecords of data stored on a magnetic mediumsuch as a tape. This space helps preventinginterference between the two records. It canalso contain markers marking the beginningand end of the records.

interaction region a region where an elec-tron beam and microwaves transfer energy toeach other.

interaural attribute attribute of ear in-put signals (e.g., localization in the horizontalplane) that depend on differences between, orratios of measures of, two ear input signals.

interconnect the physical manifestationof an interconnection network.

interconnection density number of in-terconnection channels per unit area or unitvolume. The interconnection density of opti-cal interconnects is high because unlike elec-trons, light beams can cross each other with-out any change.


interconnection network the combina-tion of switches and wires that allows multi-ple processors in a multiprocessor or multi-computer to communicate.

interconnections and packaging (I&P)elements of the electrical signal transmissionpath from the driver chip to the receiver chip.Various elements that make up I&P are chipwirebonds and package pins, circuit boards,connectors, motherboards, and cables.

interdigital capacitor a capacitor con-sisting of two pieces of electrodes that havemultiple fingers. Each finger of these elec-trodes is adjacently arranged in turn with anarrow gap. It utilizes the gap capacitancebetween adjacent fingers.

interdigital filter filter consisting of par-allel coupled transmission line resonatorswhere each resonator is grounded on one sideand is either open-circuited or capacitivelyloaded to ground on the other side. Adja-cent resonators are grounded on the oppositeends.

interface the set of rules specified forcommunicating with a defined entity.

interface state for a semiconductor devicesuch as a transistor, a region where trappingof charge carriers can occur at the interfacebetween regions of differing electrical prop-erties.

interface trap energy level at the insu-lator/semiconductor interface that can trapelectrons or holes. The occupation of thesetraps follow Fermi statistics.

interference (1) a process in which twowaves interact. Interference occurs when thecomplex properties of the wave are combinedinstead of simply their two amplitudes. Thewave interference can be constructive whenthey both have the same phase, or destructivewhen they have opposite phases.

(2) a disturbance on a communication sys-tem caused by the presence of a signal, orsmall number of signals, of man made origin.The interference may be due to the presenceof signals external (other systems) or internal(e.g., multiple access interference) to the sys-tem of interest.Compare withbackgroundnoise.

(3) an interaction whereby a process thatis supposed to be independent (and isolated)from another process does indeed performsome action that has an effect visible from thesecond process. Interference can be causedby direct writing to the state of the secondprocess or by indirect effects, such as by mo-nopolizing the use of a system module.

(4) any external electrical or electromag-netic disturbance which causes an undesiredresponse or degradation of the desired signal.

(5) the mutual influence of two wavesor vibrations of any kind, producing cer-tain characteristic distorting phenomena, isknown as interference. For example, radio-frequency interference (RFI) or crosstalk andintersymbol interference (ISI) caused by in-terference between successive pulses in anoptical fiber.

interference cancellation a signal pro-cessing technique where an interfering sig-nal is estimated, regenerated based on the es-timate, and then canceled from the receivedsignal, leaving a potential interference-free,desired signal. Examples are multiple ac-cess interference cancellation, co-channel in-terference cancellation, echo cancellation.

interference channel a multiple transmit-ter, multiple receiver communications sys-tem in which each received signal is a (pos-sibly non-deterministic) function of two ormore transmitted signals.See alsobroadcastchannel, multiple access channel.

interference to noise ratio (INR) the totalinterference measured at a receiver dividedby the receiver noise power. The INR is auseful measure for comparing the interfer-


ence distributions between different types ofradio systems.

interframe referring to the use of morethan a single image frame (e.g., an algorithmthat differences successive image frames ordetects frame-to-frame motion).See alsointraframe.

interframe coding (1) the process ofcoding the difference of consecutive videoframes based on motion estimation, insteadof the frames themselves.

(2) image coding schemes that utilize thetemporal correlation between image frames.

interframe coding 3-D in situationswhere motion picture film or television im-ages need to be coded, three-dimensionaltransform coding is used. In this approach,the frames are codedL at a time and par-tioned intoL×L×L blocks per pixel priorto transformation.

interframe hybrid transform codingL × L blocks of pixels are first 2-D-transformed. The resulting coefficients arethen DPCM (interframe) coded and transmit-ted to the receiver. This is then reversed at thereceiver for reconstruction. In the interframehybrid transform coding, the DC componentsare determined exclusively by the amount ofspatial detail in the picture.

interframe prediction this techniqueuses a combination of pixels from the presentfield as well as previous fields for prediction.For scenes with low detail and small motion,field difference prediction appears to performbetter than frame difference prediction.

interharmonic component the frequencycomponent that is not an integer value of theperiodic supply waveform.

interior permanent magnet machine apermanent magnet machine in which the per-manent magnets are buried in the rotor iron.With this construction the airgap is small, al-

lowing the airgap flux to be reduced by cur-rent control. This allows an operating modein which the top speed of the motor can be asmuch as five times the base speed, and is oneof the main features of this motor.

interlaced Seeinterlaced scanning.

interlaced scanning a system of televi-sion and video scanning whereby the down-ward rate of travel of the scanning electronbeam is increased by sending every other linewhereby field 1 (odd field) is superimposedon field 2 (even field) to create one pictureframe. Interlaced scanning is used to reduceflicker.

interleaved memory a memory systemconsisting of several memory modules (orbanks) with an assignment of addresses thatmakes consecutive accesses fall into differ-ent modules. This increases the effectivememory bandwidth, as several memory re-quests can be satisfied (concurrently by sev-eral modules) in the same time as it wouldtake to satisfy one memory requests. Forexample, a simple arrangement (sequentialinterleaving), which favors sequential ac-cess, is to assign addressk to memory bankk modN , whereN is the number of mod-ules employed: the reference stride is 1, soconsecutive requests are to different bands.

There are other schemes for assigning ad-dresses to banks, all of which schemes aimto (for nonunit strides) give the same per-formance as with a stride of 1. These in-clude skewed addressing, which is similar tosequential interleaving but with a fixed dis-placement (the skew) added to the chosenmodule number; dynamic skewed address-ing, in which the skew is variable and de-termined at run-time; pseudo-random skew-ing, in which the module number is chosenon a pseudo-random basis; prime-number in-terleaving, in which the number of modulesemployed is prime relative to the degree ofinterleaving; and superinterleaving, in whichthe number of modules exceeds the degree ofinterleaving.


interleaved storage the notion of break-ing storage into multiple pieces that may beaccessed simultaneously by the same request,increasing the bandwidth from the storage.

interleaving (1) regular permutation ofsymbols usually used to split bursts of errorscaused by channels with memory (e.g., fad-ing channels) making a coding scheme de-signed for memoryless channels more effec-tive. The two main categories areblock inter-leavingandconvolutional interleaving. Theinverse process is called deinterleaving.

(2) a characteristic of a memory subsys-tem such that successive memory addressesrefer to separately controlled sections orbanks of memory. Sequential accesses to in-terleaved memory will activate the banks inturn, thus allowing higher data transfer rates.

interlock the mechanism that stalls apipeline while a result needed in the pipelineis being produced.

intermediate frequency (IF) (1) the fre-quency produced when two signals are com-bined in a mixer circuit.

(2) an operating frequency of an internalstage used in a superheterodyne system. Thisfrequency is converted from radio frequencyby using a local oscillator and a mixer forreceivers. From this frequency, radio fre-quency is converted by using a local oscil-lator and a mixer for transmitters.

intermediate-level vision the set of visualprocesses related to the perception/descrip-tion of surfaces and of their relationships ina scene.

intermittent fault a fault that appears,disappears, and then reappears in a repeatedmanner.

intermodal distortion distortion in thetemporal shape of an optical signal transmit-ted through an optical fiber caused by the dif-ferential time delays between the numerousmodes propagating in a multimode fiber.

intermodulation the heterodyning (ormixing) of at least two separate and distinctelectrical signals within a nonlinear system.For two waves mixing together, the resultantintermodulation products are given by

FIM = PF1 +QF2

whereFIM is the resultant intermodulationproduct frequency,P and Q are integers(zero, positive, or negative), and the order,R, of the intermod is given by

R =| P | + | Q |

intermodulation distortion in optical de-vices, nonlinear transfer functions that resultfrom the interaction of multiple signal com-ponents among each other. Strong signalcomponents lead to levels of distortion thatlimit the low end of a device dynamic range.

intermodulation intercept point a figureof merit for intermodulation product suppres-sion, measured in decibels.

internal bus a bus used to connect internalcomponents of a computer system, such asprocessors and memory devices.

internal forwarding a mechanism in apipeline that allows results from one pipelinestage to be sent directly back to one or morewaiting pipeline stages. The technique canreduce stalls in the pipeline.

internal fragmentation in paging, the ef-fect of unused space at the end of a page. Onaverage the last page of a program is likely tobe 50% full. Internal fragmentation increasesif the page size is increased.

internal memory main memory of a com-puter.

internal model control (IMC) a feed-back control structure in which the outputfrom a model of the process is subtractedfrom the process output to form the feed-back signal. Using this loop configuration as


Internal model control structure.

a theoretical construct, the design of closedloop controllers (which is a difficult nonlin-ear mathematical problem) becomes equiva-lent to the design of an open loop controller(which is a simpler linear problem). IMC isa special case of Q-parametrization or Youlaparametrization of all stabilizing controller,and is closely related to the Smith Predic-tor invented in the late 1950s. Unstable pro-cesses need to be stabilized first by an innerloop before application of this theory. Prac-tical designs for processes subjected to in-put disturbances should be checked for inter-nal performance.See alsoinverse responsecompensator.

internal model principle in process con-trol, one of the most common problems isregulation. This implies the importance ofconsidering the attenuation of disturbancesthat act on the process. One of the manyways of dealing with disturbances is calledinternal model principle.

Using the internal model principle, thepole placement procedure can be modifiedto take disturbances into account. In manycases, the disturbances have known charac-teristics, which can be captured by assumingthat the disturbance in the process model isgenerated by some dynamical system. Us-ing internal model principle implies that amodel of the disturbance is included in thecontroller.

internal performance defines the perfor-mance of a feedback system to stimuli thatenter the loop at internal points. Naive de-signs may produce automatic control loopsthat have good performance in the traditionalsense, yet exhibit poor internal performance.As an example, the process modeled by the

transfer function

g(s) = 4

1 + 7s

when controlled in a unity feedback config-uration by the digital control law

k(z) = 0.794(z− 0.842)

z− 1

illustrates clearly the difference betweenthe traditional concept of performance andthat of internal performance. The controlloop has good performance in the traditionalsense, as its response to a setpoint change(at timet = 0) and an output disturbance (att = 10) is much faster than the open loop sys-tem, yet its internal performance is poor sinceit rejects input disturbances sluggishly (aftert = 20). The concept is particularly impor-tant for processes subjected to input distur-bances in which the closed loop is faster ormore damped than the open loop process it-self. See alsointernal stability.

Illustration of the difference between performance

and internal performance.

internal photoemission emission of en-ergetically excited electrons or holes over aenergy barrier into its contacting collector bya process in which the emitting electrons orholes are optically excited from the groundstate of the contacting emitter.

internal quantum efficiency the prod-uct of injection efficiency and radiative effi-


ciency corresponds to the ratio of power ra-diated from the junction to electrical powersupplied.

internal stability defines the stability ofa system to signals that enter the loop at allpossible points. The difference between thetraditional concept of stability and the mod-ern concept of internal stability is illustratedby the open loop control configuration, whichis internally unstable yet might easily havebeen classified as stable under the traditionaldefinition. The relevant equations are

y = g(s)k(s)× e + g(s)× v

= 1

s + 1× e + 1

s − 1× v

where blocksg(s) andk(s) are transfer func-tions, signalse andv are inputs, signalu is aninternal variable, and signaly is the output.Clearly the relationship betweeny ande isstable, while that betweeny andv remainsunstable. The traditional definition of stabil-ity would have missed the latter phenomena.Exactly the same effect is observed when theabove system is put into closed loop. Theconcept of internal stability relates closely tothose of controllability and observability instate space theory of dynamic systems.Seealso internal performance.

Illustration of internal stability.

International Electro-technical Commis-sion (IEC) the international standards andconformity assessment body for all fields ofelectrotechnology.

International Radio Consultative Com-mittee (CCIR) one of two internationalcommittees that exist for the purpose of car-rying out studies of technical and other prob-lems related to the interworkings of their

respective national telecommunication sys-tems to provide a worldwide telecommuni-cations network. It operates under the aus-pices of the International Telecommunica-tion Union (ITU). After each plenary assem-bly, the ITU publishes recommendations thatdeal with point-to-point radio relay systems.A purpose of those recommendations is tomake the performance of such systems com-patible with metallic line systems that followthe CCITT (The International Telegraph andTelephone Consultative Committee) recom-mendations.

internet a network formed by the inter-connection of networks.

interpixel redundancy the tendency ofpixels that are near each other in time or spaceredundancy to have highly correlated graylevels or color values. Reducing interpixelredundancy is one way of compressing im-ages.See alsoimage compression.

interpolation (1) the process of finding avalue between two known values on a chartor graph.

(2) the reconstruction of missing signalsamples based on observed samples. Fre-quently, the conversion of a discrete signal toa discrete signal on a finer time step or (in thelimit) to a continuous signal.

interpolative coding coding schemes thatinvolve interpolation.

interpolative vector quantization (IRVQ)technique in which a subsample of the in-

terpolated original image is used to form thepredicted image, and then a residual imageis formed based upon this prediction. Thisapproach reduces blocking artifacts by usinga smooth prediction image.

interpole a set of small poles located mid-way between the main poles of a DC ma-chine, containing a winding connected in se-ries with the armature circuit. The interpoleimproves commutation by neutralizing the


flux distortion in the neutral plane caused byarmature reaction.

interpreter a computer program thattranslates and immediately performs in-tended operations of the source statementsof a high-level language program.

interprocess communication (IPC) thetransfer of information between two cooper-ating programs. Communication may takethe form of signal (the arrival of an event) orthe transfer of data.

interrupt an input to a processor thatsignals the occurrence of an asynchronousevent. The processor’s response to an inter-rupt is to save the current machine state andexecute a predefined subprogram. The sub-program restores the machine state on exit,and the processor continues in the originalprogram.

interrupt descriptor table used in pro-tected mode to store the gate descriptors forinterrupts and exceptions.

interrupt disable (1) a state in which in-terrupts to the CPU are held, but not pro-cessed. In most systems it is possible todisable interrupts selectively, so that certaintypes of interrupts are processed, while oth-ers are disabled and held for later processing.

(2) the operation that changes the interruptstate of the CPU from enabled to disabled.

interrupt enable (1) a state in which in-terrupts to the CPU can be processed.

(2) the operation that changes the interruptstate of the CPU from disabled to enabled.

interrupt handler a predefined subpro-gram that is executed when an interrupt oc-curs. The handler may perform input or out-put, save data, update pointers, or notify otherprocesses of the event. The handler must re-turn to the interrupted program with the ma-chine state unchanged.

interrupt line a wire carrying a signal tonotify the processor of an external event thatrequires attention.

interrupt mask a bit or a set of bits thatenables or disables the interrupt line to betransmitted at the interrupt detector circuit(inside or outside the CPU). The mechanismsof masking are typically implemented intointerrupt controllers.

interrupt priority a value or a specialsetting that specifies the precedence to servethe corresponding interrupt with respect toother interrupt signals.

interrupt service the steps that make upthe operation by which the CPU processesan interrupt. Briefly, the CPU suspends ex-ecution of its current program and branchesto a special program known as an interrupthandler or interrupt routine to take appropri-ate action. Upon completion of the interruptservice, the CPU will take one of a number ofactions, depending on circumstances: it can

1. return to the previous task if conditionspermit;

2. process other pending interrupts, or

3. request further action from the operat-ing system.

interrupt-driven I/O an input/output(I/O) scheme where the processor instructsan I/O device to handle I/O and proceeds toexecute other tasks; when the I/O is complete,the I/O device will interrupt the processorto inform completion. An interrupt-drivenI/O is more efficient than a programmed I/Owhere the processor busy-waits until the I/Ois complete.

interruptible load a load, typically ofa commercial customer, which by contractmay be interrupted by the utility for purposesof system stability.

interrupting capacity Seeinterruptingrating.


interrupting rating for a circuit breaker,fuse, or switch, the maximum fault level thatthe device can safely interrupt. The interrupt-ing rating can be expressed in terms of ampsor volt-amps.

interruption Seesustained interruption,momentary interruption, temporary interrup-tion.

intersection of fuzzy sets a logic opera-tion, corresponding to the logical AND oper-ation, forming the conjunction of two sets. Ina crisp (non-fuzzy) system, the intersectionof two sets contains the elements which arecommon to both. In fuzzy logic, the intersec-tion of two sets is the set with a membershipfunction which is the smaller of the two.

Denoted∩, fuzzy intersection is more rig-orously defined as follows. LetA and B betwo fuzzy sets in the universe of discourseX with membership functionsµA(x) andµB(x), x ∈ X. The membership functionof the intersectionA ∩ B, for all x ∈ X, is

µA∩B(x) = min{µA(x), µB(x)}

See alsofuzzy set, membership function.

intersubband laser a type of quantumwell laser in which the lasing transition cou-ples states that are accessible to a single typeof carrier, (i.e., either electrons or holes).

intersymbol interference (ISI) the dis-tortion caused by the overlap (in time) of ad-jacent symbols.

intersymbol interference postcursor in-tersymbol interference caused by the causalpart of the channel impulse responseh(t)(i.e.,h(t), t > 0). For each transmitted sym-bol, postcursor intersymbol interference iscontributed from preceding transmitted sym-bols.

intertie another name for tie line.

interval polynomial a polynomial

p(s,a) = a0 + a1s + a2s2 + · · · + ans

n

with the uncertain coefficient vector

a = [a0, a1, a2, . . . , an]

that may take values from then + 1 dimen-sional rectangle

A = {a|ai ∈ [a−

i ; a+i

], i = 0,1,2, . . . , n

}intra-cell interference interference aris-ing in a cellular communication system thatdue to other transmitters (base stations or mo-bile terminals) in the same cell.Comparewith inter-cell interference.

intra-chip optical interconnect opticalinterconnect in which the source and the de-tector are connected to electronic processingelements in a single chip.

intrafield predictor for television a tech-nique used to deal with problems caused inTV signal prediction. In particular, prob-lems due to lack of correlation because of thepresence of chrominance components. Onemethod of overcoming such difficulties is tosample the composite signal at integral multi-ples of subcarrier frequencies. Predictors canbe designed to match the bimodal spectrumof the composite signal to predict the base-band luminance and modulated chromiancewith one predictor.

intraframe image or video processingmethod that operates within a single framewithout reference to other frames.See alsointerframe.

intraframe coding image coding schemesthat are based on an intraframe restriction,that is, to separately code each image in asequence.

intramodal distortion distortion in thetemporal shape of an optical signal transmit-ted through an optical fiber caused by the


wavelength dispersion of a propagating modedue to the finite spread of wavelengths in theoptical source used to transmit the signal.

intrinsic term associated with the insideor interior. In devices and device modeling,intrinsic refers to that part of the device ormodel associated with the active semicon-ductor structures that control device opera-tion, or provide the desired functions.

intrinsic coercive force the demagnetiz-ing field required to reduce the intrinsic in-duction to zero; the x-intercept of the intrin-sic demagnetization curve. This quantity isused to gage the field required to magnetizea material and its ability to resist demagneti-zation.

intrinsic demagnetization curve the sec-ond quadrant portion of the hysteresis loopgenerated when intrinsic induction (Bi) isplotted against applied field (H ), which ismathematically related to the normal curve;most often used to determine the effects ofdemagnetizing (or magnetizing) fields.

intrinsic fiber optic sensor a fiber op-tic sensor where a property of the fiber itselfresponds to the measured parameter. Exam-ples include microbend and interferometricsensors.

intrinsic image an image onto whichphysical properties of the imaged scene (e.g.,range, orientation, reflectance) are mapped.

intrinsic impedence (1) the impedancepresented when a source is open-circuited.

(2) a characteristic parameter associatedwith a medium that is the ratio of the mag-nitudes of the transverse components of theelectric field intensity and magnetic field in-tensity for a wave propagating in a given di-rection. It has units of ohms.

intrinsic induction the vector differencebetween the magnetic induction and the ap-plied external magnetic field, or the magnetic

field established by the magnetic material it-self.

intrinsic permutation symmetry Seesymmetries of nonlinear susceptibility.

intrinsic semiconductor a semiconduc-tor material in which the number of holesequals the number of electrons.

invalid entry an entry in a table, regis-ter or module that contains data that is notconsistent with the global state. Invalidity isused in cache coherence mechanisms.

invariant subspace of generalized 2-Dmodel a linear subspaceV of the localstate spaceX that satisfies certain conditions.It is called an(E,A,B)- invariant subspaceof the generalized 2-D model

Exi+1,j+1 = A1xi+1,j + A2xi,j+1

+ B1ui+1,j + B2ui,j+1

i, j ∈ Z+ (the set of nonnegative integers) if[A1A2

]V ⊂ EV × EV + Im

[B1B2

]in product spaceX × X wherexij ∈ Rn isthe local semistate vector,uij ∈ Rm is theinput vector, andE,Ak, Bk (k = 1,2) arereal matrices. There exists a state feedbackuij = Kxij , i, j ∈ Z+ such that([

A1A2

]+[B1B2

]K

)V ⊂ EV × EV

if and only ifV is an(E,A,B)-invariant sub-space of the model.

inverse Chebychev filter filter that ex-hibits an inverse Chebychev response in thestop-band region. That is, in the stop-bandregion of the filter, the gain of the filter as afunction of frequency is equal ripple in mag-nitude. This is in contrast with many othertypes of filters, which in the filter stop-bandregion exhibit a gain response that is mono-tonic with frequency.


inverse dynamics the inverse dynamicsproblem consists of determining the jointgeneralized forcesτ(t) (acting at the jointsof the manipulator) needed to generate themotion specified by the joint generalized po-sitionq(t), velocitiesq(t), and accelerationsq(t). The inverse dynamics solution is usedin model based control algorithms for manip-ulators.

inverse dynamics linearizing control in-verse dynamics linearizing control is an oper-ational space control scheme that uses a feed-back control signal that leads to the systemof double integrators. In this case, nonlineardynamics of the manipulator is compensatedand the inverse dynamics linearizing controlallows full trajectory tracking in the opera-tional space.

inverse filter for a linear time invariant(LTI) filter, the filter which, when cascadedwith it produces white noise output for awhite noise input. The inverse filter of anLTI filter with impulse responseh(n) has im-pulse responsea(n), whereh(n) ∗ a(n) =δ(n). Alternatively, in the frequency domain,the inverse filterA(z) satisfies the relationA(z)H(z) = 1. The inverse filterA(z) issaid to whiten the filterh(n) to produceδ(n)

inverse Fourier transform SeeFouriertransform.

inverse homogeneous transformation ma-trix in general, the inverse of the homo-geneous transformation matrix describes thereference coordinate frame with respect tothe transformed frame and has the form

T −1 =

nx ny nz −p · nox oy oz −p · oax ay az −p · a0 0 1 1

where “·” denotes a scalar product betweentwo vectors.

inverse kinematics problem for a desiredposition and orientation of the end-effector

of the manipulator and the geometric linkparameters with respect to the reference co-ordinate system, calculates a set of desiredjoint variables vector. The inverse kinemat-ics problem has usually multiple solutions.In general, the inverse kinematics can besolved using algebraic, interactive, or geo-metric methods.

inverse Laplace transform SeeLaplacetransform.

inverse Nyquist array (INA) an array ofNyquist plots of the elements of the inversefrequency response matrixQ(jω) of a multi-input–multi-output transfer function modelQ(s) of an open loop multivariable sys-tem. Each diagonal graph can be viewed asa single-input–single-output inverse Nyquistdiagram from which the stability of the mul-tivariable system can be deduced, under cer-tain conditions relating to diagonal domi-nance. A typical INA diagram, for a MIMOsystem with two inputs and two outputs,clearly shows the individual polar plots andthe Gershgorin circles forming a band withcanters along the diagonal elements. Rosen-brock, H. H.,Computer-Aided Control Sys-tem Design,Academic Press, London, 1974.See alsoGershgorin circle.

inverse Nyquist plot a graphical repre-sentation of the complex function

q(jw)

which is the inverse of the frequency responsemodel of the open loop transfer function sys-temq(s) that is connected in a unity feedbackcontrol loop. The stability of the resultingclosed loop system is deduced by the princi-ple of the argument applied to this complexfunction.

inverse problem essentially, the prob-lem of inverting a “forward” system. Morespecifically, suppose we have some systemS(x) parameterized in terms ofx, and sup-pose we can define observations

y = O(S(x)).


Typical inverse Nyquist array diagram.

Then the “forward” problem is the deductionof y (or its statistics) based on knowledge ofx. The “inverse” problem is the deduction ofx (or its statistics) based on knowledge ofy(and possibly given prior statistics forx).

inverse response compensator a feed-back control configuration in which the dy-namics of the process are canceled by a cas-cade pre-compensator that includes a modelof the process. The block diagram for an in-verse response compensator is shown in thefigure. Other configurations based on the

Block diagram for an inverse response compen-

sator.

model are also possible. For example, the in-ner control loop might easily have been a statespace control law, with or without asymptoticsetpoint tracking.

inverse scattering theory yields physicalparameters of the inhomogeneous mediumfrom knowledge of the scattered intensitydistribution.

inverse system the inverse of a systemS1 that transforms input signalsx into outputsignalsy is the systemS2 that, when placedin series withS1, results in a larger cascade

that leaves the input signal unchanged (x in,x out). See alsoinvertible system.

inverse translation buffer (ITB) a deviceto translate a real address into its virtual ad-dresses, to handle synonyms. Also called areverse translation buffer and an inverse map-per.

inverse wavelet transform a computationprocedure that calculates the function usingthe coefficients of the wavelet series expan-sion of that function.

inverse-time circuit breaker a circuitbreaker in which the allowed current and timeare inversely proportional. It contains a ther-mal element and a magnetic element in se-ries. The thermal element is designed totrip as a result of heating over time in re-sponse to overload currents, while the mag-netic element is designed to trip magnetically,with no intentional time delay, in response toshort-circuit currents. Also called a thermal-magnetic circuit breaker.

inversion when a positive (negative) volt-age is applied between a conductor and a p-type (n-type) semiconductor separated by athin dielectric layer, the majority carrier holes(electrons) are repelled and minority carrierelectrons (holes) are trapped at the surface.

inversion layer an equilibrium layer ofminority carriers at a semiconductor surfaceor interface.


inversion population usually the densityof atoms or molecules in the higher state of alaser transition minus the density in the lowerstate.

inversion symmetry the study of howthe physical properties of a material changeunder the hypothetical inversion of the co-ordinate of each particle through the ori-gin of the coordinate system. See alsocentrosymmetric medium.

inversionless laser a laser that operateswithout the need for a population inversion.In general, there are two types of inversion-less lasers: those that have a hidden inver-sion in some eigenstate basis and those thatdo not. Raman and Mollow lasers can be in-cluded in the category of hidden inversion.Lasers in the category of no hidden inver-sion are possible because it is polarization,not population, that produces gain. Theselasers generally require at least three quan-tum mechanical states (for example, lambdaand vee systems) and make simultaneous useof Raman-like gain as well as induced trans-parency of the type associated with coher-ent population trapping or Fano interference.Oscillators based on parametric gain, for ex-ample optical parametric oscillators or phaseconjugate oscillators, are generally not con-sidered inversionless lasers.

inverted magnetron a radial magnetronin which the anode structure is at inside andthe cathode structure is at outside.

inverted page table the number of en-tries in a conventional page table in a virtualmemory system is the number of pages in thevirtual address space. In order to reduce thesize of the page table, some systems use aninverted page table that has only as many en-tries as there are pages in the physical mem-ory, and a hashing function to map virtual tophysical addresses in nearly constant time.

inverter (1) switching circuit that con-verts direct current to alternating current.

An inverter can be classified as a natural-commutation type or a self-commutationtype. The output AC of a natural-commutated inverter is synchronized to theAC line. This type of inverter is a thyris-tor rectifier in a reversed order. The outputAC of a self-commutated inverter is indepen-dent, i.e., both the frequency and amplitudemay be controlled. Commonly used self-commutated inverters include square-waveinverters and PWM inverters.

(2) a physical or logical gate that changesa signal from high to low, or low to high.

invertibility of generalized 2-D linear sys-tem let

T = T (z1, z2)

= C [Ez1, z2 − A0 − A1z1 − A2z2]−1

× (B0 + B1z1 + B2z2)+D

be the transfer matrix of the generalized 2-Dmodel


+ B0uij + B1ui+1,j

+ B2ui,j+1

yi,j = Cxij +Duij

i, j ∈ Z+ (the set of nonnegative integers)wherexij ∈ Rn is the local semistate vector,uij ∈ Rm is the input vector,yij ∈ Rp is theoutput vector andE,Ak,Bk (k = 0,1,2) C,D are real matrices withE possibly singular.The model is called(d1, d2)-delay invertibleif there exists a generalized 2-D model.

Ezi+1,j+1 = F0zij + F1zi+1,j + F2zi,j+1

+G0yij +G1yi+1,j

+G2yi,j+1

uij = Hzij + Jyij

with transfer matrix

TL = TL(z1, z2)

= H[Ez1, z2 − F0 − F1z1 − F2z2

]−1

× (G0 +G1z1 +G2z2)+ J


such thatTLT = z−d11 z

−d22 , wherezij ∈ Rn′

is the local semistate vector,E, Fk,Gk (k =0,1,2) H , J real matrices, andd1, d2 arenonnegative integers.

invertible system a systemS1 is invert-ible if there exists a systemS2 such that,whenS1 andS2 are placed in series, the largercascaded system leaves the input unchanged(x in, x out). For example, the multipliery(t) = k ∗ x(t) is invertible with an inversesystemy(t) = x(t)/k, while the absolutevalue functiony(t) = |x(t)| is not. See alsoinverse system.

ion beam lithography lithography per-formed by exposing resist with a focusedbeam of ions.

ion implantation a high-energy process,usually greater than 10 keV, that injects ion-ized species into a semiconductor substrate.Often done for introducing dopants for de-vice fabrication into silicon with boron, phos-phorus, or arsenic ions.

ionic transition coupling of energy levelsin an ion by means of absorption or emissionprocesses.

ionized vapor a gaseous vapor in whichthe atoms are ionized, usually by electron im-pact or by photoionization.

ionizing radiation the process in whichsufficient energy is emitted or absorbed tocause the generation of electron-hole pairs.

ionosphere consists of several layers ofthe upper atmosphere, ranging in altitudefrom 100 to 300 km, containing ionized par-ticles. The ionization is due to the effectof solar radiation, particularly ultraviolet andcosmic radiation, on gas particles. The iono-sphere serves as reflecting layers for radio(sky) waves.

IPC Seeinterprocess communication.

IPP Independent Power Producers, whoproduce electrical energy but who are notowned by utility companies.

IR Seeinfrared.

IRE (1) a unit equal to 1/140 of the peak-to-peak amplitude of a video signal, which istypically 1 V. The 0 IRE point is at blankinglevel, with the sync tip at−40 IRE and whiteextending to +100 IRE.

(2) Institute of Radio Engineers,an orga-nization preceding the IEEE, which definedthe unit IRE.

iron core transformer a transformerwhere the magnetic core is iron or principallyiron. Two classes of iron cores are iron al-loy and powdered iron. Iron alloy cores areusually restricted to low-frequency applica-tions (2 kHz or less) because of eddy cur-rent losses. Powdered iron cores are usedat higher frequencies because they consist ofsmall iron particles electrically isolated fromeach other with significantly greater resistiv-ity than laminated cores, and thus lower eddycurrent losses.

irradiated fuel nuclear fuel which hasbeen part of a nuclear reaction and is thushighly radioactive.

irreducible polynomial a polynomialf (x)with coefficients in Gallois fieldGF(q)that can not be factored into a product oflower-degree polynomials with coefficientsin GF(q).

irreversible loss a reduction in the magne-tization of a permanent magnet that can onlybe recovered by remagnetizing the material,usually caused by temperature extremes, re-versing fields, or mechanical shock.

IRVQ Seeinterpolative vector quantization.

IS-54 an interim standard of the U.S.-based Telecommunications Industries Asso-


ciation (TIA) for a digital cellular communi-cation system. One of a set of digital cellularsystems commonly classified as second gen-eration cellular systems. A standard origi-nally designed to succeed the AMPS cellu-lar standard. This interim standard has nowbeen finalized in the standard known as IS-136. The standard is based on the use of thesame channel structure as the AMPS stan-dard (30 kHz channels), similar frequencyreuse and concept of hand-off, but with theuse of a digital modulation scheme, knownasπ/4-shifted DQPSK, to transmit a digitalsignal at the rate of approximately 47 kbpson a 30 kHz channel. This signal typicallycarries three digital voice signals encoded atthe rates of 8 kbps, excluding additional over-head bits that are used for the purpose of errorcorrection.

IS-95 an interim standard of the U.S.-based Telecommunications Industries Asso-ciation (TIA) for digital cellular communi-cation systems. One of a set of systemscommonly classified as second-generationcellular systems. The standard is basedon spread spectrum modulation employingCDMA. This standard makes use of the con-cepts of power control, soft hand-offs, pi-lot signals, pilot phase offsets, orthogonalCDMA, and Walsh functions.

ISDN See integrated services digitalnetwork.

ISEF Seeinfinite symmetric exponentialfilter.

ISI Seeintersymbol interference.

ISMA Seeinhibit sense multiple accessprotocol.

ISO (1) International Standards Organi-zation.

(2) Independent System Operator, an en-tity created to operate power generation andtransmission systems but which does not ownthese facilities.

iso-dense print bias the difference be-tween the dimensions of an isolated line anda dense line (a line inside an array of equallines and spaces) holding all other parametersconstant.

isodata a special elaboration of K-meansused for clustering.

isofocal bias the difference between thedesired resist feature width and the isofocallinewidth.

isofocal linewidth the resist feature width(for a given mask width) that exhibits themaximum depth of focus.

isokeraunic map a map which denotes thevariation of lightning activity over an area bymaking use of observations of the incidenceof thunder.

isolated a power electronic circuit that hasohmic isolation between the input source andthe load circuit.

isolated buck-boost converter Seeflybackconverter.

isolated I/O an I/O system that is elec-trically isolated from where it is linked to.Isolated I/O is often found in embedded sys-tems used for control.

isolation (1) the separation of a part fromother parts of the system so that the effects ofundesired changes in the system are not seenby the separated part.

(2) a figure (usually expressed in decibels)to describe how well a transmitting device(source) and receiving device (receiver) areseparated electrically; the amount of (usuallyundesirable) signal appearing at the receiverfrom the transmitting source from an undesir-able signal path such as leakage in a coupler,mutual coupling, or multipath.

isolation switch Seebypass switch.


isolation transformer a transformer, typ-ically with a turns ratio of 1:1, designed toprovide galvanic isolation between the inputand the output.

isolator (1) a two-port passive devicebased on the Faraday effect. Signals passingin one direction suffer minimal loss, however,in the reverse direction they are strongly at-tenuated.

(2) in optics, a device inserted in an opticalfiber that prevents optical power from flowingin the reverse direction from the transmittedpower. Optical fiber isolators can be madeeither dependent or independent of the polar-ization direction of the optical energy in thefiber.

isotope a variant of an element whose nu-cleus contains a greater or lesser number ofneutrons but the same number of protons asthe base element.

isotropic antenna a fictitious referenceantenna which radiates power equally in alldirections. Such an antenna cannot be real-ized exactly in practice.

isotropic medium medium in which theconstitutive parameters are not a function ofthe vector direction of the electric or mag-netic fields. In such a medium, the consti-tutive parameters are scalar quantities, andthere is no magnetoelectric coupling.

ISS Seeimpedance standard substrate.

issue the act of initiating the performanceof an instruction (not its fetch). Issue policiesare an important design consideration in sys-tems that use parallelism and execution outof program order to achieve more speed.

ITB Seeinverse translation buffer.

iterative coordination coordination pro-cess concerned with exchange of informationbetween the coordinator and the local deci-sion (control) units, which — after a numberof iterations — leads to the local solutionsbeing satisfactory from the overall (coordi-nator) point of view; iterative coordination isused in multilevel optimization, in particularin the direct method and the price method.

iterative decoding decoding techniquethat uses past estimates to provide addi-tional information to improve future esti-mates. Used in the decoding of turbo-codes.

iterative selection a thresholding algo-rithm based on a simple iteration. First se-lect a simple threshold (say, 128) and seg-ment the image, computing the mean of theblack and white pixels along the way. Next,calculate a new threshold using the midpointbetween the black and white pixels as justcomputed. Next, repeat the entire process,using the new threshold to perform the seg-mentation. The iteration stops when the samethreshold is used on two consecutive passes.

IWP Seeindirect weld parameters.


JJacobian inverse control one of theCartesian-based control schemes. The Jaco-bian inverse control is presented in the figure.In this scheme, the end-effector location inthe operational space is compared with de-sired quantity and an error1X is computed.Assuming that deviation is small1X can betransformed into corresponding joint spacedeviation,1q, through the inverse of the ma-nipulator analytical Jacobian. The gains cor-respond to the joint stiffness, which is con-stant for each joint.See alsoCartesian-basedcontrol.

Jacobian inverse control scheme.

Jacobian transpose control scheme.

Jacobian of the manipulator a matrixthat maps the joint velocities into end effectorvelocities.

Jacobian transpose from a mathemati-cal point of view, the transpose of the Ja-cobian. Jacobian transpose can be used ininverse kinematics algorithm with operationspace error as a feedback.

Jacobian transpose control depicted inthe figure. In this case the operationalspace error,1X, is calculated first. Out-put of the gains block,F , can be treatedas the elastic force generated by a general-ized spring. This operational space force

is transformed into the joint space general-ized forces,τ , through the transpose of theJacobian.See alsoCartesian-based control,Jacobian inverse control.

JANTX a prefix denoting that the mili-tary specification device has received extrascreening and testing, such as a 100% 168-hburn-in.

JANTXV a JANTX part with an addedprecapsulation visual requirement.

Jensen’s inequality for a functionf , con-vex on the supportX and a random variableX ∈ X ,

Ef (x) ≤ f (EX) ,

provided the expectation exists. Named afterits discoverer, Danish mathematician JohanL. W. V. Jensen (1859–1925).

jet a gas turbine peaking generatorSeepeaking unit.

JFET Seejunction field-effect transistor.

JFET oscillator device often used as theactive element in any of the LC-oscillators.A special attention deserves using JFET asthe active element of the tuned-gate oscilla-tor. In this configuration, the tuned circuitis not loaded by the FET, its loading dueto the transistor output impedance is excep-tionally small, and the Miller effect is alsominimized. This oscillator can be clamp-biased, and the oscillation amplitude may bestabilized with the FET operating completelywithin its square-law region.

jitter a signal sample is temporarily dis-placed by an unknown, usually small, inter-val.

JK flip-flop device that uses two inputs(J andK) to control the state of itsQ andQ′ outputs. A negative edge-sensitive clockinput samplesJ andK and changes the out-


puts accordingly. The simplified truthtable for theJK flip flop is as follows:

J K Q Q′0 0 0 00 0 1 10 1 0 00 1 1 01 0 0 11 0 1 11 1 0 11 1 1 0

Johnson noise Seethermal noise.

joint a lower-pair joint connects two linksand it has two surfaces sliding over one an-other while remaining in contact. Six dif-ferent lower-pair joints are possible: revo-lute (rotary), prismatic (sliding), cylindrical,spherical, screw, and planar. Of these, onlyrotary and prismatic joints are common inmanipulators. Each has one degree of free-dom.

joint detection Seemulti-user detection.

Joint Electron Device Engineering Coun-cil (JEDEC) a part of the EIA.

joint flexibility small deformations char-acterized as being concentrated at the jointsof the manipulator. We often refer to this sit-uation by the term “elastic joints” in lieu offlexible joints. As an example, joint flexibil-ity can be a dynamic, time-varying displace-ment that is introduced between the positionof the driving actuator and that of the drivenlink. This is due to presence of transmissionelements such as harmonic drives transmis-sion, belts, and long shafts.

joint offset the relative displacement be-tween two successive links in a robotic arm.

Joint Photographic Experts Group (JPEG)(1) group that standardizes methods for

still picture compression. Baseline JPEGcompression uses an 8× 8 discrete cosine

transform, quantization and zigzag order-ing of transform coefficients, combined run-length/value coding, and Huffman or Arith-metic lossless coding.

(2) a still picture compression method thatuses the spectral components of an image toperform compression. The JPEG baselinemust be present in all JPEG modes of opera-tion which use DCT.See alsoJPEG.

joint source-channel coding Seecombinedsource-channel coding.

joint space a position and orientation of amanipulator ofn degrees of freedom is spec-ified as a result of the forward kinematicsproblem. This set of variables is called asthen× 1 vector. The space of all such jointvectors is referred to as joint space.

joint space control depicted in the fig-ure. The joint space control scheme consistsof the trajectory conversion, which recalcu-lates the desired trajectory from the opera-tional space(Xd, Xd, Xd) into the joint space(qd, qd , qd ). Then, a joint space controlleris designed that allows tracking of the refer-ence inputs. Notice that only joint inputs arecompared with joint positions, velocities, andaccelerations. The joint space control doesnot influence the operational space variableswhich are controlled in an open-loop fashionthrough the manipulator mechanical struc-ture. Joint space control is usually a single-input–single-output system; therefore, ma-nipulator withn degrees of freedom hasnsuch independent systems. A single-input–single-output system has PI (proportional-integral), PD (proportional-derivative), orPID (proportional-integral-derivative) struc-ture. These schemes do not include dynam-ics of the manipulator. Dynamics of the ma-nipulator can be included in, for example,computed torque control.See alsocomputertorque control.

joint stiffness Seestiffness of a manipulatorarm.


Joint space control scheme.

Joint Tactical Information DistributionSystem (JTIDS) system that uses spreadspectrum techniques for secure digital com-munications; used for military applications.

joint time frequency analysis (JTFA)techniques the aim of which is to repre-sent and characterize a signal in time andfrequency domain simultaneously by usingdifferent kinds of transformation and kernelfunctions.

joint transform correlator a type of opti-cal correlator that employs two parallel paths,one for each signal, instead of an in-line cas-cade.

joystick an input device in the form of acontrol lever that transmits its movement intwo dimensions to a computer. Joysticks areoften used in games for control. They mayalso have a number of buttons whose statecan be read by the computer.

JPEG SeeJoint Photographic ExpertsGroup.

JPEG baseline See Joint PhotographicExperts Group baseline.

JPEG DCT there are two DCT modes ofJPEG, the sequential DCT mode and the pro-gressive DCT-based mode. In the sequentialmode the image components are coded indi-vidually or in blocks by a single scan in onepass. In the progressive mode several scansare taken to code parts of the quantized DCTcoefficients.

JTFA Seejoint time frequency analysis.

JTIDS SeeJoint Tactical InformationDistribution System.

jump instruction an instruction thatcauses an unconditional transfer of control toa different instruction sequence in memory.

jump linear quadratic problem opti-mal control problem in which a controlledprocess is modeled by linear system withMarkov jumps and a control objective isto minimize an average quadratic criteriongiven by

J = E{∫ T

0[x′(t)Q(ξ(t))x(t)

+ u′(t)R(ξ(t))u(t)]dt |x(0), ξ(0)}where x(t), u(t), ξ(t) denote, respectively,process state, process control, and mode,and matricesR,Q are real valued, symmet-ric respectively positive definite and positivesemidefinite weighting matrices of respectivedimensionsT is finite or infinite control hori-zon andE is an averaging operator. In the fi-nite time case the optimal control law is givenas

u(t) = p(x(t), ξ(t), t) = −P(ξ(t), t)x(t)

for eachξ(t) taking value in finite setS ={1,2, . . . , s} where

P(ξ, t) = R(ξ)−1B ′(ξ)K(ξ, t)x(t)

where for eachξ(t) = i matricesK(i, t) arethe unique positive semidefinite solutions ofdifferential coupled Riccati equations. As-suming that the system is stochastically sta-bilizable and for eachξ(t) = i the pairs(A(i),

√Q(i)) are observable, then for in-

finite control intervalT → ∞ the solutionof the jump linear quadratic (JLQ) problemis given by the optimal steady state control:

u(t) = p(x(t), ξ(t))

= −P (ξ(t))x(t)= −R(ξ)−1B ′(ξ)K(ξ)x(t)

where for eachξ(t) = i matricesK(i) arethe unique positive semidefinite solutions ofalgebraic coupled Riccati equations. This so-lution may be also found as the steady-state


value ofK(i, t). See alsolinear systems withMarkov jumps.

jumper a plug or wire used for settingthe configuration of system. It can be usedfor changing the hardware configuration byforcing some line to be high or low. It is usedto change software configuration (especiallyon embedded systems) when the status of thejumper is read by the microprocessor.

jumping jack a voltmeter.

junction capacitance change in chargeof immobile ions in the depletion region ofa diode corresponding to a change in reversebias voltage on a diode.

junction field-effect transistor (JFET) atype of FET in which the high input resistance

at the gate is achieved by use of a reversebiased p-n junction between the gate and thedrain-source channel.

junction-to-case thermal resistance theproportionality constant at thermal equilib-rium between the temperature difference ofthe bipolar device junction (Tjunction) and aspecified case surface (Tcase) to the dissipatedpower in the junction (pw), in units ofC◦/W .The specified surface is usually the most ef-fective in reducing the temperature. It in-cludes the thermal resistance of the chip, dieattach material (solder or adhesive), packag-ing and mounting medium, as applicable.

θjc = Tjunction − Tcase

Pw

∣∣∣∣equilibrium


KK (1) symbol for the Linville stability fac-tor, a dimensionless quantity.

(2) abbreviation for 1024 (not for 1000).

K-factor rating an indication of a trans-former’s capability to provide a specifiedamount of harmonic content in its load cur-rent without overheating. K-factor ratedtransformers typically use thinner lamina-tions in the core to reduce the eddy currentlosses, larger conductors to reduce skin ef-fect, and for three-phase transformers a largerneutral conductor to conduct zero-sequenceharmonic currents.

K-means algorithm (1) a clustering al-gorithm known from the statistical literature(E. W. Forgy, 1965), that relies on the sameprinciples as the Lloyd algorithm and the gen-eralized Lloyd algorithm.

Given a set of patterns andK, the numberof desired clusters, K-means returns the cen-troid of theK clusters. In a Pascal-like, thealgorithm runs as follows:

function K-Means(Patterns:set-of-pattern):Centroids;

begin repeatfor h = 1 to N doAssign point h to class k for which the distance(xh −

µk)2 is a minimum

For k = 1 to K doComputeµ[k] (average of the assigned points)

until no further change in the assignment

end

The algorithm requires an initializationthat consists of providingK initial points,which can be chosen in many different ways.Each of theN given points is assigned toone of theK clusters according to the Eu-clidean minimum distance criterion. The av-erage in the clusters is computed and the al-gorithm runs until no further reassignment ofthe points to different clusters occurs.

(2) statistical clustering algorithm.N ,m-component training vectorsv1 . . . vN are tobe approximated byk prototypesw1 . . . wk,for example, in vector quantization. Afterinitialization of thewi , randomly, uniformlyor in some other way, the algorithm iteratesthe following two steps: 1. associatevi withprototypes according to a closest distance cri-terion (e.g., minimum euclidean distance);2. find the mean vector of all thevi associatedwith a particular prototypewj , and move thatprototype to the mean point. Iteration ceaseswhen there is no more movement of proto-types or according to some other stoppingcondition.

K-nearest neighbor algorithm a methodof classifying samples in which a sample isassigned to the class which is most repre-sented among thek nearest neighbors; an ex-tension of the nearest neighbor algorithm.

Kaczmarz’s algorithm the recursiveleast-squares algorithm has two sets of statevariables, parameters vector and covariancematrix, which must be updated at each step.For large dimension, the updating of the co-variance matrix dominates the computing ef-fort. Kaczmarz’s projection algorithm is onesimple solution that avoids updating the ma-trix at the cost of slower convergence. Theupdating formula of the least squares algo-rithm using the Kaczmarz’s algorithm has theform

θ (t) = θ (t − 1)

+ φ(t)

φT (t)φ(t)(y(t)− φT (t)θ (t − 1))

whereθ is parameter’s estimates vector,φ isregressor vector. This approach is also calledthe normalized projection algorithm.

Kaiser window a function defined by

w[n] ={I0[β(1−[(n−α)/α]2) 1

2 ]I0(β)

if 0 ≤ n ≤ M

0 otherwise


whereM is the length of the filter minus one,α = M

2 andβ is a term that depends on thedesired ripple level of the filter.

Kalman decomposition of linear systemsone of several decompositions for linear sys-tems.

Every linear, stationary finite-dimensionalcontinuous-time dynamical system can bedecomposed into four subsystems: control-lable and observable, controllable and un-observable, uncontrollable and observable,uncontrollable and unobservable. This de-composition does not depend on any non-singular transformations in the state spaceRn. The transfer function matrix always de-scribes only the controllable and observablepart of the dynamical system and does not de-pend on any nonsingular transformations inthe state spaceRn. Similar statements holdtrue for linear stationary finite-dimensionaldiscrete-time dynamical systems.

Kalman filter the method of recursivelyestimating the state vector of a linear dynamicsystem based on noisy output measurements.This method combines the knowledge of sys-tem parameters, statistical characteristics ofinitial state, process and measurement noises,and the current measurement to update theexisting estimate of the state. Kalman filteris the minimum variance filter if the initialstate and the noises are Gaussian distributed;otherwise, it is the linear minimum variancefilter. In addition to providing the expectedvalue of the state conditional on the measure-ment, this filter generates the estimation errorcovariance, which signifies the uncertaintyassociated with this estimate. This covari-ance can be calculated off-line before runningthe filter.

Kanerva memory a sparse distributedmemory developed for the storage of high-dimensional pattern vectors. Memory ad-dresses are randomly generated patterns, anda given pattern is stored at the address that isclosest in terms of Hamming distance. Pat-terns are usually bipolar, and those stored at

the same address are added together. Suit-able thresholding of this sum allows a closeapproximation to individual patterns to be re-called, so long as the memory is not over-loaded.

Karhunen–Loeve transform (KLT) anoptimal image transform in an energy-packing sense.

If the images are transformed byy = Ax,then the corresponding quadratic term asso-ciated with the covariance matrix becomesx′A′RAx. Hence the image uncorrelation isreduced to findingA such that

A′RA = 3 = diag(λ1, . . . , λn) .

This equation can be solved by finding inthe following three steps.

1. Solve equation|R − λI | = 0 (find theR’s eigenvalues).

2. Determine then solutions of (R −λi I )ai = 0 (eigenvectors).

3. Create the desired transformationA asfollows:

A = [a1, . . . ,an] .

In this transform a limited number oftransform coefficients are retained. Thesecoefficients contain a larger fraction of thetotal energy in the image. This transform isheavily dependent on the image features andrequires a covariance function estimate forperforming the transform.See alsodiscretecosine transform.

Karnaugh map a mapping of a truth tableinto a rectangular array of cells in which thenearest neighbors of any cell differ from thatcell by exactly one binary input variable.

KCL SeeKirchoff’s current law.

kcmil Seecircular mil.

KDP Seepotassium dihydrogen phosphate.


kernel (1) the term for the portion of theoperating system that operates in privilegedmode.

(2) the set of basis functions for a trans-formation.

(3) the set of convolution weights for fil-tering.

Kerr coefficients describes the quadraticelectro-optic effect, particularly for gases andliquids.

Kerr effect the quadratic electro-optic ef-fect, particularly in the case for gases andliquids.

key (1) in a table, the value used to selectthe desired entry (or entries).

(2) in an access control system, a valueheld by a process to permit it to make accessto certain objects within the system.

key point detection a technique usuallyemployed in specialized linear or morpho-logical filters designed for measuring gray-level changes in several directions. A keypoint is an isolated image point correspond-ing to a peculiar physical or geometrical phe-nomenon in the scene from which the imagearises; it can be for example a corner, a linetermination, a junction, a bright or dark spot,etc. Key points are distinguished from edgesby two properties: they are sparse and displaystrong gray-level variations in two or more di-rections (while edge points are grouped intolines, and have gray-level variations essen-tially in the direction normal to that line).Seeedge. Seealso salient feature.

key punch a device with a keyboard usedfor storing data in paper cards or paper tapesby punching holes. The pattern of holespunched across these cards or tapes representthe data keyed in. Now obsolete.

keyboard an input device with a set of but-tons (or keys) through which characters areinput to a computer. In addition to the keysfor inputing characters, a keyboard may also

have function keys and special keys, such aspower-on or print-screen.

keyboard controller the device controllerthat processes keyboard input. Because of itsimportance, I/O from the keyboard controlleris often handled differently from other I/Oprocesses, with its own direct connection tothe CPU.

keystone distortion a distortion thatpresents an image in the shape of a trape-zoidal. For projection television displays,off-axis projections of the red and blue tubescan cause keystone distortion. For directview television displays, unequal deflectionsensitivities for the two sections of the deflec-tion yoke can result in keystone distortion. Ahorizontal trapezoid is the result of the verti-cal yoke deflection sensitivities. Similarly, avertical trapezoid is the result of the horizon-tal yoke deflection sensitivities.

KGD Seeknown good die.

Kharitonov theorem the most popularnecessary and sufficient condition of robuststability for characteristic interval polynomi-als. It states that robust stability of uncertainlinear time-invariant system with character-istic interval polynomialp(s,a) is stable ifand only if the following four polynomialshave all roots endowed with negative parts:

p1(s) = a+0 +a−

1 s+a−2 s

2+a+3 s

3+a+4 s

4+...

p2(s) = a+0 +a+

1 s+a−2 s

2+a−3 s

3+a+4 s

4+...

p3(s) = a−0 +a+

1 s+a+2 s

2+a−3 s

3+a−4 s

4+...

p4(s) = a−0 +a−

1 s+a+2 s

2+a+3 s

3+a−4 s

4+...

wherea−i , a

+i are bounds for coefficients of

the characteristic interval polynomial. Thepolynomials are called Kharitonov polyno-mials, and the required property could bechecked by four Routh–Hurwitz tests. Somegeneralizations of this result for dependent


coefficient perturbations and more generalzero location regions are known.

Kilby, Jack St. Clair (1923– ) Born:Jefferson City, Missouri, U.S.A.

Kilby is best known as the person whofirst suggested and then implemented theconcept of integrating transistors, with ca-pacitors and resistors, on a single siliconwafer. This concept, simple-sounding to-day, formed the basis for the first early in-tegrated circuits. Kilby’s employer, TexasInstruments, launched a patent suit againstFairchild Industries to establish their andKilby’s claim to the technology. The courtruled in favor of Robert Noyce and Fairchild.Despite this, Kilby is universally recognizedfor his pioneering work, especially in thepractical implementation of IC technology.

kil ovar See VAR.

kilovolt-ampere (KVA) a measure of ap-parent power, often in the rating of a pieceof equipment or the measure of an electricalload, which is obtained by multiplying thedevice voltage in kilovolts by the current inamperes.

kilowatt-hour (KWH) a measure of elec-trical energy: 1000 watts delivered for onehour.

kinematic calibration a procedure tofindingaccurateestimatesof Denavit–Harten-berg parametersfromaseriesofmeasurementsof themanipulator’send-effector locations.See also Denavit–Hartenberg notation.

kinematic singularity a point for whichthe geometric Jacobian is not invertible. Inother words, those configurations at whichJacobian is rank-deficient are termed kine-matic singularities. As a consequence, in sin-gular configurations infinite solutions to theinverse kinematics problem may exist. Frompractical point of view small velocities in theoperation space will cause large velocities inthe joint space.

kinematically simple manipulator a ma-nipulator where some or all link offset anglesequal 0,−90◦, or 90◦ and link distances andoffsets equal zero. Kinematically simple ma-nipulators are found in most of the industrialrobots. For these robots, forward kinematicsproblem can be solved in a closed form solu-tion and dynamics equations are the simplest.See alsoDenavit–Hartenberg notation.

kinetic energy energy of motion.

kinetic energy conservation principleany change of kinetic energy of an electronis transformed into a change in electric po-tential energy with which the electron is in-teracting, and vice versa.

kineto-static duality the static relation-ship combined with the differential kinemat-ics equation defines a kineto-statics duality.For the description of differential kinematics.Seedifferential kinematics. The static rela-tionship describes the end point force andequivalent joint generalized force balance as-suming that we neglect gravity and frictionat the joints. This relationship is describedby the transpose of the manipulator Jacobianand has the formτ = J T F , whereF is avector of exerted end point forces andτ is avector of joint generalized forces. As a con-sequence of the last definition, the followingproperties can be defined.

1.The range ofJ T is the subspaceR(JT )in Rn of the joint forces and/or torques thatcan balance the end-effector forces, in thegiven manipulator posture.

2.The null ofJ T is the subspaceN(JT ) inRT of the end-effector forces and/or torquesthat do not require any balancing joint gener-alized forces, in the given manipulator pos-ture.

Kirchhoff, Gustav Robert (1824–1887)Born: Konigsberg, Germany

Kirchhoff is best known for discoveringthe laws that govern electric flow in networks,known now as Kirchhoff’s laws. Kirchhoff isalso famous as the inventor of spectroscopy,


and his theoretical work led to the quan-tum theory of matter. Kirchhoff first held ateaching post in Berlin, followed by appoint-ments in Breslau and Heidelberg. In the lattertwo universities, Kirchhoff had very fruitfulcollaborations with Robert Bunsen. Kirch-hoff’s studies in electricity provided a foun-dation for Maxwell and Lorenz’s descriptionof electromagnetic theory.

Kirchoff’s current law (KCL) a funda-mental law of electricity that states that thesum of the currents entering and exiting a cir-cuit node must be equal to 0.

Kirchoff’s laws laws that govern the rela-tionships between voltages/currents in a cir-cuit/network. See alsoKirchoff’s voltagelaw, Kirchoff’s current law.

Kirchoff’s voltage law (KVL) a funda-mental law of electricity that states that thesum of the voltage drops and rises in a closedloop must equal 0.

Kleinman symmetry Seesymmetries ofnonlinear susceptibility.

KLT SeeKarhunen–Loeve transform.

klydonograph a measuring instrumentsfor high-voltage impulses which makes useof Lichtenburg figures impressed on photo-graphic film.

klystrode an amplifier device for UHF-TV signals that combines aspects of a tetrode(grid modulation) with a klystron (velocitymodulation of an electron beam). The resultis a more efficient, less expensive device formany applications. (Klystrode is a trademarkof EIMAC, a division of Varian Associates.)

knife-edge diffraction classical diffrac-tion model defined by the interaction ofa propagating electromagnetic wave with aperfectly conducting, infinitely thin obstruct-ing boundary. Knife-edge diffraction is often

used is terrain propagation calculations as anapproximation for hills and ridges.

knob-and-tube wiring a form of residen-tial wiring, now obsolete, in which lightly-insulated wires are supported on porce-lain insulators (knobs) or porcelain bushings(tubes) through joists or studs.

knowledge engineering the process ofdeveloping an expert system.

Knowlton’s technique a pyramidal typeof hierarchical approach, in which a re-versible transformation takes adjacent pairsof k-bit pixel values and maps it into a k-bitcomposite value and a k-bit differentiator.

known good die (KGD) bare silicon chips(die) tested to some known level.

Kogelnik transformation See ABCDlaw.

Kohler illumination a method of illu-minating the mask in a projection imagingsystem whereby a condenser lens forms animage of the illumination source at the en-trance pupil (entrance aperture) of the objec-tive lens, and the mask is at the exit pupil ofthe condenser lens.

Kohonen network a 2-dimensional arrayof neurons with each neuron connected to thefull set of network inputs. Learning is unsu-pervised: the neuron whose vector of inputweights is closest to an input vector adjustsits weights so as to move closer to that inputvector; the neuron’s neighbors in the arrayadjust their input weights similarly. As a re-sult, clusters of weights in different parts ofthe array tend to respond to different types ofinput.

Kolmogorov complexity the minimumlength description of a binary string thatwould enable a universal computer to recon-struct the string.


Kraft inequality a theorem from infor-mation theory that sets a restriction on instan-taneous codes (codes where no codeword isa prefix of any other codeword, i.e., a codecontaining 0 and 01 is not an instantaneouscode). The Kraft inequality states that for aninstantaneous code over an alphabet with sizeD and codeword lengthsl1, l2, l3, . . . , lm thefollowing must be true:

m∑i=1

D−li <= 1.

Kramer drive an electric drive systemin which the output of the driven frequencyconverter is fed to the slip rings of the woundrotor. Unlike its predecessor Leblanc system,the variable transformer is connected to thewound rotor slip rings instead of to the line.

Kramer’s generalization a sampling the-ory based on other than Fourier transformsand frequency.

Kramers–Kronig relations relates thereal and imaginary components of the indexof refraction of a medium.See alsoHilberttransform.

Kronecker delta function The discrete-time unit impulse function defined asδij = 1for i = j , δij = 0 for i 6= j . See alsodeltafunction.

ku-band frequency band of approxi-mately 11–12 GHz.

Kubo formula a fundamental relation-ship, developed by Ryogo Kubo, between,e.g., the conductance and the current fluctu-ations. The Kubo formula is an example ofthe fluctuation–dissipation theorem in statis-tical physics, in which the average value ofthe current (conductance) is determined byan integral over the correlation function ofthe current fluctuations.

Kullback–Liebler distance Seerelativeentropy.

Kuroda’s identities four identities areused to achieve practical microwave filter im-plementation by using redundant transmis-sion line sections.

kVA Seekilovolt-ampere.

Kvhler illumination a method of illu-minating the mask in a projection imagingsystem whereby a condenser lens forms animage of the illumination source at the en-trance pupil (entrance aperture) of the objec-tive lens, and the mask is at the exit pupil ofthe condenser lens.

KVL SeeKirchoff’s voltage law.

KWH Seekilowatt-hour


LL-band frequency band of approximately1–2 GHz.

L-L Seeline to line fault.

label a tag in a programming language(usually assembly language, also legal in C)that marks an instruction or statement as apossible target for a jump or branch.

labeling (1) the computational problemof assigning labels consistently to objects orobject components (segments) appearing inan image.

(2) a technique by which each pixel withina distinct segment is marked as belonging tothat segment. One way to label an imageinvolves appending to each pixel of an im-age the label number or index of its segment.Another way is to specify the closed contourof each segment and to use a contour fillingtechnique to label each pixel within a contour.

ladder diagram (1) the connection of thecoils and contacts used in a control circuitshown one line after the other that looks likea ladder.

(2) a visual language for specifying theBoolean expressions, which are the core ofthe control law of PLC.

laddertron a microwave vacuum tube os-cillator with a slow-wave structure coupledto a single-cavity resonator.

lag the inability of an imaging tube to re-spond to instantaneous changes in light. Formeasurement purposes, lag has two compo-nents: rise lag is the response time from darkto light, whereas decay lag is the responsetime from light to dark. Lag is a very short-

term effect, and should not be confused withimage retention, image burn, or sticking.

lag circuit a simple passive electronic cir-cuit designed to add a dominant pole to com-pensate the performance of a given system. Alag circuit is generally used to make a systemmore stable by reducing its high-frequencygain and/or to improve its position, veloc-ity, or acceleration error by increasing thelow frequency gain. A nondominant zero isincluded in the lag circuit to prevent unduedestabilization of the compensated system bythe additional pole.

lag network a network where the phaseangle associated with the input–output trans-fer function is always negative, or lagging.

lag-lead network the phase shift versusfrequency curve in a phase lag-lead networkis negative, or lagging, for low frequenciesand positive, or leading, for high frequencies.The phase angle associated with the input–output transfer function is always positive orleading.

Lagrange formulation a formulationwhere the equations of motion are derived in asystematic way by choosing a set of general-ized coordinates, forming the Lagrangian ofthe mechanical system (as a difference of to-tal kinetic energy and potential energy of thesystem) and by solving the Lagrange equa-tions

d

dt

∂L∂qi

− ∂L∂qi

= τi i = 1, . . . , n

whereL stands for Lagrangian,qi is the gen-eralized coordinate,qi is its derivative, andτi is a generalized force, andn denotes num-ber of degrees of freedom of the mechani-cal system. Last equations establish the rela-tions existing between the generalized forcesapplied to the manipulator and the joint po-sitions, velocities, and accelerations in socalled closed form.See alsoNewton–Eulerrecursive algorithm.


Lagrange stable state Seebounded state.

Lagrangian interpolation a classic inter-polation procedure used in numerical analy-sis. The sampling theorem is a special case.

Laguerre polynomial a solution to thedifferential equationxy′′ + (1 − x)y′ +ny = 0. Laguerre polynomialsL0(x) =1, L1(x) = 1 − x, L2(x) = 1 − 2x + x2/2,andL3 = 1−3x+3x2/2−x3/6. AdditionalLaguerre polynomials may be obtained fromthe recursion formula(n+1)Ln+1(x)−(2n+1 − x)Ln(x)+ Ln−1(x) = 0.

Laguerre–Gaussian beam electromag-netic beam solution of the paraxial waveequation in which the field is a product ofa Laguerre polynomial and a Gaussian func-tion of distance from the beam axis.

Lamb dip decrease in output power of aDoppler-broadened standing-wave laser os-cillator as a function of length tuning whenthe resonant frequency is within approxi-mately one homogeneous linewidth of gaincenter; results from the interaction of boththe right and left travelling waves with thesame atoms for line-center tuning.

lambda system a 3-level system in whichthe lowest two energy states are coupled byelectromagnetic fields to a common interme-diate state of higher energy. This system isso named because schematic representationsof it often look like the capital Greek letterlambda,3.

Lambert’s cosine law a law stating that,for a ideal matte (Lambertian) surface, theapparent brightness of the surface is propor-tional to the cosine of the angle of incidenceand independent of both the angle of reflec-tion and the phase angle between the incidentand reflected beams.

Lambertian source a source whose di-rectional emission pattern follows Lambert’slaw; a cosine variation.

Lambertian surface a surface with per-fect diffusion properties, i.e., for which thereflectance function depends only on the an-gle of incidence of illumination.

laminate multi-chip module (MCM-L)a multi-chip module built using advancedPCB manufacturing techniques.

lamination a thin sheet of metal used tobuild up the core of an electromagnetic de-vice. Laminations are insulated from eachother to reduce the losses associated witheddy currents.

land pattern a combination of lands in-tended for the mounting, interconnection,and testing of a particular component.

Landauer formula describes the con-ductance as a fundamental property of wave(electron) transmission through a structure.

Lange coupler four coupled lines usedwith interconnections to provide tight cou-pling. A practical implementation to increasethe coupling between edge-coupled lines byusing several lines parallel to each other, sothat the fringing fields at edges of the linecontribute to coupling.

Langevin, Paul (1872–1946) Born:Paris, France

Langevin is best known as the developer ofecholocation, which is the precursor to mod-ern sonar. Langevin was the first to describeparamagnetism and diamagnetism.

LAOS Seelight-amplifying optical switch.

lap winding an armature winding on a DCmachine in which the two ends of each coilare connected to adjacent bars on the com-mutator ring. The lap winding provides “P ”parallel paths through the armature winding,whereP is the number of poles in the ma-chine.


Laplace’s equation a partial differen-tial equation mathematically described by∇2φ = 0, where∇2 is the Laplacian andφ is the equation’s solution.

Laplace, Pierre-Simon, Marquis de(1749–1827) Born: Beaumont-en-Auge,Normandy, France

Laplace is best known for his develop-ment of basic tools of mathematical anal-ysis including the Laplace transform, theLaplace theorem, and the Laplace coeffi-cients. Laplace studied in Paris with the greatmathematician Jean d’Alembert. Laplacewas heavily involved in politics throughouthis career and held many government posts.Laplace’s theoretical work was heavily in thefield of celestial mechanics. He helped to es-tablish the mathematical basis for the fieldand in doing so confirmed significant partsof Newton’s work.

Laplace transform the transform of afunctionf (t) given by

F(s) =∫ ∞

−∞f (t)e−st dt

wheres = a+jω is a complex variable. Theone sided or unilateral Laplace transform isgiven by the same equation except that thelower limit is 0 and not−∞. The regionof convergence of the Laplace integral is avertical stripR in the s-plane. The inverseLaplace transform is given by

f (t) = 1/(2πj)∫L

F(s)est ds

whereL is a vertical line in R. The FourierTransform off (t) is given byF(jω).

Laplacian operator the second-order op-erator, defined inRn as ∇2 = ∂2/∂x2

1 +· · · + ∂2/∂2

n . The zero crossings of an im-age to which the Laplacian operator has beenapplied usually correspond to edges, as insuch points a peak (trough) of the first deriva-tive components can be found. Also simplycalled the Laplacian.

Laplacian pyramid a set of Laplacian im-ages at multiple scales used in pyramid cod-ing. An input imageG1 is Gaussian lowpassfiltered and downsampled to formG2. Typi-callyG2 is one quarter the size ofG1, i.e., itis downsampled by a factor of 2 in each direc-tion. G2 is upsampled and Gaussian lowpassfiltered to formR1 which is then subtractedfromG1 to giveL1. The process then repeatsusingG2 as input. The sets of multiresolu-tion images so generated are called “pyra-mids”: G1 . . .Gn form a Gaussian pyramid;L1 . . .Ln form a Laplacian pyramid.

lapped orthogonal transform (LOT) acritically sampled block transform, where theblocks overlap, typically by half a block.Equivalently the LOT is a critically sampledfilter bank, where typically the filter lengthsare equal to twice the number of channels orfilters. The LOT was motivated by reducingthe blocking effect in transform coding by us-ing overlapping blocks. A cosine modulatedfilter bank is a type of LOT.

large cell cell with the radius of 5–35 km(such as those found in Groupe Special Mo-bile systems).See alsocell.

large disturbance a disturbance forwhich the equation for dynamic operationcannot be linearized for analysis.

large-scale integration (LSI) (1) termusually used to describe the level of integra-tion at which entire integrated circuits can beplaced on a single chip.

(2) an integrated circuit made of hundredsto thousands of transistors.

large-scale process (system) partitionedcomplex process (system) composed of sev-eral sub-processes (subsystems) that are ei-ther physically interconnected or must beconsidered jointly due to the nature of thecontrol objectives.

LASCR See light-activated siliconcontrolled rectifier


laser acronym that stands for light ampli-fication by stimulated emission of radiation.Usually refers to an oscillator rather than anamplifier; commonly also refers to similarsystems that operate at non-optical frequen-cies or with nonelectromagnetic wave fields.

laser amplifier usually refers to a mediumthat amplifies light by the process of stimu-lated emission; sometimes refers to amplifi-cation of some other field (nonoptical elec-tromagnetic, phonon, exciton, neutrino, etc.)or some other process (nonlinear optics, Bril-louin scattering, Raman scattering, etc.).

laser array systematic distribution oflasers intended to provide more power than asingle laser.

laser beam localized electromagneticfield distribution produced by a laser.

laser efficiency output power from a laserdivided by the input power (sometimes thepump power into the laser medium and some-times the wall-plug power).

laser medium the material in a laser thatemits light; it may be a gas, solid, or liquid.

laser oscillator oscillator usually produc-ing an optical frequency output and usuallybased on amplification by stimulated emis-sion in a resonant cavity.

laser pumping mechanism for obtaininga population inversion in a laser medium; theuse of a laser beam to pump another laser.

laser threshold the condition under whichthe round-trip gain in a laser is equal to theround-trip loss.

laser transient a time-dependent laser be-havior such as mode-locking, loss switch-ing, spontaneous pulsations, relaxation os-cillations.

laser transition transition in a mediumthat has the capability of exhibiting morestimulated emission than absorption or spon-taneous emission.

last-in-first-out (LIFO) Seefirst-in-last-out.

latch a small temporary holding cell for avalue, the value on the input wires is bufferedupon occurrence of some event, such as aclock pulse or rising edge of a separate latchsignal.

latency (1) total time taken for a bit topass through the network from origin to des-tination.

(2) the time between positioning a read/writehead over a track of data and when the be-ginning of the track of data passes under thehead.

lateral (1) a lateral on a primary distribu-tion line is a short tap from the main distri-bution line which serves a local set of loads.Single phase laterals are common in residen-tial districts.

(2) a three-phase or single-phase powerline which supplies the distribution trans-formers along a street.Seefeeder.

lateral inhibition in the human visualsystem, the inhibitory effect between nearbycells which acts to enhance changes (tempo-ral or spatial) in the stimulus.

lateral superlattice refers to a lithograph-ically defined structure in which a periodic(superlattice) potential is induced onto thesurface of a normal semiconductor (or metal-lic) system. Since the periodic potential is in-duced in the lateral variations of the surface,it is called a lateral superlattice.

lateral wave wave generated by a beam ofbounded extent incident at an angle close tothe critical angle. It manifests by producinga lateral shift of the bounded reflected wave.


lattice constant the length of the sides ofthe three dimensional unit cell in a crystal.

lattice structure a filter used in linear pre-diction that has two outputs, the forward pre-diction errorfm[n] and the backward predic-tion errorgm[n]. These two error signals aredefined recursively:

fm[n] = fm−1[n] +Kmgm−1[n− 1]gm[n] = K∗

mfm−1[n] + gm−1[n− 1]where theKm are the called the lattice coef-ficients of the predictor.

lattice vector quantization (1) a struc-tured vector quantizer where the reproduc-tion vectors are chosen from a highly regulargeometrical structure known as a “lattice.”The method is employed mainly because ofthe reduction in storage capacity obtained(compared to optimal vector quantization).

(2) vector quantization, where codewordsrepresent prototype vectors arranged in a reg-ular lattice inn-dimensional space.

lattice VQ Seelattice vector quantization.

law of excluded middle a logical law stat-ing that for Boolean variableX, X must beeitherZ or notZ.

law of the first wavefront SeeHaaseffect.

layout specifies the position and dimen-sion of the different layers of materials asthey would be laid on the silicon wafer.

LC-oscillator a type of oscillator that ispractical at the frequencies above 50 kHz andup to 500 MHz, where it is rational to usehigh-Q tuned circuits for frequency selec-tion (as a result, sometimes they are calledtuned-circuit oscillators). It is impossible toclassify all circuits of LC-oscillators, yet awide group of LC-oscillators are reduced tothe circuit that includes one active device (aFET characterized by transconductancegm

and drain-source impedancero is shown asan example; it can be substituted by a bipolartransistor; this essentially does not changethe results), a load resistor, and three reac-tances (ifX1 andX2 are coils, the mutualreactanceXm should be considered as well).

The frequency of oscillation is given bythe condition

X1 +X2 +X3 + 2Xm = 0

and the threshold condition for self-startingis

gmroRL

ro + RL≥ X2 +Xm

X1 +Xm

If X1 is a tuned coil in the gate,X2 is the draincoil, 2Xm is the mutual inductance betweenthe two coils, the circuit is called tuned-gateoscillator, if the single tuned circuit is movedinto the drain, the circuit is called tuned-drainoscillator. Other frequently used configura-tions are Colpitts oscillator and Hartley os-cillator.

LCD Seeliquid crystal display.

LCI Seeload-commutated inverter.

LCLV Seeliquid crystal light valve.

LDD Seelightly doped drain.

lead a conductive path, usually self-supporting; the portion of an electrical com-ponent that connects it to outside circuitry.

lead circuit a simple passive electroniccircuit designed to add a dominant zero tocompensate the performance of a given sys-tem. It is generally used to make an oscilla-tory system more damped. A nondominantpole must be included to make the circuitcausal. This pole also limits the high fre-quency gain of the lead circuit, thereby avoid-ing excessive amplification and transmissionof undesirable noise.


lead frame the metallic portion of the de-vice package that makes electrical connec-tions from the die to other circuitry.

lead lanthanum zirconium titanate (PLZT)a quadratic electro-optic material where therefractive index changes quadratically withapplied electric field. Commonly availableas a hot-pressed polycrystalline ceramic, al-though single-crystal film is being developed.

leader an elongated region of ionized gasthat extends from one electrode to anotherjust before a high-voltage breakdown.

leader-follower game SeeStackelbergequilibrium.

leading-edge triggered pertaining to adevice that is activated by the leading edgeof a signal.

leakage the flux in a magnetic circuit thatdoes not do any useful work.

leakage flux the flux that does not link allthe turns of a winding or in coupled circuits,flux that links one winding but not another.For example, the magnetic flux produced bythe primary winding of a transformer that isnot coupled to the secondary winding.

leakage reactance the amount of induc-tive reactance associated with leakage flux.The leakage flux is the flux which traverses inpaths farther from the designated paths suchas the magnetic core in transformers and theair gap in electric machines and constitutesthe non-useful flux. The electric circuit sym-bol of leakage reactance isXl . It is a functionof the leakage inductance and the frequencyof operation. Higher values of leakage reac-tance affect the regulation and efficiency ofthe system.Xl is expressed in ohms.

leaky feeder an antenna consisting of a ca-ble that continuously radiates a signal fromall points along the entire length of the cable.Such a cable is typically used as the radiating

antenna in places such as tunnels and mineswhere the range of radio propagation is lim-ited. Compare withdistributed antenna.

leaky modes Seetunneling modes.

leaky wave a wave that radiates sig-nal power out of an imperfectly closed andshielded system. Open waveguides areable, under certain circumstances, to pro-duce leaky waves, which can be used for ra-diation purposes. In this case, a leaky an-tenna couples power in small increments perunit length, either continuously or discretely,from a bound mode of the open waveguide.

learning in neural networks, the collec-tion of learning rules or laws associated witheach processing element. Each learning lawis responsible for adapting the input–outputbehavior of the processing element transferfunction over a period of time in response tothe input signals that influence the process-ing element. This adaptation is usually ob-tained by modification of the values of vari-ables (weights) stored in the processing ele-ment’s local memory.

Sometimes neural network adaptation andlearning can take place by creating or de-stroying the connections between processingelements. Learning may be also achieved byreplacing the transfer function of a process-ing element by a new one.

learning law Seelearning rule.

learning rate a parameter in a learningrule which determines the amount of changefor a parameter during the current iteration.While a large value may increase the rate oflearning, it may also prevent a learning algo-rithm from converging.

learning rule in neural networks, an equa-tion that modifies the connection weights inresponse to input, current state of the pro-cessing element, and possible desired outputof the processing element.


learning vector quantization (LVQ) asupervised learning algorithm first proposedby Kohonen that uses class information tomove the Voronoi vectors slightly, so as toimprove the quality of the classifier decisionregions.

The training algorithm for LVQ is simi-lar to its unsupervised counterpart with su-pervised error correction. After the winneris found, the weights of the winner and itsneighbors will be updated according to thefollowing rules:

If the class is correct,

wi(t + 1) = wi(t)+ α · (xi(t)− wi(t)) ,

otherwise

wi(t + 1) = wi(t)− α · (xi(t)− wi(t)) ,

See alsoself-organizing system,self-organiz-ing algorithm.

least mean square (LMS) algorithmin some cases of parameter estimation forstochastic dynamic systems, it is not feasi-ble to use the least squares based algorithmsdue to the computational effort involved inupdating and storing theP(t) (probability)matrix. This is especially so when the num-ber of parameters is large. In this case, itis possible to use a variant of the stochasticgradient algorithm — the least mean square(LMS) algorithm, which has the form

θ (t + 1) = θ (t)+ φ(t)(y(t + 1)

− φT (t + 1)θ(t))

whereθ is a parameter’s estimates vector,φ

is a regressor vector.

least recently used (LRU) algorithm areplacement algorithm based on program lo-cality by which the choice of an object (usu-ally a page) to be removed is based on thelongest time since last use. The policy re-quires bookkeeping of essential informationregarding the sequence of accesses, whichmay be kept as LRU bits or as an LRU stack.

least squares an approach to determin-ing the optimal set of free parametersEwof an input-output mappingEy = EF(Ex, Ew),whereby the square of the difference betweenthe output of the functionEy and the desiredoutput Ed is minimized.

least squares algorithm an adaptive al-gorithm that adjusts the weights of a digitalfilter to minimize a least squares (LS) costcriterion. For equalization, this cost criterionis

i∑k=1

wi−k|bk − yk|2

wherebk is the transmitted symbol,yk is thefilter output, both at timei, and wherew isan exponential weighting factor which dis-counts past data.

least squares filter the optimal filter,in the least squares sense, for restoring acorrupted signal. Specifically, given obser-vations of a signalx[n], the least squaresfilter produces thex[n], which minimizesE[∑ |x[n]− x[n]|2]. See alsoWiener filter.

least squares solution the set of free pa-rameters that satisfies the least squares crite-rion.

least-significant bit (LSB) in a binaryword, a bit with the lowest wight associatedwith it.

least-square-error fit an algorithm or setof equations resulting from fitting a polyno-mial or other type curve, such as logarithmic,to data pairs such that the sum of the squarederrors between data points and the curve isminimized.

LED Seelight emitting diode.

LEF Seelighting effectiveness factor.

left-hand circular polarization the stateof an electromagnetic wave in which the elec-tric field vector rotates anticlockwise when


viewed in the direction of propagation of thewave.

left–right models typical hidden Markovmodels adopted in automatic speech recog-nition. The state diagram for these models isa directed acyclic graph, that is, there are nocycles apart from self-loops.

legacy system applications that are in amaintenance phase but are not ready for re-tirement.

Legendre functions a collection of func-tions, typically denoted asPν(x) andQν(x),that satisfy Legendre’s equation:

(1 − x2)d2f

dx2− 2x

df

dx+ ν(ν + 1)f = 0 ,

wheref is equal to eitherPν or Qν ; ν isthe order of the function andx is its argu-ment. Typically, Legendre functions arise inboundary value problems that are based upona spherical coordinate system.

Leibniz’s formula a formula that is satis-fied by the three types ofLie derivatives,

L[f,g]h = 〈∇h, [f ,g]〉 = LgLfh− LfLgh .

Leibniz, Gottfried Wilhelm (1646–1716) Born: Leipzig, Germany

Leibniz is best known for his work inmathematics. Leibniz’s work was an im-portant contribution to the development ofdifferential calculus and logic. Leibniz alsoimproved Pascal’s early calculating machine,extending its capacity to include multiplica-tion and division. Leibniz’ later years wereembittered by the charge that he had plagia-rized some of Isaac Newton’s work on thedevelopment of the calculus. Subsequent in-vestigations have proved this charge to bebaseless.

Lempel–Ziv coding See Ziv–Lempelcoding.

Lempel–Ziv compression See Ziv–Lempel coding.

Lempel-Ziv-Welch (LZW) coding avariant of the dictionary-based coding schemeinvented by Ziv and Lempel in 1978 (LZ78),where strings of symbols are coded as indicesinto a table. The table is built up progres-sively from the input data, such that stringsalready in the table are extended by one sym-bol each time they appear in the data.

lengthened code a code constructed fromanother code by adding message symbols tothe codewords. Thus an(n, k) original codebecomes, after the adding of one messagesymbol, an(n+ 1, k + 1) code.

lens optical element for focusing or defo-cusing electromagnetic waves.

lens aberrations any deviation of the realperformance of an optical system (lens) fromits ideal performance. Examples of lens aber-rations include coma, spherical aberration,field curvature, astigmatism, distortion, andchromatic aberration.

lens array a two-dimensional array of (of-ten small) lenses fabricated on a single sub-strate.

lens design the mathematical determina-tion of the parameters of the optical elementsin an optical instrument or system requiredto satisfy its performance goals. Rays aretraced through the system to predict its per-formance and then changes of one or moreparameters are made to improve its perfor-mance. Now performed using computer soft-ware packages.

lenslet array interconnect free-space in-terconnect that uses lenslet arrays to controloptical paths from sources to detectors. Eachlenslet images the source array onto the de-tector array. It has no property of dynamicreconfiguration.

lenslike medium beam propagation med-ium in which the gain and index of refraction


may have linear or quadratic variations withdistance away from the optical axis.

level crossing rate Seefading rate.

level-1 cache in systems with two separatesets of cache memory between the CPU andstandard memory, the set nearest the CPU.Level-1 cache is often provided within thesame integrated circuit that contains the CPU.In operation, the CPU accesses level-1 cachememory; if level-1 cache memory does notcontain the required reference, it accesseslevel-2 cache memory, which in turn accessesstandard memory, if necessary.

level-2 cache in systems with two sepa-rate sets of cache memory between the CPUand standard memory, the set between level-1cache and standard memory.

level-sensitive pertaining to a bistable de-vice that uses the level of a positive or nega-tive pulse to be applied to the control input,to latch, capture, or store the value indicatedby the data inputs.

level-triggered Seelevel-sensitive.

Levenshtein distance Seeedit distance.

LFSR Seelinear-feedback shift register.

Liapunoff SeeLyapunov, Alexandr M.

Liapunov SeeLyapunov, Alexandr M.

Lichtenburg figure a pattern produced ona powder-coated flat electrode or upon photo-graphic film held between a pair of flat elec-trodes subjected to a high voltage impulseas from a lightning stroke. Impulse polarity,magnitude and waveform can be estimatedby examination of its characteristic pattern.

lid the fuse-holder portion of a cut-out.

Lie derivative one of three types of spe-cial derivatives.

1. The derivative of a vector field withrespect to a vector field that is also knownas the Lie bracket. Iff ,g : R

n → Rn are

two differentiable vector fields, then their Liebracket is

[f ,g] = ∂f∂x

g − ∂g∂x

f ,

where∂f∂x and∂g

∂x are the Jacobian matrices ofthe vector fieldsf andg, respectively. Oftenthe negative of the above is used.

2. The derivative of a functionh with re-spect to a vector fieldf defined as

Lfh = 〈∇h, f 〉 = (∇h)T f ,

where∇h : Rn → R

n is the gradient ofh.3. The derivative ofdh = (∇h)T with

respect to the vector field,

Lf (dh) =(∂∇h∂x

f)T

+ (∇h)T ∂f∂x

.

See alsoLeibniz’s formula.

lifetime broadening a spectral line broad-ening mechanism that is a consequence ofthe finite lifetime of the excited state. Nu-merically, the lifetime broadened linewidthis equal to the inverse lifetime of the excitedstate.

LIFO Seefirst-in-last-out.

lift-off process a lithographic process bywhich the pattern transfer takes place by coat-ing a material over a patterned resist layer,then dissolving the resist to “lift off” the ma-terial that is on top of the resist.

light emitting diode (LED) a forward-biased p-n junction that emits light throughspontaneous emission by a phenomenontermed electroluminescence.

light guide system of lenses, mirrors,graded index, or graded gain media that hasthe capability of overcoming diffraction andguiding an electromagnetic wave at opticalfrequencies.


light loss factor (LLF) the ratio of the il-lumination when it reaches its lowest level atthe task just before corrective action is taken,to the initial level if none of the contributingloss factors were considered.

light pen an input device that allows theuser to point directly to a position on thescreen. This is an alternative to a mouse.Unlike a mouse, a light pen does not requireany hand/eye coordination skills, because theusers point to where they look with the pen.

light scattering (1) spreading of the lightas it passes or is reflected by an optically in-homogeneous medium.

(2) the process in which a beam oflight interacts with a material system andbecomes modified in its frequency, polar-ization, direction of propagation, or otherphysical property. See alsospontaneouslight scattering, stimulated light scattering,Brillouin scattering, Raman scattering.

light valve Seespatial light modulator.

light-activated silicon controlled rectifiera silicon controlled rectifier in which the

gate terminal is activated by an optical sig-nal rather than an electrical signal.

light-amplifying optical switch (LAOS)vertically integrated heterojunction photo-transistor and light emitting diode that haslatching thyristor-type current-voltage char-acteristics.

lighting effectiveness factor (LEF) theratio of equivalent sphere illumination to or-dinary measured or calculated illumination.

lighting system any scheme used for illu-minating a scene, usually for acquisition by adigital system. Illumination is crucial to dig-ital images, since even illumination gradientsthat cannot be perceived by the eye can havean influence on the results of digital process-ing. For inspection tasks and document dig-itization, a uniform, reproducable, high level

of lighting is usually required. Other applica-tions have other requirements for uniformity,frequency, and intensity.

Structured lighting schemes are used tocollect multiple images of a scene each hav-ing different illumination .Strobe lights can be used to effectivelyfreezemotion, andare useful for many visualinspection tasks. See also structured light.

lightly doped drain (LDD) in an MOStransistor, an extension to the source/draindiffusion that is separated slightly from thegate region which contains lower dopingthan that used for the source/drain diffusions.Since the source and drain diffusions areheavily doped, the lightly doped extensiontends to increase the width of the depletion re-gion around the drain which lowers the elec-tric field intensity, to increase the breakdownvoltage of the drain region of deep submicrondevices.

lightning arrestor a voltage-dependentresistor which is connected in parallel withlightning-susceptible electrical equipment.It provides a low-resistance electrical path toground during overvoltage conditions, thusdiverting destructive lightning energy aroundthe protected equipment.

lightning choke one of several arrange-ments of conductors, usually a single ormulti-turn coil, used to reduce lightning cur-rents by increasing a power line’s impedanceat lightning frequencies.

lightwave communications optical com-munications techniques that used guidedwave optical devices and fiber optics.

lightwave technology technology basedon the use of optical signals and optical fiberfor the transmission of information.

likelihood ratio the optimum processorfor reducing a set of signal-detection mea-


surements to a single number for subsequentthreshold comparison.

likelihood ratio test a test using the likeli-hood ratio that can be used along with thresh-old information to test different information-content hypotheses. An example of a signaldetection problem is the demodulation of adigital communication signal, for which thelikelihood ratio test may be used to decidewhich of several possible transmitted sym-bols has resulted in a given received signal.

limit cycle undamped but bounded oscil-lations in a power system caused by a distur-bance.

limited-look-ahead control predictivecontrol policy whereby — unlike in the caseof open-loop-feedback control — the deci-sion mechanism used at each interventioninstant takes into account one or more, butnot all, remaining future interventions; thisin particular means that such decision mech-anism requires the usage of at least two sce-narios of the future free input values over theconsidered prediction interval.

limiter an equipment or circuit that hasa function to keep output power constant. Itcan also be used to protect other circuits notto be overdriven.

limiting spatial resolution for an imag-ing photodetector, the maximum number ofblack and white bar pairs of equal width andspacing that can be resolved per unit length,usually given in units of line-pairs per mil-limeter.

Linde–Buzo–Gray (LBG) algorithm (1)an algorithm for vector quantizer design, dueto Y. Linde, A. Buzo and R. M. Gray (1980).The procedure is based on the principles ofthe generalized Lloyd algorithm and the K-means clustering algorithm.

(2) an iterative method for designing acodebook for a vector quantizer using a setof training data that is representative of the

source to be coded. Otherwise known as thegeneralized Lloyd algorithm or K-means al-gorithm.

line (1) on a bus structure, one wire ofthe bus, which may be used for transmitting adatum, a bit of an address, or a control signal.

(2) in a cache, a group of words fromsuccessive locations in memory stored incache memory together with an associatedtag, which contains the starting memory ref-erence address for the group.

(3) a power-carrying conductor or groupof conductors.

line broadening nonzero spectral widthof an absorbing or emitting transition; causedby many physical effects.

line code modification of the source sym-bol stream in a digital communication systemto control the statistics of the encoded sym-bol stream for purposes of avoiding the oc-currence of symbol errors that may arise dueto limitations of practical modulation and de-modulation circuitry. Also called recordingcodes or modulation codes.

line conditioner Seepower conditioner.

line detection the location of lines or linesegments in an image by computer. Oftenaccomplished with the Hough transform.

line drop compensator a multiply-tapped autotransformer equipped with a load-sensing relay which will adjust the line volt-age to compensate for the impedance drop inthe circuit between the device and the loadcenter.

line hose split rubber tubing which is ap-plied over energized electric conductors astemporary insulation to protect nearby work-ers.

line impedance stabilization network(LISN) a network designed to present adefined impedance at high frequency to a de-


vice under test, to filter any existing noiseon the power mains, and to provide a 50-�

impedance to the noise receiver.

line of sight (LOS) the shortest possi-ble straight line that can be envisioned, re-gardless of possible obstacles in the way, be-tween a transmitter and a receiver. If a line ofsight between transmitter and receiver is notblocked, the strongest signal will be receivedfrom the line-of-sight direction.

line outage distribution factor a ratioused in contingency analysis. Given two par-allel lines in a power system called x and y,assume that line y is removed from service.The line outage distribution factor of line xfor the outage of line y is the ratio of thechange in power flow on line x to the flow online y before the outage.

line rate Seehorizontal rate.

line shape function shape of the spectrumof an emission or absorption line.

line spread function the response of asystem to a 2-D input consisting of a singleline. If the system is linear and space invari-ant, its output to any image with a 1-D patternis a sum of weighted line spread functions.Such patterns are often used in vision experi-ments.See alsolinear shift invariant system,point spread function, space invariance.

line to line fault a fault on a three phasepower line in which two conductors have be-come connected.

line width width of the spectrum of anemission or absorption line; often full widthat half maximum, but other definitions alsoused.

line-connected reactor Seeshunt reactor.

line-current harmonic Seeelectromagneticinterference filter.

line-impact printer a printer that printsa whole line at a time (rather than a singlecharacter). An impact printer has physicalcontact between the printer head and the pa-per through a ribbon. A dot-matrix printer isan impact printer, whereas an ink-jet printeris not. A line-impact printer is both a lineand impact printer.

line-reflect-match (LRM) calibrationan error correction scheme (calibration)where the calibration standards used area transmission line, a reflect load, and amatched load. Line-reflect-match (LRM)two-port calibration requires a line standard(1 picosecond line), reflect standard (opencircuit is preferred, short circuit is optional),and match standard (50 ohm load). The LRMtechnique is similar to the TRL technique, ex-cept the reference impedance is determinedby a load instead of a transmission line. Ithas the advantages of self-consistency, re-quires only two standards to contact, and hasno bandwidth limitations.

line-reflect-reflect-match (LRRM) calibra-tion an extension of line-reflect-match(LRM) calibration. The second reflect stan-dard in LRRM is used for correcting the in-ductance caused by the match standard.

line-to-line run-length difference codinga coding scheme for graphics. In this ap-

proach, correlation between run-lengths insuccessive lines is taken into account. Dif-ferences between corresponding run-lengthsof successive scan lines are transmitted.

line-to-line voltage a voltage measure-ment of a three phase line made between anytwo conductors.

linear a circuit or element in whichthe output spectrum is proportional throughgain(s), attenuation(s) and delay(s) to the in-put spectrum, and in which no spectral shift,conversion or generation takes place. Truelinearity is seldom encountered in the realworld, but often used in approximate descrip-


tions, thereby promoting understanding andsimplifying computation.

linear approximation any technique usedfor the purpose of analysis and design of non-linear systems. For example, one way of an-alyzing the stability of a system described bynonlinear differential equations is to linearizethe equations around the equilibrium point ofinterest and check the location of the eigen-values of the linear system approximation.

linear block code a block coding schemefor which the mapping can be described by alinear transformation of the message block.The transformation matrix is referred to asthe generator matrix.

linear code a forward error control codeor line code whose code words form a vec-tor space. Equivalently, a code where theelement-wise finite field addition of any twocode words forms another code word.

linear constant-coefficient equation ageneralN th-order linear constant-coefficientdifferential equation is of the form

N∑0

akdky(t)

dtk=

M∑0

bkdkx(t)

dtk

while anN th-order linear constant-coefficientdifference equation is of the form

N∑0

aky[n− k] =M∑0

bkx[n− k]

Linear constant-coefficient equations at ini-tial rest are linear time invariant and causal,and can be conveniently analyzed usingtransform techniques.

linear dynamic range of Bragg cellregime of cell operation where the amplitudeof the principal diffracted beam is approx-imately proportional to the acoustic signalamplitude modulating the acousto-optic in-teraction medium.

linear filter a filter whose output signalis a linear function of the input (that is, inputand output are related via a convolution).Seealsoconvolution, linear system.

linear generator Seelinear machine.

linear interpolation linear interpolationis a procedure for approximately reconstruct-ing a function from its samples, wherebyadjacent sample points are connected by astraight line.

linear least squares estimator (LLSE)the linear estimatorx = Ky + c,

where matrix K and vector c are cho-sen to minimize the expected squared errorE[(x − x)T (x − x)

]. The general LLSE

solution to estimate a random vectorx basedon measurementsy is given by

x(y) = E[x] + cov(x, y)

· cov(y, y)−1 · (y − E[y])

where “cov” represents the covariance op-eration. See alsoleast squares, covariance,expectation,minimum mean square estimator.

linear load an electrical load with a cur-rent that is linearly proportional to the voltagesupplied.

linear machine a machine in which themoving member constitutes linear motion in-stead of the more conventional rotary motion.Each of the rotary machine types can be pro-duced in linear versions. The most widelyknown use of linear motors are in the fieldof transportation, where the stator is usuallythe moving vehicle and the conducting rotorsare the rails. In these machines, the inducedcurrents provide levitation in addition to pro-viding the main propulsion.

linear medium (1) medium in which theconstitutive parameters are not functions ofthe electric or magnetic field amplitudes.


(2) medium in which any response is di-rectly proportional in magnitude to the mag-nitude of the applied field.

linear motor Seelinear machine.

linear multistep method this is a classof techniques for solving ordinary differen-tial equations which is widely used in circuitsimulators.

linear network a network in which theparameters of resistance, inductance and ca-pacitance are constant with respect to voltageor current or the rate of change of voltage orcurrent and in which the voltage or currentof sources is either independent of or propor-tional to other voltages or currents, or theirderivatives.

linear phase system where the phase shiftproduced by the filter at frequencyw is a lin-ear function ofw (6 H(w) = dw). If a signalx(t) is passed through a unit magnitude, lin-ear phase filter with sloped, the output signalwill be x(t+d), the input signal time-shiftedby d seconds.

linear polarization a polarization state ofa radiated electromagnetic field in which thetip of the electric field vector remains on aline and does not rotate as a function of timefor a fixed position.

linear prediction for a stochastic process,the prediction of its samples based upon de-termining a linear model capable of estimat-ing the samples with minimal quadratic error.

linear prediction based speech codingSeelinear predictive coding.

linear predictive coding speech codingmethods where short-term redundancy of thespeech signal is removed by linear predictionanalysis prior to encoding.See alsoadaptivedifferential pulse code modulation.

linear predictor a predictor that usesa weighted sum ofK previous samples ofthe original signal withαi, i = 1, . . . , K asweights.

linear quadratic control for a linear de-terministic plant, the problem of determiningthe control structure that minimizes the per-formance index.

Given a plant in the form of the state-spaceequation

x(t) = Ax(t)+ Bu(t)

The associated performance index is thequadratic form

J = 1

2x(T )S(T )x(T )

+ 1

2

∫ T

t0

(xT (t)Qx(t)+ uT (t)Ru(t)dt

with symmetric weighting matricesS(T ) ≥0, Q ≥ 0, R > 0. Both plant and weight-ing matrices can be functions of time. Lin-ear quadratic control is the problem of deter-mining the controluo(t) on (t0, T ) that min-imizes the performance indexJ with x(T )free andT fixed. Also called LQ control.

linear quadratic Gaussian control fora plant, the problem of finding the controlstructure that minimizes the expected cost.

The linear stochastic plant is given in thefollowing state-space form:

x(t) = Ax(t)+ Bu(t)+Gw(t)

with white noisew(t) and x(t0) a randomvariable. The associated performance indexis the quadratic form

J = 1

2xT (T )S(T )x(T )

+ 1

2

∫ T

t0

(xT (t)Qx(t)+ uT (t)Ru(t)dt

with symmetric weighting matricesS(T ) ≥0,Q ≥ 0, R > 0. The plant and weighting


matrices can be functions of time. It is de-sired to determine the controluo on (t0, T )which minimizes the expected cost

j = E(J )

with x(T ) free andT fixed. This problem iscalled the linear quadratic Gaussian control.

linear response the characteristic of manyphysical systems that some output propertychanges linearly in response to some appliedinput. Such systems obey the principle oflinear superposition.

linear scalar quantization Seeuniformscalar quantization. Also known as linearSQ.

linear scrambler a linear one-to-onemapping of a codeword,c, of lengthn ontoa new codeword,cc, also of lengthn. cc isdetermined ascc = S c, whereS is the linearscrambler matrix. This matrix is full rankand, for the binary case, has all binary en-tries. In the binary case, modulo-2 arithmeticis performed.

linear separation the process of deter-mining the hyper plane that separates a givenset of patterns according to their member-ship. Of course, such separation can only beobtained once the patterns are linearly sepa-rable.

linear shift invariant (LSI) system a lin-ear discrete-time systemT [], with y[n] =T [x[n]] and y[n − n0] = T [x[n − n0]].See alsolinear system, linear time invariantsystem, shift invariance.

linear SQ Seeuniform scalar quantization.

linear susceptibility the coefficientχ re-lating the polarizationP of a material sys-tem (assumed to exhibit linear response) tothe applied electric strengthE according toP = χE.

linear system the systems in whichthe components exhibit linear characteristics,i.e., the principle of superposition applies.Strictly speaking, linear systems do not existin practice; they are idealized models purelyfor the simplicity of theoretic analysis anddesign. However, the system is essentiallylinear when the magnitude of the signals in acontrol system are limited to a range in whichthe linear characteristics exist.

More formally, consider a system withzero initial conditions such that any two in-put/output signal pairs{f1, y1} and{f2, y2}satisfy the system equation. The system isadditive if the input/output pair{f1+f2, y1+y2} also satisfies the system equation. Thesystem is homogeneous if for any real con-stantC, the input/output pair{Cf1, Cy1} sat-isfies the system equation. The system islinear if it is additive and homogeneous. Inother words, for any real constantsC1 andC2,the input/output pair{C1f1+C2f2, C1, y1+C2y1} satisfies the system equation. If a sys-tem is not linear, then it is nonlinear. Thetheory of linear systems is well-developed,hence many tools exist for the analysis of sys-tem behavior. In practice, nonlinear systemsare often approximated by a linear model sothat the tools may be exploited.

linear systems with Markov jumps aclass of piecewise deterministic processesthat follows linear dynamics between randomjumps of parameters that in turn can be de-scribed by finite-state Markov processes. Inthe continuous-time case, linear systems withMarkov jumps could be modeled by linearstate equations of the form:

x(t) = A(ξ(t))x(t)+ B(ξ(t))u(t)

where t ∈ [0, T ], T being finite or infi-nite control horizon,x(t) ∈ Rn is the pro-cess state,u(t) ∈ Rm is the process control,A,B are real valued matrices of respectivedimensions depending on the random pro-cess{ξ(t)}. This process is a continuous-timediscrete-state Markov process taking valuesin a finite setS = {1,2, . . . , s} called mode


with transition probability matrixP = {pij }given by

pij = Pr(ξ(t + δt)

= j |ξ(t) = i) = qij δt +O(δt)

if i 6= j and

pii = 1 + qiiδt +O(δt)

whereδt > 0, qij is the intensitivity or tran-sition rate fromi to j if i 6= j and

qii = −qi = −s∑

j=1,j 6=iqij

Linear systems with Markov jumps mayserve as models for continuous-time pro-cesses subject to abrupt changes in param-eter values because of component failures,sudden shifts in environment or subsystemsconnections. Generally, linear systems withjumps are hybrid in the sense that their statecombines a part taking values from continu-ous space (process state) and a part that takesvalues from discrete space (mode).

linear threshold unit a neural elementthat computes the weighted sum of its inputsand compares that sum to a threshold value.If the sum is greater than (or equal to) thethreshold, the output of the element takes onthe value+1. Otherwise, the output takes onthe value 0 (in a binary system) or−1 (in abipolar system).

linear time invariant (LTI) system a lin-ear systemT [], with y(t) = T [x(t)] andy(t − t0) = T [x(t − t0)]. See alsolinearsystem, linear shift invariant system (LSI),shift invariance.

linear time-invariant lumped-parameter(LTIL) system a continuous system thatcan be described by an ordinary differentialequation with real constant coefficients. Asingle loop RLC circuit is an example of alinear time-invariant lumped-parameter sys-tem (LTIL), with the following differential

equation relating the input voltagef (t) tothe output loop currenty(t):

Ld2y

dt2+ R

dy

dt+ 1

Cy = df

dt

A discrete time system is a LTIL system ifit can be modeled by a difference equationwith real constant coefficients. For exam-ple, the following LTIL difference equationrepresents a discrete time approximation of adifferentiator:

y[k] = 1

T(f [k] − f [k − 1])

A large number of physical systems are LTILsystems. Since the theory for modeling andanalysis of such systems is well developed,LTIL systems are often used to approxi-mately model physical processes in order tosimplify the mathematical analysis.

linear transformation a transformationoperatorA which satisfies superposition,

A(x1 + x2) = Ax1 + Ax2

and homogeneity

A(λx1) = λAx1.

For a discrete linear transformA is a ma-trix and x1 and x2 are vectors. Any ma-trix transform is linear. Seelinear system,superposition.

linear-feedback shift register (LFSR) ashift register formed by D flip-flops and ex-clusive OR gates, chained together, with asynchronous clock.

linear-quadratic game one of a classof noncooperative infinite dynamic gameswith state equations linear with respect tostate variablesx and players actionsui; i =1,2, . . . , N and cost functions quadratic withrespect to those variables with weighting ma-trices semipositive definite for the state andpositive definite for the players actions. Theexistence and uniqueness of the open-loop


Nash equilibrium in such games is guaran-teed under assumptions on the existence ofa unique solution for respectively definedcoupled Riccati equations. The equilibriumstrategies appear to be linear functions of theassociated state trajectories. The existenceand linear form for the feedback Nash strate-gies could be guaranteed by the existence ofpositive semidefinite solution to the relevantcoupled Riccati equations but it does not at-tribute an uniqueness features to the solu-tion set. For the zero-sum linear-quadraticgames, the situation becomes simpler bothin discrete-time and continuous-time cases.The saddle point strategies could be foundby solving standard Riccati equations, andwhenever both open-loop and closed-loopsolutions do exist they generate the same statetrajectories. For example, in the continuous-time zero-sum linear-quadratic game definedby the state equation

x = Ax + B1u1 + B2u2

and the quadratic cost functional

J = 1

2

∫ T

0(x′Qx + u′

1u1 + u′2u2)dt

the saddle point strategies are given by

ui = (−1)iB ′iK(t)x; i = 1,2

whereK(t) is a unique symmetric boundedsolution to the matrix differential Riccatiequation

K+A′K+KA+Q−K(B1B′1−B2B

′2) = 0

with K(T ) = 0.

linearity a property of a system if thatsystem obeys the principle of superposition.In other words, if the outputy(t) is a functionof the inputx(t), i.e., y(t) = f (x(t)), andif x(t) = αx1(t) + βx2(t), then for a linearsystem,y(t)will be = αy1(t)+βy2(t)wherey1(t) = f (x1(t)) andy2(t) = f (x2(t)).

linearization approximation of a nonlin-ear evolution equation in a small neighbor-

hood of a point by a linear equation. This isobtained by keeping only the first order termsin the Taylor series expansion of the nonlin-earities about this point.

linearized machine equations state equa-tions obtained by linearizing the nonlinearvoltage and electromagnetic torque equa-tions of induction or synchronous machines.The state variables can be either currents, orflux linkages. The linearization can be ac-complished using a Taylor series expansionof the machine variables about an operatingpoint, i.e.,

g(f ) = g(f0)+ g′(f0)1f

+ higher order terms

wheref = f0 +1f (andf0 is the value off at a given operating point).

The small displacement characteristicsare then approximated as

g(f )− g(f0) = 1g = g′(f0)1f

The linearized equations are typically ma-nipulated into a standard state model form

dx

dt= Ax + Bu

y = Cx +Du

and are used for Eigensystem, stability, andcontrol analysis and design.

linearizer an equipment or circuit that isused to reduce distorted components gener-ated in nonlinear amplifiers such as traveling-wave tube amplifier (TWTA) or solid statepower amplifier (SSPA). There are variouskinds of linearizers, such as predistortion,feedback, and feedforward types.

lineman utility employee working on pri-mary facilities, distribution class equipment,as opposed to customer service level facili-ties.

linguistic hedge Seemodifier.

linguistic variable variable for which val-ues are not numbers, but words or sentences


in a natural or artificial language. In fuzzyset theory, the linguistic values (or terms) ofa linguistic variable are represented by fuzzysets in an universe of discourse. In the exam-ple shown in the figure, the linguistic variableis speed,the universe of discourse is associ-ated to the base variables, and the fuzzy sets(through the plot of their membership func-tions) associated to the linguistic valueslow,medium,andhigh.

Definition of the linguistic variable speed.

See alsomembership function, fuzzy set.

link (1) the portion of the compilation pro-cess in which separate modules are placed to-gether and cross-module references resolved.

(2) a linkage (joint) in a manipulator arm.

link inertial parameters for a manipu-lator arm, consists of six parameters of theinertia tensor, three parameters of its centerof mass multiplied by mass of the link (moreprecisely, three components of the first or-der moment) and mass of the link. Dynamicproperties of each link are characterized by10 inertial parameters. They appear in thedynamic equations of motion of the manipu-lator.

linkage flux also called magnetizing ormutual flux. In a magnetically coupled cir-cuit such as a transformer, the linkage fluxis the flux that links all the windings. Forexample, in a transformer the magnetic fluxproduced by the primary winding which iscoupled to the secondary winding.

linker a computer program that takes oneor more object files, assembles them into

blocks that are to fit into particular regions inmemory, and resolves all external (and pos-sibly internal) references to other segmentsof a program and to libraries of precompiledprogram units.

Linville stability factor the inverse of theRollett stability factor (K),C is a measure ofpotential stability in a 2-port circuit operat-ing under small signal conditions, but stand-alone is insufficient to guarantee stability. A2-port circuit that is matched to a positivereal source and load impedance is uncondi-tionally stable if 0< C < 1,B1 > 0 (port 1stability measure) andB2 > 0 (port 2 stabil-ity measure). The design must provide suffi-cient isolation from the RF input and outputports to the bias ports to allow a reasonableinterpretation of the “2-port” device criteria.

C = 1

K= 2· | s12 · s21 |

1− | s11 |2 − | s22 |2 + | 1 |2

where1 = s11 · s22 − s12 · s21.

Lipschitz condition for a vector func-tion f : <n −→ <n , where<n is then-dimensional real Euclidean space, the condi-tion

‖f (x)− f (y)‖ ≤ b‖x − y‖where‖.‖ is any vector norm inn and b,which is called the Lipschitz constant, is apositive real number.

Lipschitz continuous system Seeincre-mental gain.

liquid crystal class of organic polymericmaterials made up of elongated moleculesthat show various degrees of order in one,two, and three dimensions.

liquid crystal display (LCD) the screentechnology commonly used in notebook andsmaller computers.

liquid crystal light valve (LCLV) a typeof optically addressed spatial light modulator


that uses twisted nematic liquid crystal ma-terial and a photosensitive layer for opticalinputs.

liquid crystal on silicon (SLM) a type ofelectrically addressed spatial light modulatorusing liquid crystal material on top of a VLSIsilicon circuit used for electrical signal inputand signal preprocessing.

liquid laser laser in which the activemedium is a liquid, dye lasers being the mostcommon example.

LISN Seeline impedance stabilizationnetwork.

list decoding decoding procedure inwhich the decoder, instead of producing asingle estimate of the transmitted codeword,yields a list of candidate codewords, for ex-ample theL most likely.

list of capabilities usually associated witha process, defining a set printed circuit boardthat can be plugged into a main board to en-hance the functionality or memory of a com-puter.

literal a data type consisting of alphanu-meric data.

lithium niobate (LiNbO 3) a strong linearelectro-optic material, also strongly piezo-electric and possessing high acousto-opticfigure of merit with low acoustic attenuation.

lithography (1) for a semiconductor man-ufacturing process, the process of printingimages of the various circuit layers on thewafer via a photographic technique. Thetechnique uses a radiation source, such aslight, electrons, or X-rays, to generate a pat-tern in a radiation-sensitive material. Theradiation-sensitive material is illuminatedthrough a mask, that prevents certain portionsof the material from being exposed. Exposedmaterial is changed chemically such that it is

either removed by or resistant to a solutionused to develop the image.

(2) a method of producing three-dimensionalrelief patterns on a substrate (from the Greeklithos,meaning stone, andgraphia,meaningto write).

little endian a memory organizationwhereby the byte within a word with thelowest address is the least significant, andbytes with increasing address are succes-sively more significant. Opposite of big en-dian. Sometimes believed (with no merit)to be either the “right” or the “wrong” mem-ory organization, hence the name (cf. Swift’sGulliver’s Travels).

For example, in a 32-bit, or four-byte wordin memory, the most significant byte wouldbe assigned addressi, and the subsequentbytes would be assigned the addressesi − 1,i − 2, andi − 3. Thus, the least significantbyte would have the lowest address ofi − 3in a computer implementing the little endianaddress assignment.See alsobig endian.

live insertion the process of removingand/or replacing hardware components (usu-ally at the board level) without removal ofsystem power and without shutting down themachine.

live tank breaker a power circuit breakerwhere the tank holding the interruptingchamber is not at ground potential. SF6 cir-cuit breakers, for example, are typically livetank breakers.

livelock a condition where attempts bytwo or more processes to acquire a resourcerun indefinitely without any process succeed-ing.

LLF Seelight loss factor.

Lloyd–Max scalar quantization a scalarquantizer designed for optimum performance(in the minimum mean squared error sense).The method, and the corresponding designalgorithm, are due to S. P. Lloyd (1957) and


J. Max (1960). Also referred to as PDF-optimized quantization, since the structureof the scalar quantizer is optimized to “fit”the probability density function (PDF) of thesource.

Lloyd–Max SQ SeeLloyd–Max scalarquantization.

LLSE Seelinear least squares estimator.

LMS Seeleast mean square algorithm.

LMS algorithm Seeleast mean squarealgorithm.

LNA Seelow noise amplifier.

load balancing the process of trying todistribute work evenly among multiple com-putational resources.

load break device any switch, such asa circuit breaker or sectionalize capable ofdisconnecting a power line under load.

load buffer a buffer that temporarily holdsmemory-load (i.e., memory-write) requests.

load bypass a read (or load) request thatbypasses a previously issued write (store)request. Read requests stall a processor,whereas writes do not. Therefore high-performance architectures permit load by-pass. Typically implemented using write-buffers.

load center the geographic point withina load area, used in system calculations,at which the entire load could be concen-trated without affecting the performance ofthe power system.

load flow study Seepower flow study.

load frequency control the purpose ofload frequency control is to maintain thepower system frequency at its nominal valuewhile maintaining the correct outputs on in-dividual generators to satisfy the loading on

the system. As the load varies, the inputsto the generator prime movers must be con-trolled to keep the generation in balance withthe loads.

load instruction an instruction that re-quests a datum from a virtual memory ad-dress, to be placed in a specified register.

load line with a slope, also known as thepermeance coefficient, determined solely bythe geometry of the magnetic circuit, this lineintersects with the normal demagnetizationcurve to indicate a magnet’s operating point.

load mismatch the load impedance doesnot match the device output impedance, re-sulting in power reflection. A perfect matchoccurs when the real parts of the load anddevice output impedance are equal and thereactive parts cancel or resonate, resultingin maximum power transfer. The magnitudeof the load mismatch is usually expressed interms VSWR, reflection coefficient or returnloss.

load tap changer (LTC) a tapped trans-former winding combined with mechanicallyor electronically switched taps that can bechanged under load conditions. The load tapchanger is used to automatically regulate theoutput of a transformer secondary as load andsource conditions vary.

load torque the resisting torque appliedat the motor shaft by the mechanical load thatcounterbalances the shaft torque generatedby the motor and available at the shaft.

load-break device any switch which canbe opened while the circuit is loaded

load-commutated inverter (LCI) an in-verter in which the commutating voltages aresupplied by the load circuit.

load-pull the systematic variance of themagnitude and phase of the load terminationof a device under test.


load-pull measurement Seeactive load-pull measurement, harmonic load-pull mea-surement.

load/store architecture a system designin which the only processor operations thataccess memory are simple register loads andstores.

load/store unit a computer based on theload/store architecture.

loaded Q dimensionless ratio of theaverage over any period of time (T =1/frequency) of the ratio of the maximum en-ergy stored (Umax) to the power absorbed ordissipated (pabsorbed= pin − pout) in a pas-sive component or circuit, including externalloading effects, expressed as a dimensionlessratio. For most applications, the higher theQ, the better the part.

local area network a network of comput-ers and connection devices (such as switchesand routers) that are located on a single site.The connections are direct cables (such asUTP or optical fiber) rather than telecommu-nication lines. The computer network in auniversity campus is typically a local areanetwork.

local bus the set of wires that connects aprocessor to its local memory module.

local controllability of generalized 2-Dmodel the generalized 2-D model



i, j ∈ Z+ (the set of nonnegative integers)is called locally controllable in the rectangle[0, N1] × [0, N2] if for admissible bound-ary conditionsxi0 for i ∈ Z+ andx0j forj ∈ Z+, there exists a sequence of inputsuijfor 0 ≤ i ≤ N1 + n1 and 0≤ j ≤ N2 + n2such thatxN1N2 = 0 wherexij ∈ Rn isthe local semistate vector,uij ∈ Rm is theinput vector,E, Ak, Bk (k = 0,1,2) arereal matrices withE possibly singular, and

(n1, n2) is the index of model.See alsolocalreachability of generalized 2-D model.

local decision unit control agent or a partof the controller associated with a given sub-system of a partitioned system; local decisionunit is usually in charge of the local decisionvariables and is a component of a decentral-ized or a hierarchical control system.

local decision variable control inputs as-sociated with a given subprocess (subsystem)of the considered partitioned process (sys-tem); local decision variables can be eitherset locally by local decision unit, or globallyby a centralized controller.

local field effect effect associated withthe distinction that occurs in condensed mat-ter between the spatially averaged electricfield and the field that acts on a represen-tative molecule of the material system. Aconsideration of local field effects leads tothe Lorentz–Lorenz and Clausius–Mossottirelations.

local memory memory that can be ac-cessed by only one processor in a multipro-cessor or distributed system. In many multi-processors, each processor has its own localmemory.See alsoglobal memory.

local minimum a minimum of a functionthat is not the global minimum.

local mode oscillation this mode of os-cillation is associated with the swinging ofunits at a generating station with respect tothe rest of the power system. The oscilla-tions are thus localized to within a small partof the system.

local observability of 2-D Fornasini–Marchesini model the 2-D Fornasini–Marchesini model

xi+1,j+1 = A1xi+1,j + A2xi,j+1

+ B1ui+1,j + B2ui,j+1

yij = Cxij


i, j ∈ Z+ (the set of nonnegative integers)is called locally observable in the rectangle[0, N1] × [0, N2] if there is no local statesx10 6= 0 andx01 6= 0 such that for zero inputsuij for 0 ≤ i ≤ N1 and 0≤ j ≤ N2 and zeroboundary conditionsxi0 = 0 for i ≥ 2 andx0j for j ≥ 2 the output is also zeroyij = 0for 0 ≤ i ≤ N1 and 0≤ j ≤ N2. The modelis locally observable in[0, N1] × [0, N2] ifand only if

rank

CT10...

CTN1−1,0CT01...

CT0,N2−1CT11...

CTN1−1,N1−1

[A1, A2] = 2n

where the transition matrixTij of the modelis defined by

Tij =

In for i = j = 0Tij = A1Ti−1,j + A2Ti,j−1for i, j ≥ 0 (i + j 6= 0)Tij = 0 for i < 0 or/and j < 0

local observability of 2-D Roesser modelthe 2-D Roesser model[xhi+1,jxvi,j+1

]=[A1 A2A3 A4

][xhijxvij

]+[B1B2

]uij

yij = C

[xhijxvij

]i, j ∈ Z+ (the set of nonnegative integers)is called locally observable in the rectangle[0, N1] × [0, N2] if there is no local ini-tial statex00 6= 0 such that for zero inputsuij = 0 for 0 ≤ i ≤ N1 and 0≤ j ≤ N2 andzero boundary conditionsxh0j = 0 for j ≥ 1andxvi0 = 0 for i ≥ 1, the output is also zeroyij = 0 for 0 ≤ i ≤ N1 and 0≤ j ≤ N2wherexhij ∈ Rn1 andxvij ∈ Rn2 are the hori-zontal and vertical local state vectors, respec-tively, uij ∈ Rm is the input vector,yij ∈ Rp

is the output vector, andA1,A2,A3,A4, B1,B2,C are real matrices. The model is locallyobservable in[0, N1] × [0, N2] and only if

rank

C

CT10CT01...

CTij...

CTN1N2

= n

where the transition matrix is defined by

Tij =

I for i = j = 0[A1 A20 0

]for i = 1, j = 0 ;

[0 0A3 A4

]for i = 0, j = 1 and

T10Ti−1,j + T01Ti,j−1 fori, j ∈ Z+ (i + j 6= 0)

Tij = 0 for i < 0 or/and j < 0

local oscillator (1) an oscillator or circuitthat produces a periodic signal whose func-tion is to be utilized in the demodulation ofa received radio signal. This periodic sig-nal is typically a sinusoid and the oscillatoris typically located in a radio receiver. Thetuning of the radio to a given channel, or sta-tion, typically involves the tuning of the localoscillator. The local oscillator is part of theradio frequency (RF) front end of a radio re-ceiver and is an important component in aheterodyne receiver.

(2) the signal applied to a mixer circuitthat is of a sufficient level to bias the diodeswithin the mixer into a nonlinear region sothat the mixing process may occur.

local oxidation of silicon (LOCOS)masked oxidation of silicon to provide elec-tronic isolation between devices. Made pos-sible by relatively slow oxidation of siliconnitride, which is used as a mask.


local reachability of generalized 2-D modelthe generalized 2-D model


+ A2xi,j+1 + B0uij

+ B1ui+1,j + B2ui,j+1

i, j ∈ Z+ (the set of nonnegative integers)is called locally reachable in the rectangle[0, N1] × [0, N2] if for admissible bound-ary conditionsxi0, i ∈ Z+ andx0j , j ∈ Z+and every vectorxf ∈ Rn there exists a se-quence of inputsuij for 0 ≤ i ≤ N1 + n1and 0≤ j ≤ N2 + n2 such thatxN1N2 = xf ,wherexij ∈ Rn is the local semistate vec-tor, uij ∈ Rm is the input vector,E, Ak, Bk(k = 0,1,2) are real matrices withE possi-bly singular. The model is locally reachablein [0, N1] × [0, N2] if and only if

rank[M0,M11, . . . ,M1

N1,M2

1, . . . ,M2N2,

M11, . . . ,M1N2,M21, . . . ,MN1,N2

]= n

M0 = TN1−1,N2−1B0,

M1p := TN1−p,N2−1B1 + TN1−p−1,N2−1B0

for p = 1, . . . , N1 = N1 + n1

M2q := TN1−1,N2−qB2 + TN1−1,N2−q−1B0

for q = 1, . . . , N2 = N2 + n2

Mpq := TN1−p−1,N2−q−1B0

+ TN1−p,N2−q−1B1 + TN1−p−1,N2−qB2

for

{p = 1, . . . , N1

q = 1, . . . , N2

and the transition matrixTpq is defined by

ETpq =

A0T−1,−1 + A1T0,−1

+A2T−1,0 + I for p = q = 0A0Tp−1,q−1 + A1Tp,q,−1

+A2Tp−1,qfor p 6= 0 and/or q 6= 0

and

[Ez1z2 − A0 − A1z1 − A2z2]−1

=∞∑

p=−n1

∞∑q=−n2

Tpqz−(p+1)1 z

−(q+1)2

pair (n1, n2) of positive integersn1, n2 suchthatTpq = 0 for p < −n1 and/orq < −n2is called the index of the model.

local stability Seestable state.

local wavelength distance between thephase fronts of a non-planewave signal in-ferred from measurements that are local inspace; 2 pi over the magnitude of the lo-cal propagation constant, which is the gradi-ent of the total phase.See alsowavelength,instantaneous frequency.

locality one of two forms of programmemory relationships.

1. Temporal locality: if an object is beingused, then there is a good chance that theobject will be reused soon.

2.Spatial locality: when an object is beingused, there is a good chance that objects inits neighborhood (with respect to the memorywhere these objects are stored) will be used.

These two forms of locality facilitate theeffective use of hierarchical memory. Reg-isters exploit temporal locality. Caches ex-ploit both temporal and spatial locality. In-terleaved memories exploit spatial locality.See alsosequential locality.

localization refers to the “trapping” of anelectron into a potential well minimum, sothat the wave function ceases to be describ-able by a propagating wave. This localizationcan be “strong” localization such as when anelectron is trapped by an ionized donor orother ionized potential, or “weak” localiza-tion in which it is induced by a “self” inter-ference effect.See alsoweak localization.

lock a synchronization variable, used inshared-memory multiprocessors, that allowsonly one processor to hold it at any one time,thus enabling processors to guarantee thatonly one has access to key data structures orcritical sections of code at any one time.

lock range the range of frequencies in thevicinity of the voltage controlled oscillator


(VCO) free-running frequency over whichthe VCO will, once locked, remain synchro-nized with the signal frequency. Lock rangeis sometimes called tracking bandwidth.

lock-in amplifier a system for detect-ing weak, noisy periodic signals based onSynchronous detection, and incorporating allthe other components necessary for record-ing the amplitude profile of the weak in-coming signal, including input AC amplifier,diode or other detectors, low-pass filter, DCamplifier, and any special filters. Such in-struments are nowadays constructed with in-creasing amounts of digital and computerizedcircuitry, depending on the frequency of op-eration.

lock-out phenomenon exhibited duringchannel switching that results from a fast au-tomatic gain control (AGC) system interact-ing with the horizontal automatic frequencycontrol (AFC), thereby reducing the pull-inrange of the AFC system.

lock-up-free cache See nonblockingcache.

locked-rotor current the current drawnby an induction motor when the shaft is notmoving and rated voltage is applied. Thestarting current is essentially equal to thelocked rotor current and may be as much aseight times the rated current of the machine.

locked-rotor torque the torque producedin an induction motor when the rotor is lockedand rated AC voltage is applied to the stator.

locking Seebus locking.

lockout the condition following faultclearing when the circuit will not attempt areclose. Transformers, generators, and busestypically trip once and lockout immediately.Transmission lines and distribution lines willgenerally attempt one or more recloses, andwill lockout if the fault remains following thelast reclose in the sequence.

lockout relay an auxiliary relay whichis operated by protective relay(s) that inturn opens the appropriate circuit breakers orother fault clearing devices. The lockout re-lay will remain in the trip position until man-ually reset, and is used in protective zoneswhere temporary faults are unusual and thepotential for equipment damage is high.

LOCOS Seelocal oxidation of silicon.

log periodic antenna broadband antennadesigned using physical dimensions (lengths,spacings, diameters, etc.) that vary logarith-mically. The result of such designs is an an-tenna whose performance parameters (e.g.,input impedance) is periodic with respect tothe logarithm of the frequency.

log-likelihood function the likelihoodfunction of y given x is the conditionalPDF, p(y|x). The log-likelihood functionis the logarithm of the likelihood function,log(p(y|x)).

log-normal distribution probability dis-tribution with density

f (x) = 1/(√(2π)σx)e−((logx−µ)2/(2σ2))

whereµ andσ are the mean and standarddeviation of the logarithm.

logarithmic quantization a method fornon-linear scalar quantization where the in-put signal is transformed logarithmically andthen coded using uniform quantization. Thetransformation is utilized to enhance perfor-mance for sources having nonuniform prob-ability distribution, and to give robustness to-wards varying input signal dynamics.

logic analyser a machine that can be usedto send signals to, and read output signalsfrom, individual chips or circuit boards.

logic circuit a circuit that implements alogical function, such as AND, OR, NAND,NOR, NOT, or XOR. (DB)


logic gate a basic building block for logicsystems that controls the flow of pulses.

logic level the high or low value of voltagevariable that is assigned to be a 1 or 0state.

logical operation the machine-level in-struction that performs Boolean operationssuch as AND, OR, and COMPLEMENT.

logical register Seevirtual register.

logical shift a shift in which all bits ofthe register are shifted.See alsoarithmeticshift.

long code in a spread-spectrum system,a (periodic) spreading code (spreading se-quence) with a period (substantially) longerthan a bit duration.See alsoshort code.

long duration Seevoltagevariation.

long integer an integer that has doublethe number of bits as a standard integer ona given machine. Some modern machinesdefine long integers and regular integers tobe the same size.

long-term stability a measure of a powersystem’s long-term response to a disturbanceafter all post-disturbance transient oscilla-tions have been damped out, often associatedwith boiler controls, power plant and trans-mission system protection, and other long-period factors.

longitudinal excitation laser pumpingprocess in which the pump power is intro-duced into the amplifying medium in a direc-tion parallel to the direction of propagation ofthe resulting laser radiation.

longitudinal mode term (somewhat mis-leading) used in referring to the longitudinalstructure or index of the mode of a laser os-cillator.

longitudinal modelocking forcing thelongitudinal modes of a laser oscillator to beequally spaced in frequency and have a fixedphase relationship; useful for obtaining veryshort and intense pulsations.

longitudinal phase velocity phase veloc-ity of microwave propagation in the axial di-rection of a slow-wave structure of travelingwave tube.

longitudinal redundancy check an errorchecking character written at the end of eachblock of information on a track of magnetictape. The character is calculated by countingthe number of one on a track and adding ei-ther an additional one or zero in the characterso that the total number of ones in the blockis even.

longitudinal section electric and mag-netic modes (LSE, LSM) are alternativechoices for the electromagnetic potentials.For a waveguiding structure, the most com-mon choices are potentials directed in thepropagation direction (i.e., TE and TMmodes). Sometimes, especially when lay-ered dielectric are present, it is more con-venient to consider the LSE, LSM potential,perpendicular to the longitudinal section.

look-ahead carry the concept (frequentlyused for adders) of breaking up a serial com-putation in which a carry may be propa-gated along the entire computation into sev-eral parts, and trying to anticipate what thecarry will be, to be able to do the computa-tion in parallel and not completely in series.

look-up table model a model in whichmeasured data is stored in a data base, allow-ing the user to access the data and interpo-late or extrapolate performance based on thatdata. S-parameter data is an example that isencountered frequently in RF and microwavedesign. Sophisticated look-up tables may re-duce the data by polynomial curve fitting andretaining only the coefficients. Look-up table


models generally have a very low modelingvaluation coefficient.

lookahead for conditional branches, astrategy for choosing a probable outcome ofthe decision that must be made at a condi-tional branch, even though the conditions arenot known yet, and initiates “speculative” ex-ecution of the instructions along the corre-sponding control path through the program.If the chosen outcome turns out to be incor-rect, all effects of the speculative instructionsmust be erased (or, alternately, the effects arenot stored until it has been determined thatthe choice was correct). Some lookaheadmechanisms attempt execution of both possi-bilities from a conditional branch, and otherswill cascade lookahead choices from severalconditional branches that occur in close prox-imity in the program flow.

lookbehind for an instruction buffer, ameans of holding recently executed instruc-tions in a buffer within the control unit of aprocessor to permit fast access to instructionsin a loop.

loop (1) a set of branches forming a closedcurrent path, provided that the omission ofany branch eliminates the closed path.

(2) a programming construct in which thesame code is repeated multiple times until aprogrammed condition is met.

loop analysis Seemesh analysis.

loop antenna a two-terminal, thin wireantenna for which the terminals are close toone another and for which the wire forms aclosed path.

loop feeder a closed loop formed by anumber of feeders. The resulting configura-tion allows a load area to be served withoutinterruption if one feeder should fail.

loop filter the filter function that followsthe phase detector and determines the systemdynamic performance.

loop gain the combination of all DC gainsin the PLL.

loop network Seering network.

loop primary feeder a feeder in a distri-bution system which forms a loop around theload area.

loop system a secondary system of equa-tions using loop currents as variables.

loop-set Seecircuit-set.

loosely coupled multiprocessor a systemwith multiple processing units in which eachprocessor has its own memory and commu-nication between the processors is over sometype of bus.

Lorentz force the mutually perpendicu-lar force acting on a current-carrying elementplaced perpendicular to a magnetic field.

Lorentz medium a frequency-dependentdielectric whose complex permittivity is de-scribed by an equation with a second-orderLorentz pole involving a resonant frequencyand a damping constant. Some optical mate-rials and artificial dielectrics may require sev-eral Lorentz poles to describe their behaviorover the frequency band of interest.

Lorentz theorem for isotropic and recip-rocal media, the Lorentz reciprocity theoremstates that, in a source-free region boundedby a surfaceS, the electromagnetic fields pro-duced by two sources, denoted by subscriptsa, b, satisfy the equation∮

S

(Ea × Hb − Eb × Ha) dS = 0

Lorentz, Hendrik Antoon (1853–1928)Born: Arnhem, Holland

Lorentz is best known for his work inelectromagnetic theory. Along with PieterZeeman, he shared the 1902 Nobel Prize forPhysics. His theoretical work was in the ex-tension and refinement of Maxwell’s theory.


Lorentz’s work was an essential cornerstonein Einstein’s, Special Theory of Relativity.

Lorentzian line shape the shape of aspectral line that results from certain physi-cal mechanisms, such as the finite lifetime ofthe upper level of an atomic transition. It ischaracterized by broad wings, that decreaseas the inverse square of the frequency sepa-ration from line center.

Lorentzian lineshape function spectrumof an emission or absorption line that is aLorentzian function of frequency; character-istic spectrum of homogeneously broadenedmedia.

Lorenz, Ludwig Valentin (1829–1891)Born: Elsinore, Denmark

Lorenz was not well known. He did, how-ever, do significant work on electromagnetictheory, on the continuous loading method forcables, and for the acceptance of the ohm asthe resistance standard.

LOS Seeline of sight.

loss (1) decrease of intensity of an electro-magnetic wave due to any of several physicalmechanisms.See alsoattenuation.

(2) a term for electric power which doesnot register on the consumer’s electric me-ter, e.g., through ohmic losses in transmissionlines, iron losses in transformers, or theft.

loss coefficient a factor used in economicdispatch calculations that relates power linelosses to the power output of generatingplants.

loss factor the product of the dielectricconstant and the power factor.

loss of service the complete loss of elec-tric power exclusive of sags, swells, and im-pulses.

loss tangent Seedissipation factor.

loss-of-field relay a protection relay usedto trip a synchronous generator when the ex-citation system is lost. Loss of excitationcauses the generator to run as an inductiongenerator drawing reactive power from thesystem. This can cause severe system volt-age reductions and damage to stator due toexcessive heating.

lossless coding See lossless sourcecoding.

lossless compression compression pro-cess wherein the original data can be re-covered from the coded representation per-fectly, i.e., without loss. Lossless compres-sors either convert fixed length input sym-bols into variable length codewords (Huff-man and arithmetic coding) or parse the in-put into variable length strings and outputfixed-length codewords (Ziv-Lempel cod-ing). Lossless coders may have a static oradaptive probability model of the input data.See alsoentropy coding.

lossless predictive coding a lossless cod-ing scheme that can encode an image at abit rate close to the entropy of themth-orderMarkov source. This is done by exploitingthe correlation of the neighboring pixel val-ues.

lossless source coding source codingmethods (for digital data) where no informa-tion is “lost” in the coding, in the sense thatthe original can be exactly reproduced fromits coded version. Such methods are used, forexample, in computers to maximize storagecapacity.Compare withlossy coding.

lossless source encoding See losslesssource coding.

lossy coding Seelossy source coding.

lossy compression compression whereinperfect reconstruction of the original data isnot possible. Aims to minimize signal degra-dation while achieving maximum compres-


sion. Degradation may be defined accordingto signal error or subjective impairment.

lossy encoding Seelossy source coding.

lossy source coding refers to noninvert-ible coding, or quantization. In lossy sourcecoding information is always lost, and thesource data cannot be perfectly reconstructedfrom its coded representation.Compare withlossless source coding.

lossy source encoding Seelossy sourcecoding.

LOT Seelapped orthogonal transform.

low byte the least-significant 8 bits in alarger data word.

low level waste nuclear waste such asgloves and towels which have comparativelylow radioactivity.

low noise amplifier (LNA) (1) an am-plifier that boosts low-level radio/microwavesignal received without adding substantialdistortions to the signal.

(2) an amplifier in which the primarycause of noise is due to thermally excitedelectrons and is designed to introduce mini-mum internally generated noise.

low order interleaving in memory inter-leaving, using the least significant addressbits to select the memory module and mostsignificant address bits to select the locationwithin the memory module.

low output capacitance Plumbicon tubea picture tube designed to reduce the capaci-tance of the target to ground, resulting in animproved signal-to-noise ratio.

low-pass filter (1) filter exhibiting fre-quency selective characteristic that allowslow-frequency components of an input sig-nal to pass from filter input to output unat-

tenuated; all high-frequency components areattenuated.

(2) a filter that passes signal componentswhose frequencies are small and blocks (orgreatly attenuates) signal components whosefrequencies are large. For the ideal case, ifH(ω) is the frequency response of the filter,thenH(ω) = 0 for |ω| > B, andH(ω) = 1for |ω| < B. The parameterB is the band-width of the filter. A filter whose impulseresponse is a low-pass signal.

low resistance grounded system an elec-trical distribution system in which the neutralis intentionally grounded through a low resis-tance. Low resistance grounding will limitground fault current to a value that signifi-cantly reduces arcing damage but still per-mits automatic detection and interruption ofthe fault current.

low side pertains to the portion of a cir-cuit which is connected to the lower-voltagewinding of a power transformer.

low state a logic signal level that has alower electrical potential (voltage) than theother logic state. For example, the low stateof TTL is defined as being less than or equalto 0.4 V.

low voltage holding coil a holding coilthat keeps the main-line contactor closed onlow voltage conditions. Controllers that con-tain this feature are used in places where themotor is vital to the operation of a process,and it is necessary to maintain control of themotor under low voltage conditions.

low-level transmitter a transmitter inwhich the modulation process takes place at apoint where the power level is low comparedto the output power.

low-level vision the set of visual processesrelated to the detection of simple primitives,describing raw intensity changes and/or theirrelationships in an image.


low-pass equivalent (LPE) model amethod of representing bandpass signals andsystems by low-pass signals and systems.This technique is extremely useful when de-veloping discrete time models of bandpasscontinuous-time systems. It can substan-tially reduce the sampling rate required toprevent aliasing and does not result in anyloss of information. This, in turn, reduces theexecution time required for the simulation.This modeling technique is closely relatedto the quadrature representation of bandpasssignals.

low-pass signal a signal whose Fouriertransform has frequency components that aresmall for frequencies greater than some inter-mediate frequency value. To define mathe-matically, letX(ω) be the Fourier transformof the signal thenX(ω) = 0 for |ω| > B,for someB > 0. In the strict sense, ifB isthe smallest value for which the above holds,thenB is the bandwidth of the signal.

low-power TV (LPTV) a television ser-vice authorized by the FCC to serve specificconfined areas. An LPTV station may typ-ically radiate between 100 and 1000 W ofpower, covering a geographic radius of 10 to15 miles.

low-pressure discharge a discharge inwhich the pressure is less than a torr or afew torrs; low-pressure gases can be easilyexcited, giving spectra characteristic of theirenergy structure.

lower frequency band edge the lowercutoff frequency where the amplitude is equalto the maximum attenuation loss across theband.

lower side frequency the difference fre-quency that is generated during the het-erodyning process or during the amplitude-modulating process. For example, if a 500kHz carrier signal is amplitude-modulatedwith a 1 kHz frequency, the lower side fre-quency is 499 kHz.

LPE model See low-pass equivalentmodel.

LPTV Seelow-power TV.

LQ control Seelinear quadratic control.

LRM calibration Seeline-reflect-matchcalibration.

LRRM calibration See line-reflect-reflect-match calibration.

LRU Seeleast recently used algorithm.

LRU bits a set of bits that record therelative recency of access among pairs ofelements that are managed using an LRUreplacement policy. Ifn objects are be-ing managed, the number of bits required isn(n − 1)/2. Upon each access,n bits areforced into certain states; to check one of theobjects to determine whether it was the leastrecently accessed,n bits need to be exam-ined.

LRU replacement Seeleast recently usedalgorithm.

LRU stack a stack-based data structure toperform the bookkeeping for a least recentlyused (LRU) management policy. An objectis promoted to the top of the stack when it isreferenced; the object that has fallen to thebottom of the stack is the least recently usedobject.

LS algorithm Seeleast squares algorithm.

LSB Seeleast-significant bit.

LSI Seelarge-scale integration.

LTC See load tap changerand tapchanging under load.

LTI Seelinear time invariant system.


LTIL system Seelinear time-invariantlumped-parameter system.

lumen the SI unit of illumination mea-surement. Also, the hollow interior of ablood vessel or airway.

luminance (1) formally, the amount oflight being emitted or reflected by a surface ofunit area in the direction of the observer andtaking into account the spectral sensitivity ofthe human eye.

(2) the amount of light coming from ascene.See alsocandela.

(3) part of the video signal that providesthe brightness information; often designatedas the “Y ” component in RGB systems whereY = .3R + .59G+ .11B.

luminance ratio the ratio between the lu-minance of two areas in the visual field.

luminescence emission of light caused byrelaxation of an electron excited to a higherenergy level. Excitation of the electron mayhave occurred as a result of light absorption(fluorescence or phosphorescence), a chem-ical reaction (chemiluminescence), or fromsome living organisms (bioluminescence).

luminosity the ratio of luminous flux (to-tal visible energy emitted) to the correspond-ing radiant flux (total energy emitted) usuallyin lumens per watt.

luminous efficiency the measure of thedisplay output light luminance for a giveninput power, usually measured in lumens perwatt, which is equivalent to the nit.

lumped element a circuit element of in-ductors, capacitors, and resistors; its dimen-sion is negligible relative to the wavelength.

LVQ Seelearning vector quantization.

Lyapunov, Alexandr M. (1857–1918)also transliterated as Liapunov or Liapunoff.Russian mathematician noted for his contri-

butions to the stability of dynamical systems.His most influential work is titledThe Gen-eral Problem of the Stability of Motion,and itwas published in 1892 by the Kharkov Math-ematical Society. Lyapunov was a studentof Chebyshev at St. Petersburg, and taughtat the University of Kharkov from 1885 to1901. He became academician in appliedmathematics at the St. Petersburg Academyof Sciences.See alsodirect method.

Lyapunov equation in statistics and lin-ear systems theory a set of equation-classesthat describe system evolution. In each case,P,Q are symmetric positive definite;A,Bare arbitrary. The continuous-time algebraicLyapunov equation:

0 = AP + PAT + BQBT .

The discrete-time, time-dependent Lyaponovequation:

P(t+1) = A(t)P (t)AT (t)+B(t)Q(t)BT (t).The discrete-time algebraic Lyanpunov equa-tion:

P = APAT + BQBT .

See alsoRiccati equation.

Lyapunov function corresponding to anequilibrium statexeq ∈ R

n of the systemx =f (t, x) is a continuously differentiable func-tion V = V (t, x) such thatV (t, xeq) = 0,it is positive in a neighborhood ofxeq , thatis, V is positive definite with respect toxeq ,and the time derivative ofV evaluated on thetrajectories of the systemx = f (t, x) is neg-ative semidefinite with respect toxeq . Forthe discrete-time model,x(k + 1) = f (k, x),continuous differentiablity is replaced withcontinuity, and the time derivativeV is re-placed with the first forward difference1V .The existence of a Lyapunov function for agiven equilibrium state implies that this equi-librium is stable in the sense of Lyapunov. Ifthe time derivative, respectively the forwarddifference, ofV is negative definite with re-spect to the equilibrium, then this equilibrium


is asymptotically stable in the sense of Lya-punov.

Lyapunov function candidate for anequilibrium state ofx = f (t, x), its Lyapunovfunction candidate is any continuously dif-ferentiable functionV = V (t, x) that is pos-itive definite with respect to the equilibriumxeq . For an equilibrium state of the discrete-time systemx(k + 1) = f (k, x), a Lyapunovfunction candidate is any continuous functionV = V (k, x) that is positive definite with re-spect to the equilibrium state.

Lyapunov stability also known as stabil-ity in the sense of Lyapunov; concerned withthe behavior of solutions of a system of differ-ential, or difference, equations in the vicinityof its equilibrium states. The concept of sta-bility in the sense of Lyapunov can be relatedto that of continuous dependence of solutionsupon their initial conditions. An equilibriumstate is stable in the sense of Lyapunov if thesystem trajectory starting sufficiently close tothe equilibrium state stays near the equilib-rium state for all subsequent time. Formally,the equilibrium statexeq ∈ R

n of the systemx = f (t, x), or the systemx(k+1) = f (k, x),is stable, in the sense of Lyapunov, if forany ε > 0 there exists aδ = δ(t0, ε) suchthat if the initial conditionx(t0) is within δ-neighborhood ofxeq , that is,

‖x(t0)− xeq‖ < δ ,

then the system trajectoryx(t) satisfies

‖x(t)− xeq‖ < ε

for all t ≥ t0, where‖ · ‖ is any Holder normonR

n.

Lyapunov surface the set of all pointsx ∈<n (n-dimensional real Euclidean space) thatsatisfyV (x) = positive real number, whereV (x) is a Lyapunov function.

Lyapunov’s direct method SeeLyapu-nov’s second method.

Lyapunov’s first method the method thatallows one to assess the stability status of anequilibrium point of a nonlinear system basedon stability investigation of the equilibriumpoint of the linearized version of that system.This method is also called Lyapunov’s indi-rect method.

Lyapunov’s indirect method SeeLyapu-nov’s first method.

Lyapunov’s second method the methodof stability assessment that relies on the use ofenergy-like functions without resorting to di-rect solution of the associated evolution equa-tions. This is also called Lyapunov’s directmethod.

LZ77 refers to string-based compressionschemes based on Lempel and Ziv’s 1977method. An input string of symbols thatmatches an identical string previously (andrecently) transmitted is coded as an offsetpointer to the previous occurrence and a copylength.

LZ78 refers to string-based compressionschemes based on Lempel and Ziv’s 1978method. A dictionary of prefix strings isbuilt at the encoder and decoder progres-sively, based on the message. The encodersearches the dictionary for the longest stringmatching the current input, then encodes thatinput as a dictionary index plus the literalsymbol which follows that string. The dic-tionary entry concatenated with the literal isthen added to the dictionary as a new string.


MM common notation for the number ofmodes in astep index fiber, given by

M = V 2

2

whereV is the fiber parameter.

M (mega) abbreviation for 1,048,576 (notfor 1 million).

M-algorithm reduced-complexity breadth-first tree search algorithm, in which at mostM tree nodes are extended at each stage ofthe tree.

m-ary hypothesis testing the assessmentof the relative likelihoods ofM hypothesesH1, H2, . . . , HM . Normally we are givenprior statisticsP(H1), . . . , P (HM) and ob-servationsy whose dependencep(y|H1),

. . . , p(y|HM) on the hypotheses are known.The solution to the hypothesis testing prob-lem depends upon the stipulated criterion;possible criteria include maximizing the pos-terior probability (MAP) or minimizing theexpected “cost” of the decision (a costCijis assigned to the selection of hypothesisjwhen i is true). See alsobinary hypothesistesting. See alsoconditional statistic, a pri-ori statistics,a posteriori statistics,maximuma posteriori estimator.

m-phase oscillator See multi-phaseoscillator.

m-sequence maximal length sequence. Abinary sequence generated by a shift registerwith a given number of stages (storage el-ements) and a set of feedback connections,such that the length of the sequence periodis the maximum possible for shift registerswith that number of stages over all possible

sets of feedback connections. For a shift reg-ister with n stages, the maximum sequenceperiod is equal to 2n − 1.

Mach band a perceived overshoot (lighterportion) on the light side of an edge and anundershoot (darker portion) on the dark sideof the edge. The Mach band is an artifactof the human visual system and not actu-ally present in the edge.See alsobrightness,simultaneous contrast.

machine code the machine format of acompiled executable, in which individual in-structions are represented in binary notation.

machine interference the idle time ex-perienced by any one machine in a multiple-machine system that is being serviced by anoperator (or robot) and is typically measuredas a percentage of the total idle time of all themachines in the systems to the operator (orrobot) cycle time.

machine language the set of legal instruc-tions to a machine’s processor, expressed inbinary notation.

machine vision Seerobot vision.

macro Seemacroprogram.

macro cell a cell in a cellular communica-tion system that has a size that is significantlylarger than the cell size of a typical cellularsystem. Macro cells are sometimes designedwith the base station being located in a satel-lite. Macro cells typically cover large areassuch as rural areas in cases where the userdensity is low.

macro diversity a diversity technique,used in a cellular communication system,that is based on the transmission of multi-ple copies of the same signal from transmit-ters that have a separation that is a signifi-cant fraction of the coverage area of a cell.Macro diversity also refers to the concept ina cellular system where coverage at a par-


ticular location may be obtained from mul-tile transmitters, typically utilizing differentfrequency channels. Macro diversity is typi-cally utilized to overcome the type of signalfading typically known as shadow fading.

macro shadowing shadowing due to theobstacles in the propagation path of the radiowave.See alsomicro shadowing.

macroinstruction a short code-like text,defined by the programmer, that the assem-bler or compiler will recognize and that willresult in an inline insertion of a predefinedblock of code into the source code.

macroprogram a sequence of macroin-structions.

macrotiming diagram a graphical dis-play showing how the waveforms vary withtime but with a time scale that does not havesufficient resolution to display the delays in-troduced by the individual basic elements ofthe digital circuit.

MAD Seemaximal area density.

made electrode a ground electrode for alightning rod which has been especially con-structed for the purpose, as opposed to usinga building frame or water pipe for the pur-pose.

magic T Seemagic tee.

magic T junction Seemagic tee junction.

magic tee a combination of E-plane (se-ries) and H-plane (shunt) tees forming a hy-brid waveguide junction. Typically used tosplit or couple two microwave signals that arein or out of phase in the same wave guide.Also called a hybrid tee.

magic tee junction a four-port microwavedevice, that couples the input signal at port 1equally into ports 2 and 3, but not into port4. Input signal at port 3 is coupled equally

into ports 1 and 4, but not into port 2, etc. Atmicrowave frequency, the waveguide junc-tion that makes up this device involves threemagnetic plane arms and one electric fieldplane arm resembling T and hence is calledmagic tee junction.

maglev Seemagnetic levitation.

magnet any object that can sustain an ex-ternal magnetic field.

magnetic actuator any device using amagnetic field to apply a force.

magnetic bearing a component of a ma-chine that uses magnetic force to providenon-contact support for another componentmoving relative to it.

magnetic bias a constant magnetic fieldon which is superimposed a variable, oftensinusoidal, perturbation magnetic field in de-vices like magnetic bearings.

magnetic brake any device using a mag-netic field to retard motion.

magnetic charge density a fictitioussource of the electromagnetic field that quan-tifies the average number of discrete mag-netic charges (also fictitious) per unit volume.The magnetic charge density is often intro-duced in problems where duality and equiv-alence concepts are employed.

magnetic circuit the possible flux pathswithin a system consisting of a source of flux(electromagnets, permanent magnets), per-meable flux carrying materials (steel, nickel)and non-flux carrying materials (aluminum,air).

magnetic clamp a device employing amagnetic field to deliver a clamping action.

magnetic core memory a persistent, di-rectly addressable memory consisting of an


array of ferrite toruses (cores) each of whichstores a single bit. Now obsolete.

magnetic current density a fictitioussource vector in electromagnetics that quan-tifies the amount of magnetic charge (alsofictitious) crossing some cross-sectional areaper unit time. The magnetic current densityis often introduced in problems where dual-ity and equivalence concepts are employed.The direction of the magnetic current densityis in the direction of magnetic charge motion.SI units are volts per square meter.

magnetic damper any device using amagnetic field to damp motion.

magnetic dipole an arrangement of oneor more magnets to form a magnet systemthat produces a magnetic field with one pairof opposite poles.

magnetic disk a persistent, random-access storage device in which data is storedon a magnetic layer on one or both surfacesof a flat disk.See alsohard disk, floppy disk,diskette.

magnetic drive See magnetic torquecoupling.

magnetic drum a persistent storage de-vice in which data is stored on a magneticlayer on the surface of a cylinder. Now ob-solete.

magnetic field magnetic force field wherelines of magnetism exist.

magnetic field integral equation (MFIE)used in the method of moments to solve forthe surface current on an object in terms of theincident magnetic field. The MFIE is validonly for closed surfaces.

magnetic field intensity a force field thatis a measure of the magnitude and directionof the force imparted upon an elemental cur-rent normalized to the elemental current’s

value. Depends on material characteristics.The units are amperes per meter.

magnetic flux the integral of the compo-nent of magnetic flux density perpendicularto a surface, over the given surface.

magnetic flux density a vector quantify-ing a magnetic field, so that a particle carry-ing unit charge experiences unit force whentraveling with unit velocity in a direction per-pendicular to the magnetic field characterizedby unit magnetic flux density. It has the unitsof volt-seconds per square meter in the SIsystem of units.

magnetic head Seeread/write head.

magnetic induction the flux densitywithin a magnetic material when driven byan external applied field or by its self demag-netizing field, which is the vector sum of theapplied field and the intrinsic induction.

magnetic leakage Seeleakage.

magnetic length the effective distance be-tween the north and south poles within a mag-net, which varies from 0.7 (alnico) to 1.0 (Nd-FeB, SmCo, hard ferrite) times the physicallength of the magnet.

magnetic levitation (1) noncontact sup-port of an object using magnetic forces. Ab-breviated as maglev.

(2) a method of melting metals withoutcontacting a surface. A cone-shaped high-frequency coil produces eddy currents in themetal which are strong enough to both sus-pend and melt it, generally in a neutral atmo-sphere.

(3) one of several techniques of suspend-ing a driveshaft within a bearing so that nocontact is made between the shaft and othersurfaces.Seemagnetic bearing.

(4) one of several techniques for suspend-ing a railroad train above its tracks so thatwheels are not needed. Typically, supercon-ducting magnets are needed, and propulsion


is by a linear induction motorcfwhose arma-ture lies along the rails.

magnetic loss losses in magnetic flux in amagnetic circuit, primarily due to magneticleakage and fringing.See alsocore loss.

magnetic moment for a current-carryingcoil in an external magnetic field, the ratioof the torque sensed by the coil to the fluxdensity of the external field. In permanentmagnets, the product of the polar flux and themagnetic length; the product of the intrinsicflux density and the magnet volume.

magnetic monopole a magnet system thatproduces a magnetic field of a single polarity.Although nonexistent, may be approximatedby one pole of a very long magnet.

magnetic motor starter motor starter thatuses electromechanical devices such as con-tactors and relays.

magnetic orientation the preferred direc-tion of magnetization for an anisotropic mag-netic material.

magnetic overload an overload sensor ina motor controller used to shut off the motorin event of an over current condition. With amagnetic overload, the sensor uses a mag-netic coil to sense the overload condition,then trips the overload contact(s).See alsooverload heater, overload relay.

magnetic permeability tensor relation-ship between the magnetic field vector andthe magnetic flux density vector in a mediumwith no hysteresis; flux density divided bythe magnetic field in scalar media.

magnetic polarization vector an auxil-iary vector in electromagnetics that accountsfor the presence of atomic circulating cur-rents in a material. Macroscopically, themagnetic polarization vector is equal to theaverage number of magnetic dipole momentsper unit volume.

magnetic quardrupole Seequadrupole.

magnetic recording air gap term refer-ring to two aspects of a magnetic recordingsystem.

1. The gap between the poles of aread/write head is often referred to as the “airgap.” Even though filled with a solid, it ismagnetically equivalent to an air gap. With arecorded wavelength ofλ, and a head gap ofd, the read signal varies as sin(πx)/x wherex = λ/d. The gap must then haved < λ/2for reliable reading.

2. The space between the head and therecording surface is also referred to as an “airgap.” The signal loss for a head at heighth

from the surface is approximately 55h/λ dB;if d = λ/5 the loss is 11 dB.

High recording densities therefore requireheads “flying” very close to the recording sur-face and at a constant height – 0.1µm orless in modern disks. Such separations areachieved by shaping the disk head so that itsaerodynamics force it to fly at the correct sep-aration.

magnetic recording code method usedto record data on a magnetic surface such asdisk or tape. Codes that have been used in-clude return-to-zero (RZ), in which two sig-nal pulses are used for every bit — a changefrom negative to positive pulse (i.e., magneti-zation in the “negative” direction) is used fora stored 0, a change from a positive to a neg-ative pulse (i.e., magnetization in the “pos-itive” direction) is used for a stored 1, anda return to the demagnetized state is madebetween bits; non-return-to-zero (NRZ), inwhich a signal pulse occurs only for a changefrom 1 to 0 or from 0 to 1; non-return-to-zero-inverted (NRZI), in which a positive or nega-tive pulse is used for — a sign change occursfor two consecutive 1s — and no pulse is usedfor a 0; double frequency (DF) (also known asfrequency modulation (FM)), which is simi-lar to NRZI but which includes an interleavedclock signal on each bit cell; phase encoding(PE), in which a positive pulse is used for a 1and a negative pulse is used for a 0; return-to-


bias (RB), in which magnetic transitions aremade for 1s but not for 0s, as in NRZI, butwith two transitions for each 1; and modified-return-to-bias (MRB), which is similar to RB,except that a return to the demagnetized stateis made for each zero and between two con-secutive 1s.

RZ is self-clocking (i.e., a clock signalis not required during readout to determinewhere the bits lie), but the two-signals-per-bitresults in low recording density; also, a 0 can-not be distinguished from a “dropout” (i.e.,absence of recorded data). NRZ is not self-clocking and there requires a clocking sys-tem for readout; it also requires some meansto detect the beginning of a record, does notdistinguish between a dropout and a storedbit with no signal, and if one bit is in error,then all succeeding bits will also be in errorup to the next signal pulse. NRZI has simi-lar properties to NRZ, except that an error inone bit does not affect succeeding bits. PE isself-clocking.

magnetic resonance imaging (MRI) (1)a form of medical imaging with tomographicdisplay that represents the density and bond-ing of protons (primarily in water) in the tis-sues of the body, based upon the ability ofcertain atomic nuclei in a magnetic field toabsorb and reemit electromagnetic radiationat specific frequencies. Also called nuclearmagnetic resonance.

(2) an imaging modality that uses a pulsedradio frequency magnetic field to selectivelychange the orientation of the magnetizationvectors of protons within the object understudy. The change in net magnetic moment asthe protons relax back to their original orien-tation is detected and used to form an image.

magnetic saturation the condition in amagnetic material when an increase in themagnetizing force does not result in a use-ful increase in the magnetic induction of thematerial.

magnetic separator a device employingmagnetic fields to separate magnetic materi-als from nonmagnetic ones.

magnetic stabilization the act of pur-posely demagnetizing a magnet with reversefields or a change in temperature so that noirreversible losses are experienced when themagnet operates under similar conditions inthe field.

magnetic susceptibility the ratio of themagnetization to the applied external field.

Tensor relationship between the magneticfield vector and the magnetization vector ina medium with no hysteresis; magnetizationdivided by the permeability of free space andthe magnetic field in scalar media. It is anindicator of how easily a material is magne-tized and has no units in the SI system of units(pure number).

magnetic suspension See magneticlevitation.

magnetic tape a polyester film sheetcoated with aferromagneticpowder, whichis used extensively in auxiliary memory. Itis produced on a reel, in a cassette, or in acartridge transportation medium. Often usedfor backups.

magnetic torque coupling any device uti-lizing a magnetic field to transmit torque.

magnetic vector potential an auxiliaryfield used to simplify electromagnetic com-putations. This field satisfies a wave equa-tion, the curl of this field is related to themagnetic field intensity vector field, and thedivergence of this field is specified by somegage which is to be specified in each problem.

magnetic-tape track each bit positionacross the width of a magnetic tape read/writehead, running the entire length of the tape.

magnetization curve See hysteresiscurve.


magnetizing current the current requiredto magnetize the different parts of a magneticcircuit. It is calculated as the ratio of the totalmagnetomotive force (F ) and the number ofturns (N ). More or less in transformers, andAC synchronous and induction machines, themagnetizing current is the current through themagnetizing inductance. Denoted byIm, itis calculated as the ratio of the induced EMFacross the magnetizing inductance to its mag-netizing reactanceXm.

magneto plasma a plasma medium thatin the presence of a static magnetic field be-haves like an anisotropic dielectric mediumwhose dielectric function is a tensor.

magnetohydrodynamic MHD machinea form of electric machine in which a streamof electrically conductive gas or liquid ispassed through pairs of orthogonally posi-tioned magnetic poles and electrodes. In anMHD generator, the fluid is forced by theprime mover to produce a DC across the elec-trodes. In the MHD motor, a current acrossthe electrodes through the fluid forces thestream to flow.

magneto-hydrodynamic (MHD) genera-tor a heat-to-electricity conversion devicewith an intermediate kinetic energy stage.In the MHD generator, a partially conduct-ing gas is heated by a fuel-fired source ora nuclear reactor to convert the heat energyto kinetic energy, and then passed betweenthe poles of an electromagnet, which con-verts some of the kinetic energy to electri-cal energy. The electrical energy is collectedthrough a pair of electrodes situated in thegas channel.

magneto-optic Bragg cell a magneticallytunable microwave signal processing devicethat uses optical Bragg diffraction of lightfrom a moving magneto-optic grating gen-erated by the propagation of magnetostaticwaves within the magnetic medium.

magneto-optic modulator any of a classof light modulators that use magneto-opticeffects, such as Faraday rotation for lightmodulation.

magneto-optical Kerr effect the rotationof the plane of polarization of a linearly po-larized beam of light upon reflection fromthe surface of a perpendicularly magnetizedmedium.

magneto-resistive head Seeread/writehead.

magnetoimpedance the change in imped-ance of a ferromagnetic conductor experienc-ing a change in applied magnetic field.

magnetomotive force (MMF) a magneticcircuit term referring to that phenomenon thatpushes magnetic flux through the reluctanceof the circuit path. MMF is analogous tothe concept of electromotive force (voltage)in an electric circuit. For a magnetic corewith a single coil ofN turns, carrying currentI , the MMF isNI , with units of amperes(sometimes expressed as ampere-turns).

magnetoresistance the change in electri-cal resistance in a conducting element expe-riencing a change in applied magnetic field.This is most pronounced when the magneticfield is perpendicular to current flow.

magnetostriction a change in the lengthof a ferromagnetic material as the fluxchanges under the influence of an appliedmagnetic field, or resulting from domain for-mation after cooling from above Curie tem-perature. In an AC device, the steel in thecore expands and contracts twice each cy-cle, creating audible noise (e.g., transformerhum).

magnetostrictive smart material one ofa class of materials with self-adaptively mod-ifiable elastic properties in response to a mag-netic field applied in proportion to sensedstress–strain information.


magnetotransport motion of electrons orholes in a conducting material in the presenceof an applied magnetic field.

magnetron any arrangement of magnetsin a sputter deposition or etch system thatprovides the magnetic field required to trapelectrons in closed loops near the cathode,thus enhancing deposition/etch rates.

magnitude (1) the absolute value of ascalar.

(2) the norm of a vector, i.e., the squareroot of the sum of the squares of the vectorcomponents.

magnitude response the magnitude ofthe frequency response of a system.Seefrequency response.

magnitude scaling a procedure for chang-ing the values of network elements in a filtersection without affecting the voltage ratio orcurrent ratio transfer function of the section.Resistors and inductors are multiplied by theparametera, while capacitors are divided bya. If 1 < a < ∞, then the magnitudes of theimpedances are increased.

magnitude squared coherence (MSC) ameasure of the degree of synchrony betweentwo electrical signals at specific frequencies.

magnon a polariton in a magneticmedium.

Mahalanobis distance a distance mea-sure used in certain decision rules.

Letp(x, ci) be the probability distributionthat patternx belong to classci . A com-monly used decision rule is that of lookingfor arg maxi p(x, ci). Furthermore, let us as-sume thatp(x|ci) is Gaussian, that is

p(x|ci) = 1√(2π)n detRi

exp

(−1

2(x − µi)

′ R−1i (x − µi)

)

and that∀i = 1, . . . , n Ri = R (the samecovariance matrices). Hence, the previ-ously defined decision rule reduces straight-forwardly to a distance evaluation where onelooks for arg maxi (x − µi)

′R−1i (x − µi).

This distance is commonly referred to as theMahalanobis distance.See alsoweightedEuclidean distance.

MAI Seemultiple access interference.

mailbox an operating system abstractioncontaining buffers to hold messages. Mes-sages are sent to and received from the mail-box by processes.

main beam in antenna theory, the direc-tion in which the global maximum of the ra-diation pattern occurs.

main memory the highest level of mem-ory hierarchy.

main switch a switch which controls allpower to a building’s wiring or other electricinstallation.

mains voltage European term for the volt-age at the secondary of the distribution trans-former.

mainframe a large centralized machinethat supports hundreds of users simultane-ously.

maintainability the probability that an in-operable system will be restored to an oper-ational state within the timet .

maintenance the changes made on a sys-tem to fix errors, to support new require-ments, or to make it more efficient.

major hysteresis loop for a magnetic ma-terial, the loop generated as intrinsic or mag-netic induction (Bi or B) is plotted with re-spect to applied field (H ) when the materialis driven from positive saturation to negative


saturation and back, showing the lag of in-duction with respect to applied field.

major orbit a larger helical orbit of anelectron beam in a gyrotron.

majority carrier an electron in an n-typeor a hole in a p-type semiconductor.

majority-logic decoding a simple, andin general suboptimal, decoding method forblock and convolutional codes based on theorthogonality of the parity-check sums.

male connector a connector presentingpins to be inserted into a corresponding fe-male connector that presents receptacles.

MAN Seemetropolitan-area network.

management by exception Seecoordina-tion by exception.

Manhattan distance See city-blockdistance.

manipulability measure is a kind of dis-tance of the manipulator from singular con-figuration. Mathematically it is defined asfollows: W(q) = √

det(J (q)J T (q)), whichvanishes at a singular configuration. It isclear that by maximizing this measure, re-dundancy is exploited to move away fromsingularities. Partial derivative of the manip-ulability measure with respect to a vector ofgeneralized positions allows to define an ar-bitrary vector of joint velocities in the inversekinematics problem for redundant manipula-tors.

manipulated input a quantity influencingthe controlled process from outside, availableto the controller and used to meet the controlobjectives; attributes of the inflowing streamsof material, energy, or information may serveas manipulated inputs. A manipulated in-put is defined by a continuous trajectory overgiven time interval or by a sequence of values

at given time instants. Also known as controlinput.

manipulator workspace a manipulatorworkspace defines all existing manipulatorpositions and orientations that can be ob-tained from the inverse kinematics problem.The lack of a solution means that the manip-ulator cannot attain the desired position andorientation because it lies outside of the ma-nipulator’s workspace.

Manley–Rowe criteria SeeManley–Rowerelations.

Manley–Rowe relations relations amongthe intensities of optical fields interacting in anonlinear material. These can be understoodin terms of the discrete nature of the trans-fer of energy in terms of the emission andabsorption of photons.

mantissa the portion of a floating-pointnumber that represents the digits.See alsofloating-point representation.

manually-controlled shunt capacitors abank of shunt capacitors that are controlledvia SCADA signals from an operating cen-ter as opposed to local automatic control byvoltage sensing.

MAP Seemaximum a posteriori estimator.

MAP estimator Seemaximum a posterioriestimator.

mapping the assignment of one locationto a value from a set of possible locations.Often used in the context of memory hierar-chies, when distinct addresses in a level ofthe hierarchy map a subset of the addressesfrom the level below.

MAR Seememory address register.

Marconi, Guglielmo (1874–1937) Born:Bologna, Italy


Marconi is best known for work that led tothe development of the commercial radio in-dustry. Marconi’s many experiments with ra-dio waves (long wavelength electromagneticradiation) led to many communications inno-vations. Marconi traveled to England to findsupport for his ideas, and there formed theMarconi Wireless Telegraph Company, Ltd.In 1909, Marconi shared the Nobel Prize forphysics with K. F. Braun. Marconi’s most fa-mous demonstration came in 1901 when hewas able to send the first Morse-coded mes-sage from the Pondhu, Cornwall, England toSt. Johns, Newfoundland, Canada.

Markov chain a particular case of aMarkov process, where the samples take onvalues from a discrete and countable set.Markov chains are useful signal generationmodels for digital communication systemswith intersymbol interference or convolu-tional coding, and Markov chain theory isuseful in the analysis of error propagation inequalizers, in the calculation of power spectraof line codes, and in the analysis of framingcircuits.

Markov model a modeling techniquewhere the states of the model correspond tostates of the system and transitions betweenthe states in the model correspond to systemprocesses.

Markov process a discrete-time randomprocess,{9k} , that satisfiesp(ψk+1|ψk,ψk−1, . . .) = p(ψk+1|ψk). In other words,the future sampleψk+1 is independent of pastsamplesψk−1, ψk−2, . . . if the present sam-ple9k = ψk is known.

Markov random field an extension of thedefinition of Markov processes to two dimen-sions. Consider any closed contour0, anddenote by0i and0o the points interior andexterior to0, respectively. Then a processψ is a Markov random field (MRF) if, con-ditioned onψ(0), the setsψ(0i) andψ(0o)

are independent. That is,

p(ψ(0i), ψ(0o)|ψ(0)) =p(ψ(0i)|ψ(0)) · p(ψ(0o)|ψ(0)).

See alsoMarkov process, conditional statistic.

Marr-Hildreth operator (1) edge-detec-tion operator, also called Laplacian-of-Gaussian or Gaussian-smoothed-Laplacian,defined by

∇2G = − 1√2πσ 3

(1 − x2 + y2

σ 2)e

−(x2+y2)2σ2 .

It generates a smoothed isotropic secondderivative. Zero crossings of the output cor-respond to extrema of first derivative and thusinclude edge points. (2) The complete edgedetection scheme proposed by Marr and Hil-dreth, including use of the∇2G operator atseveral scales (i.e., gaussian variances), andaggregation of their outputs.

Marx generator a high-voltage pulsegenerator capable of charging capacitors inparallel and discharging them in series.

maser acronym for microwave amplifica-tion by stimulated emission of radiation.

maser amplifier usually refers to amedium that amplifies microwaves by theprocess of stimulated emission; sometimesrefers to amplification of some other field(nonoptical electromagnetic, phonon, exci-ton, neutrino, etc.) or some other process(nonlinear optics, Brillouin scattering, Ra-man scattering, etc.).

maser oscillator oscillator usually pro-ducing a microwave frequency output andusually based on amplification by stimulatedemission in a resonant cavity.

mask (1) in digital computing, to specifya number of values that allow some entities ina set, and disallow the others in the set, frombeing active or valid. For example, maskingan interrupt.


(2) for semiconductor manufacturing, adevice used to selectively block photolitho-graphic exposure of sensitized coating usedfor preventing a subsequent etching processfrom removing material. A mask is analo-gous to a negative in conventional photogra-phy.

(3) a glass or quartz plate containing in-formation (encoded as a variation in transmit-tance and/or phase) about the features to beprinted. Also called a photomask or a reticle.

(4) in image processing, a small set of pix-els, such as a 3×3 square, that is used to trans-form an image. Conceptually, the mask iscentered above every input pixel, each pixelin the mask is multiplied by the correspond-ing input pixel under it and the output (trans-formed) pixel is the sum of these products. Ifthe mask is rotated 180◦ before the arithmeticis performed, the result is a 2-D convolutionand the mask represents the impulse responsefunction of a linear, space-invariant system.Also called a kernel.See alsoconvolution,impulse response, kernel.

mask aligner a tool that aligns a pho-tomask to a resist-coated wafer and then ex-poses the pattern of the photomask into theresist.

mask biasing the process of changing thesize or shape of the mask feature in order forthe printed feature size to more closely matchthe nominal or desired feature size.

mask blank a blank mask substrate(e.g., quartz) coated with an absorber (e.g.,chrome), and sometimes with resist, and usedto make a mask.

mask linearity the relationship of printedresist feature width to mask feature width fora given process.

mask programming programming asemiconductor read-only-memory (ROM)by modifying one or more of the masks usedin the semiconductor manufacturing process.

mask set consists of the dozen or so(varies with process and company) individualmasks that are required to complete a MMICwafer fabrication from start to finish. Exam-ples of masks or mask levels are “first levelmetal” (defines all the primary metal struc-ture on the circuit), “capacitor top plate” (de-fines the pattern for the metal used to form thetop plate of MIM capacitors), and “dielectricetch” (defines areas where dielectric (insula-tor) material will be removed after coatingthe entire wafer with it).

maskable interrupt interrupt that can bepostponed to permit a higher-priority inter-rupt by setting mask bits in a control register.See alsononmaskable interrupt.

masking a phenomenon in human visionin which two patternsP1 andP1 +P2 cannotbe discriminated even thoughP2 is visiblewhen seen alone.P1 is said to maskP2.

mass storage a storage for large amountsof data.

massively parallel architecture a com-puter system architecture characterized bythe presence of large numbers of CPUs thatcan execute instructions in parallel. Thelargest examples can process thousands ofinstructions in parallel, and provide efficientpathways to pass data from one CPU to an-other.

massively parallel processor a systemthat employs a large number, typically 1000or more, of processors operating in parallel.

master the system component responsi-ble for controlling a number of others (calledslaves).

master boot record a record of the diskcontaining the first code and table that areloaded at the bootstrap of the computer. It isread even before the partition table sector.


master control relay (MCR) used inprogrammable logic controllers to secure en-tire programs, or just certain rungs of a pro-gram. An MCR will override any timer con-dition, whether it be time-on or time-off, andplace all contacts in the program to a safeposition whenever conditions warrant.

master copy in coherence protocols, thecopy of the object that is guaranteed to holdthe “correct” contents for the object. Coher-ence protocols can be designed around thetagging of master copies. The master copycan be read (rather then the copy in slowermemory) to speed program execution, and itcan be written, provided that all other copiesare invalidated.See alsoMESI protocol.

master-oscillator-power-amplifier(MOPA) laser system in which the out-put from a highly stabilized low-power laseroscillator is amplified by one or more high-power laser amplifiers.

master–slave flip-flop a two-stage flip-flop in which the first stage buffers an inputsignal, and on a specific clock transition thesecond stage captures and outputs the stateof the input.

matched filter a filter matched to a certainknown signal waveform. The matched filteris the complex conjugate of the known sig-nal waveform. If the known signal waveformfalls in the matched filter, the output sampleat the optimum sampling instant gives the re-ceived signal energy.

matched load load that does not reflectany energy back into the transmission line.It could be a load equal to the characteristicimpedance of the transmission line or a struc-ture with electromagnetic absorbing proper-ties.

matched uncertainty See matchingcondition.

matching when referring to circuits, theprocess by which a network is placed be-tween a load and a transmission line in or-der to transform the load impedance to thecharacteristic impedance of said line and thuseliminate the presence of standing waves onthe line.

matching condition the condition that re-quires that the uncertainties affect the plantdynamics the same way as the control inputdoes. For example, in the following model ofa dynamical system, the matching conditionis satisfied,

x = Ax + B(u + d),

where the vector functiond models matcheduncertainties.

matching conditions conditions imposedon the structure of the uncertainty, which maybe viewed as the assumption that uncertaintydrives the state equations not stronger thana control variable. For linear systems, thematching conditions are given as the assump-tions imposed on the perturbations of systemmatrices1A,1B and disturbancesCv(t) inthe following way:

1A = BD; 1B = BG; C = BF

whereD,G,F are unknown but boundedmatrices andB is an input matrix. It enablesto describe the uncertain linear system by thefollowing state equation:

x = Ax + B(u+ e)

where e is the entire system uncertaintymoved into the input of the system.

matching elements elements such asposts, screws, etc., often used in order toachieve the desired reflection coefficient ofa given microwave component.

matching network an electric circuit de-signed to maximize the transfer of electricpower from an electrical source to an elec-trical load. Maximum power is transferred


from a source with an output impedancethat is the complex conjugate of the in-put impedance of the load it is driving. Amatching network is connected between asource and its load. Its input impedance isthe complex conjugate of the source outputimpedance, while its output impedance is thecomplex conjugate of the load impedance,thereby matching the source to the load andensuring maximum transfer of power.

matching stub matching technique thatemploys short-circuited (or open-circuited)sections of transmission lines as reactive el-ements.

material dispersion wavelength depen-dence of the pulse velocity. It is caused by therefractive index variation with wavelength ofglass.

math coprocessor a separate chip, addi-tional to the CPU, that offloads many of thecomputation-intensive tasks from the CPU.

mathematical modeling a mathematicaldescription of the interrelations between dif-ferent quantities of a given process. In par-ticular, a mathematical description of a rela-tion between the input and output variables ofthe process.See alsotruth modelanddesignmodel.

mathematical morphology an alge-braic theory of non-linear image transfor-mations based on set-theoretical (or lattice-theoretical) operations, which is generallyconsidered as a counterpart to the classicallinear filtering approach of signal process-ing. This methodology in image processingarose in 1964 through the works of Math-eron and Serra at Fontainebleau (France),was developed in the seventies by Sternbergat Ann Arbor (Mich.), and became interna-tionally known in the eighties. It has beensuccessfully applied in various fields requir-ing an analysis of the structure of materi-als from their images, for example biomedi-cal microscopy, stereology, mineralogy, pet-

rography. Seeclosing, dilation, erosion,morphological operator, opening.

matrix an n by m matrix is an array ofn×mnumbers of heightnand widthm repre-senting a linear map from anm-dimensionalspace into ann-dimensional space. An ex-

ample of a 2× 2 real matrix isA =[

2 35 6

].

See alsocirculant matrix, Hermitian matrix,orthogonal matrix, positive definite matrix,positive semi-definite matrix, singularmatrix, Toeplitz matrix.

matrix addressing the control of a dis-play panel consisting of individual light-producing elements by arranging the con-trol system to address each element in arow/column configuration.

matrix element in quantum mechanics,the expectation value of a quantum mechan-ical operator that is associated with a partic-ular pair of basis states. The term matrix el-ement derives from the fact that the expecta-tion values associated with all possible pairsof states can be written as a matrix.

matrix of configuration of information sys-tem in Pawlak’s information systemS =(U,A) whose universeU hasn membersxiand the setA consists of attributesaj , the el-ements of the universe are linked with eachother. Connections between the elements ofthe universeU may be given, for example,by a function

ϕ : U × U → {−1,0,1}.The pair (U, ϕ) is called the configurationof the information systemS. The matrix ofthe information systemS is ann× n matrixwhose elements are

cij = ϕ(xi, xj

).

matrix optics formalism for the analy-sis and synthesis of optical systems in whicheach system element is represented by a ma-


trix, the overall system being representableas a product of the elemental matrices.

matrixing in a color television trans-mitter, the process of converting the threecolor signals (red, green, blue) into the color-difference signals that modulate the chromi-nance subcarrier. In a color TV receiver,the process of converting the color-differencesignals into the red, green, and blue signals.

Mauchley, John Graham (1907–1980)Born: Cincinnati, Ohio, U.S.A.

Mauchley is best known as one of the de-signers of ENIAC, an early electronic com-puter. It was Mauchley, in 1942, who wrote aproposal to the Army for the design of a cal-culating machine to calculate trajectory ta-bles for their new artillery. Mauchley wasa lecturer at the Moore School of ElectricalEngineering at the time. The school wasawarded the contract and he and J. Pres-per Eckert were the principal designers ofENIAC. They later went on to form their owncompany. Mauchley was the software engi-neer behind the development of one of thefirst successful commercial computers, UNI-VAC I.

max operation an operation on two ormore variables where the resultant value isformed by taking the largest value, or maxi-mum, among these variables.

Max–Lloyd scalar quantization SeeLloyd–Max scalar quantization.

Max–Lloyd SQ SeeLloyd–Max scalarquantization.

max-min composition a frequently usedmethod of composition of two fuzzy relationswhich, as the name implies, makes use of themin operation followed by the max operation.

maximal area density (MAD) for a mag-netic disk, the maximum number of bits thatcan be stored per square inch. Computed by

multiplying the bits per inch in a disk tracktimes the number of tracks per inch of media.

maximally flat a circuit response in whichthe low-pass prototype attenuation loss (ex-pressed in decibels) has a smooth response.

maximally flat delay (MHD) filter a filterhaving a time delay that is as flat (constant) aspossible versus frequency while maintaininga monotonic characteristic.

maximum a posteriori probability theprobability of a certain outcome of a randomvariable given certain observations related tothe random variable, i.e.,x was transmittedandy was received, thenP(x|y) is thea pos-teriori probability. The maximuma poste-riori probability is found by considering allvalid realizations ofx.

For example, between two stringsA andB it is defined as

DMPP = Pr{B|A}

wherePr(B|A) is the probability thatA ischanged intoB. Sometimes called maximumposterior probability (MPP) even though,strictly speaking, it is a similarity rather thandistance measure.

maximum a posteriori (MAP) estimatorto estimate a randomx by maximizing itsposterior probability; that is,

xMAP

= argx maxp(x|y),

wherey represents an observation.x is ex-plicitly modeled as a random quantity withknown prior statistics.See alsomaximumlikelihood estimation, Bayesian estimation.SeealsoBayes’ rule, prior statistics, posteriorstatistics.

maximum accumulated matching a de-fuzzification scheme for a classification prob-lem in which a pattern is assigned to the classoutputed by the rules that accumulates themaximum firing degree.


maximum distance separable (MDS) codean (n, k) linear binary or nonbinary block

code, with minimum distancedmin = n −k+ 1. Except for the trivial repetition codes,there are no binary MDS codes. NonbinaryMDS codes such as Reed–Solomon codes doexist.

maximum effective aperture in antennatheory, the ratio of the time-average poweravailable at the terminals of an antenna dueto power incident in the direction of the mainbeam that is polarized for maximum recep-tion to the time-average power density of theincident field.

maximum entropy a procedure that max-imizes the entropy of a signal process.

maximum entropy inequality an infor-mation theoretic inequality, upper boundingthe possible value of entropy for probabilitydistributions subject to certain moment con-straints. A common example states that forany probability density functionf (x), sub-ject to a power (second moment) constraint,∫x2f (x) ≤ σ 2,

h(f ) ≤ 1

2h(2πeσ 2) .

See alsodifferential entropy.

maximum entropy restoration an iter-ative method of image restoration. At eachiteration, an image is chosen whose Fouriertransform agrees with that of the principal so-lution, i.e., the model of degradation, whilemaximizing the entropy.

maximum excitation limiter a controllerthat is used to limit the maximum amountof field current, or over-excitation, at a syn-chronous generator. This excitation limit isset by rotor winding heating limit.

maximum input power the maximum in-cident RF power that will be applied to a com-ponent or circuit without either damage orperformance degradation, expressed in watts.

The user normally will specify whether itis damage or degradation that is important.If performance degradation is, then how thedegradation is manifested should be speci-fied.

maximum likelihood estimation to esti-mate an unknownx by maximizing the con-ditional probability of the observations; thatis,

xML

= argx maxp(y|x)where y represents an observation.x isconsidered unknown; it is not a statisticalquantity. See alsomaximum a posterioriestimator, Bayesian estimation, Bayes’ rule,prior statistics.

maximum matching a defuzzificationscheme for a classification problem in whicha pattern is signed to the class outputed by therule that achieve the maximum firing degree.

maximum permissible exposure (MPE)the limit adopted by a standards-setting bodyfor exposures of unlimited duration. Gener-ally, some allowance is made in the standardfor time and/or space averaging of higher ex-posure conditions.

maximum posterior probability distanceSeemaximum a posteriori probability.

maximum rated FET gate-to-source volt-age the maximum gate-to-source volt-age that the device is designed to functionwithout damage due to breakdown, as deter-mined by the device manufacturer. The ac-tual breakdown voltage may be much largerthan this voltage, dependent process varia-tion and manufacturer derating criteria andmargins. This voltage rating is usually in the−2.0 volt (low noise devices) to−10 volt(power devices) range.

maximum stable gain (MSG) the max-imum gain derived from a transistor, whichis obtained after making a stable conditionby adding some excess loss elements for thecase when a simultaneous matching for input


and output of the transistor causes an unsta-ble condition.

maximum stable power gain figure ofmerit specified as the maximum value oftransducer gain for which stability factorKis equal to one. Equivalently, this representsthe maximum value of transducer gain of acircuit or device when the device is termi-nated with impedance values, or by usingother methods such as feedback, and so on,such thatK is equal to one.

maximum transducer power gain max-imum value of transducer power gain a cir-cuit or device exhibits; occurs when the in-put and output ports of the circuit are ter-minated with simultaneous conjugate matchconditions. The transducer power gain is de-fined as the ratio of power delivered to a loadto the power available from the source.

maximum-likelihood the maximum ofthe likelihood function,p(y|x) or equiva-lently, the log-likelihood function. Maximum-likelihood is an optimality criterion that isused for both detection and estimation.

maximum-likelihood decoding a schemethat computes the conditional probability forall the code words given the received se-quence and identifies the code word withmaximum conditional probability as thetransmitted word. Viterbi algorithm is thesimplest way to realize maximum-likelihooddecoding.

maxterm a Boolean sum term in whicheach variable is represented in either true orcomplement form only once. For example,x+y′+w+z′ is a maxterm for a four variablefunction.

Maxwell’s equations a set of four vectorequations published in 1873 that govern thegeneration and time evolution of electromag-netic fields of arbitrary electric source dis-tributions. Sometimes called the Maxwell–

Heaviside equations because Heaviside firstwrote them in their familiar vector form.

For fictitious magnetic current densityJand charge densityρ, electric fieldE, mag-netic fieldH , µ is the permeability andε isthe dielectric constant or permittivity.

Maxwell’s equations take on the followingform:

∇ × E = ∂B∂t

(Faraday’s law of induction)

∇ ×H = ∂D∂t

+ J (Ampere’s law)

∇ · B = 0

∇ ·D = q

whereE is electric field strength,D is electricflux density,H is magnetic field strength,Bis magnetic flux density,J is current densityandq is volume charge density.

Maxwell, James Clerk (1831–1879)Born: Edinburgh, Scotland

Maxwell is best known as the greatesttheoretical physicist of the 19th century. Itwas Maxwell who discovered, among otherthings, that light consisted of waves. He de-veloped the fundamental equations describ-ing electromagnetic fields in his work,A Dy-namical Theory of the Electromagnetic Field,published in 1864. Maxwell also gave us themathematical foundation for the kinetic the-ory of gases. Maxwell’s life was cut shortby cancer, and thus he was unable to seehis greatest theoretical propositions provenby experiment.

MBE Seemolecular beam epitaxy.

MBP Seemorphotropic phase boundary.

McCulloch–Pitts neuron originally a lin-ear threshold unit that responded with a bi-nary output at timet + 1 to an input appliedat time t . In current usage, usually a linearthreshold unit.


MCM a unit of area used to specify thecross-sectional area of a wire, equal to 1000circular mils .

MCM-D Seedeposited multi-chip module.

MCM-L Seelaminate multi-chip module.

MCP Seemotor circuit protector.

MCR Seemaster control relay.

MCT See metal-oxide semiconductorcontrolled thyristor.

MDA Seemonochrome display adapter.

MDR Seememory data register.

MDS code See maximum distanceseparable code.

MDT Seemobile data terminal.

Meachem bridge a bridge circuit whereone of the arms (the reactance arm) is a se-ries connection of an inductance, a capaci-tor, and a resistor, all three other arms areresistors. The Meachem bridge has a verysteep phase frequency characteristics in thevicinity of the reactance arm resonance fre-quency. The steepness increases with higherQ-factor of the reactance arm and better tun-ing of the bridge balance at the resonancefrequency. This combination of the bridgeand reactance selectivity properties has se-cured the main application of the Meachembridge as the feedback circuit of high preci-sion crystal oscillators.

Meachem-bridge oscillator an oscilla-tor where the Meachem bridge is used in theamplifier feedback. This circuit is usuallyused in design of high-frequency stabilitycrystal oscillators where the bridge reactancebranch is substituted by a crystal. Tuningof the bridge allows to obtain an extremely

steep phase frequency response of the bridgenear the oscillation frequency which coin-cides with the crystal series-mode frequency.

mean Seemean value.

mean delay the time interval betweentransmission of a pulse through a widebandcommunication channel and the instant cor-responding to the centroid of its power-delayprofile.

mean ergodic theorem a mathematicaltheorem that gives the necessary and/or suf-ficient conditions for a random process to beergodic in mean. Letx(n) be a wide-sensestationary random process with autocorrela-tion sequencecx(k). A necessary and suffi-cient condition forx(n) to be ergodic in themean is

limN−>∞

N−1∑k=0

cx(k) = 0.

The sufficient conditions forx(n) to be er-godic in the mean are that

cx(0) < ∞ and limk−>∞ cx(k) = 0.

mean filter a filter that takes the meanof the various input signal components, or inthe case of an image, the mean of all the pixelintensity values within the neighborhood ofthe current pixel.

mean free path the length of the straightline segment travelled by a photon betweentwo successive hits at scattering centers of aninhomogeneous medium.

mean of a random variable Seeexpectedvalue of a random variable.

mean of a stochastic process the ex-pected value of a stochastic process at somepoint in time.

mean of max method a method of de-fuzzification in which the resultant defuzzi-


fied quantity is obtained from the mean of themaximum grades of membership.

mean opinion score (MOS) (1) a subjec-tive measure of the human-perceived qualityof a particular telecommunication parameter,e.g., transmitted voice quality. Assessed bymethodical exposure of human test subjectsto stimuli corrupted with a known level oftechnical imperfection, and then requestingthe test subjects to rate the stimuli on a sub-jective quality scale.

(2) a subjective method of quantitativelyassessing the quality of signals. Subjects ratethe quality of presented signals (e.g., images)on a quality or impairment scale. The meanof these ratings is calculated as the meanopinion score.

mean pyramids the approach of form-ing hierarchies by averaging over blocks ofpixels (typically 2× 2), thus eliminating thedifficulties associated with subsampling ap-proaches.

mean squared error (MSE) measure ofthe difference between a discrete time signalxi , defined over[1 . . . n], and a degraded, re-stored or otherwise processed version of thesignalxi , defined asMSE = 1

n

∑ni=1(xi −

xi )2. MSE is sometimes normalized by di-

viding by∑ni=1(xi)

2.

mean time to failure (MTTF) given asthe expected working lifetime for a givenpart, in a given environment, as

MTTF =∫ ∞

0r(t)dt

wherer(t) is a reliability function for the part.If the failure rate1

λis constant, then

MTTF = 1

λ.

mean time to repair (MTTR) a predic-tion for the amount of time taken to repair agiven part or system.

mean value the expected value of a ran-dom variable or function. The mean value ofa functionf is defined as

mf = E(f ) =∫ ∞

−∞fp(f )df

wherep(f ) is the probability density func-tion of f . See alsoexpectation.

mean-q convergence for a stochastic pro-cess, the property that the mean of the ab-solute difference between that process andsome random variable, raised to theqthpower (q > 0) approaches zero in time.

mean-square estimation an estimationscheme in which the cost function is themean-square error.

mean/residual vector quantization(MRVQ) an image coding techniquewhere a prediction is made of the original im-age based on a limited set of data, and then aresidual image is formed by taking the differ-ence between the prediction and the originalimage. Prediction data are encoded using ascalar quantizer and the residual image is en-coded using a vector quantizer.

mean/residual VQ See mean/residualvector quantization.

measurement system the sum of all stim-ulus and response instrumentation, deviceunder test, interconnect, environmental vari-ables, and the interaction among all the ele-ments.

mechanical degree the spatial angle of thestator of a machine, expressed in radians ordegrees. Mechanical degree also representsone revolution of the rotor (360◦).

mechanical loss Seerotational loss.

mechanical power energy per unit timeassociated with mechanical motion. For lin-ear motion, it is force× speed; for rotationalmotion, it is torquex rotational speed. In SI


units,P = ωT , whereP is in watts,T is innewton-meters, andω is in rads per second.

media-access control a sublayer of thelink layer protocol whose implementation isspecific to the type of physical medium overwhich communication takes place and whichcontrols access to that medium.

medial axis a subset of blob pixels that arecenters of maximal lines, squares, or diskscontained in the blob.

letX be a non-empty bounded set in a Eu-clidean space; assume thatX is topologicallyclosed, in other words it contains its border∂X. For every pointx in X, the Euclideandistanced(x,Xc) of x to the complementXc

of X is equal to the distanced(x, ∂X) of xto the border∂X; letB(x) be the closed ballof radiusd(x, ∂X) centered aboutx; it is thegreatest closed ball centered aboutx and in-cluded inX. The medial axis transfor is theoperation transformingX into its medial axisor distance skeletonS(X) which can be de-fined in several ways:1.S(X) is the set of pointsx such thatB(x) isnot included inB(y) for any othery in X; inother wordsB(x) is, among all balls includedin X, maximal for the inclusion.2. S(X) is the set of pointsx such thatB(x)intersects∂X in at least two points.3. S(X) is the set of pointsx at which thedistance functionf (x) = d(x, ∂X) is notdifferentiable.

These three definitions coincide up to clo-sure, in the sense that the three skeletons thatthey define may be different, but have thesame topological closure. WhenX is con-nected and has a connected interior, its skele-tonS(X) is also connected and has the samenumber of holes asX.

For digital figures and a digital distanced,one uses only the first definition: the skeletonS(X) is the set of centers of maximal “balls”(in the sense of distanced) included inX;but then the skeleton of a connected set isno longer guaranteed to be connected.Seedistance, morphological skeleton.

medial axis transform (MAT) Seemedial axis.

median filter a filter that takes the me-dian of the various input signal componentsor, in the case of an image, the median of allthe pixel intensity values within the neigh-borhood of the current pixel, the median be-ing defined as the center value of the orderedsignal components.

medical imaging a multi-disciplinaryfield that uses imaging scanners to revealthe internal anatomic structure and physi-ologic processes of the body to facilitateclinical diagnoses. See alsoX-ray, CT,magnetic resonance imaging, ultrasound,positron emission tomography, and radiog-raphy.

medium-scale integration (MSI) (1) anearly level of integration circuit fabricationthat allowed approximately between 12 and100 gates on one chip.

(2) a single packaged IC device with 12 to99 gate-equivalent circuits.

megacell a cell with the radius of 20–100 km.See alsocell.

megaflop (MFLOP) one million floatingpoint operations per second. Usually appliedas a measurement of the speed of a computerwhen executing scientific problems and de-scribes how many floating point operationswere executed in the program.

meggar a power system device for mea-suring high-voltage insulation or ground con-nections.

mel scale the mel is the unit of pitch.

Mellin transform a transform often arisesin the study of wideband signals. The MellintransformFM(s) of a functionf (t) definedon the positive real axis 0< t < ∞ is

FM(s) =∫ ∞

0f (t)ts−1 dt.


The integral in general only exists for com-plex values ofs = a + b for a1 < a < a2,wherea1 anda2 depend on the functionf (t).

melting time the time required for currentto melt the fusible element of a fuse.

membership function a possibility func-tion, with values ranging from 0 to 1, that de-scribes the degree of compatibility, or degreeof truth, that an element or object belongs toa fuzzy set. A membership function valueof 0 implies that the corresponding elementis definitely not an element of the fuzzy setwhile a value of 1 implies definite member-ship. Values between 0 and 1 implies a fuzzy(non-crisp) degree of membership.

Rigorously, letµ(.) be a membershipfunction defining the membership value ofan elementx of an element of discourseX toa fuzzy set. IfA is a fuzzy set, andx is anelement ofA, then the membership functiontakes values in the interval[a, b] of the realline (µA(x) : X ⇒ [a, b] , a, b ∈ <).

Usually, fuzzy sets are modeled with anormalized membership function (µA(x) :X ⇒ [0,1]). Some examples of member-ship functions are shown in the figure. IfA is

Membership function: some examples of normal-

ized fuzzy sets.

a crisp (nonfuzzy) set,µA(x) is 1 if x belongstoA and 0, otherwise (µA(x) : U ⇒ {0,1}).See alsocrisp set, fuzzy set.

membership grade the degree to whichan element or object belongs to a fuzzy set. Itis also referred to as degree of membership.

membrane the functional boundary of acell. Nerve cells possess membranes that areexcitable by virtue of their nonlinear electri-cal conductance properties.

membrane probe used for performinghigh power on wafer test for large peripheryMMIC power amplifiers and discrete deviceswith non-50-W interfaces.

memory (1) area for storing computer in-structions and data for either short-term orlong-term purposes.

(2) the property of a display pixel thatallows it to remain stable in an initially es-tablished state of luminance. Memory givesa display high luminance and absence offlicker.

memory access time the time from whena read (i.e., load) request is submitted tomemory to the time when the correspond-ing data becomes available. Usually smallerthan the memory cycle time.

memory address computation the com-putation required to produce an effectivememory address; may include indexing andtranslation from a virtual to a physical ad-dress.

memory address register (MAR) a reg-ister inside the CPU that holds the address ofthe memory location being accessed whilethe access is taking place.

memory alignment matching data tothe physical characteristics of the computermemory. Computer memory is generally ad-dressed in bytes, while memories handle datain units of 4, 8, or 16 bytes. If the “memorywidth” is 64 bits, then reading or writing an8 byte (64 bit) quantity is more efficient ifdata words are aligned to the 64 bit wordsof the physical memory. Data that is notaligned may require more memory accessesand more-or-less complex masking and shift-ing, all of which slow the operations.


Some computers insist that operands beproperly aligned, often raising an exceptionor interrupt on unaligned addresses. Othersallow unaligned data, but at the cost of lowerperformance.

memory allocation the act of reservingmemory for a particular process.

memory bandwidth the maximum amountof data per unit time that can be transferredbetween a processor and memory.

memory bank a subdivision of memorythat can be accessed independently of (andoften in parallel with) other memory banks.

memory bank conflict conflict whenmultiple memory accesses are issued to thesame memory bank, leading to additionalbuffer delay for such accesses that reach thememory bank while it is busy serving a pre-vious access.See alsointerleaved memory.

memory block contiguous unit of datathat is transferred between two adjacent lev-els of a memory hierarchy. The size of ablock will vary according to the distance fromthe CPU, increasing as levels get farther fromthe CPU, in order to make transfers efficient.

memory bounds register register used toensure that references to memory fall withinthe space assigned to the process issuing thereferences; typically, one register holds alower bound, another holds the correspond-ing upper bound, and accesses are restrictedto the addresses delimited by the two.

memory cell a part in a semiconduc-tor memory holding one bit (a zero or aone) of information. A memory is typicallyorganized as a two-dimensional matrix ofcells, with “word lines” running horizontallythrough the rows, and “bit lines” running ver-tically connecting all cells in that column to-gether.See alsobit line.

memory compaction the shuffling of datain fragmented memory in order to obtainsufficiently large holes. See alsomemoryfragmentation.

memory cycle the sequence of states of amemory bus or a memory (sub-)system dur-ing a read or write. A memory cycle is usu-ally uninterruptible.

memory cycle time the time that mustelapse between two successive memory op-erations. Usually larger than the memory ac-cess time.

memory data register (MDR) a registerinside the CPU that holds data being trans-ferred to or from memory while the access istaking place.

memory density the amount of storageper unit; specifically, the amount of storageper unit surface or per chip.

memory element a bistable device or el-ement that provides data storage for a logic1 or a logic 0.

memory fragmentation See internalfragmentation, external fragmentation.

memory hierarchy See hierarchicalmemory.

memory interleaving See interleavedmemory.

memory latency the time between the is-sue of a memory operation and the comple-tion of the operation. May be less than thetime for a memory cycle.

memory management unit (MMU) ahardware device that interfaces between thecentral processing unit (CPU) and memory,and may perform memory protection, trans-lation of virtual to physical addresses, andother functions. An MMU will translate vir-


tual addresses from the processor into realaddresses for the memory.

memory mapped I/O I/O scheme inwhich I/O control and data “registers” andbuffers are locations in main memory andare manipulated through the use of ordinaryinstructions. Computers with this architec-ture do not have specific I/O commands. In-stead, devices are treated as memory loca-tions. This simplifies the structure of thecomputer’s instruction set.

memory mapping (1) the extension ofa processor-generated address into a longeraddress, in order to create a large virtual ad-dress or to extend a virtual address that istoo short to address all of real memory whentranslated.

(2) the mapping between the logical mem-ory space and the physical memory space,i.e., the mapping of virtual addresses to realaddresses.

memory module a physical componentused in the implementation of a memory.Seealsomemory bank, interleaved memory.

memory partitioning when multiple pro-cesses shares the physical main memory, thememory is partitioned between the processes.If the partitioning is static, and the amount ofmain memory for each process is not changedduring the execution, the memory partition-ing is fixed, otherwise it is said to be dynamic.

memory port an access path to a memoryunit.

memory protection a method for con-trolling access to memory; e.g., a processmay have no access, or read-only access, orread/write access to a given part of memory.The control is typically provided by a com-bination of hardware and software.

memory reference a read of one item ofdata (usually a word) from memory or a write

of one item of data to memory (same as mem-ory reference).

memory reference instruction an in-struction that communicates with virtualmemory, writing to it (store) or reading fromit (load).

memory refresh the process of recharg-ing the capacitive storage cells used in dy-namic RAMs. DRAMs must have every rowaccessed within a certain time window or thecontents will be lost. This is done as a pro-cess more or less transparent to the normalfunctionality of the memory, and affects thetiming of the DRAM.

memory select line a control line usedto determine whether a unit of memory willparticipate in a given memory access.

memory stride the difference betweentwo successive addresses presented to mem-ory. An interleaved memory with a simpleassignment of addresses performs best whenreference strides are 1, as the addresses thenfall in distinct banks.

memory swapping the transfer of mem-ory blocks from one level of the memory hi-erarchy to the next lower level and their re-placement with blocks from the latter level.Usually used to refer to pages being movedbetween main memory and disk.

memory width the number of bits storedin a word of memory. The same as the width.

memory word the total number of bitsthat may be stored in each addressable mem-ory location.

memoryless system a system whose out-put at any instant in time depends only on theinput at that instant of time. The impulse re-sponse of a linear, time-invariant, memory-less system is the impulse itself, multipliedby a constant. See also impulse,impulse re-


sponse , linear time invariant (LTI)system, system, system with memory.

MEMS Seemicroelectromechanical system.

MEM Sfor microwave Seemicroelectro-mechanical system for microwave.

mercurous chloride an acousto-opticmaterial with very slow acoustic velocityand thus very high theoretical acousto-opticfigure-of-merit. The slow acoustic velocityalso gives potential for high time-bandwidthdeflectors.

mercury delay-line memory Seeultrasonicmemory.

meridional ray a ray that is contained ina plane passing through the fiber axis.

MESFET Seemetal-electrodesemicon-ductor field-effect transistor.

mesh analysis a circuit analysis techniquein which KVL is used to determine the meshcurrents in a network. A mesh is a loop thatdoes not contain any loops within it.

mesh networks an interconnection net-work in which processors are placed in a two-dimensional grid of wires, with processors atthe intersections of the wires in the x and ydimensions. “Mesh” is occasionally used torefer to similar networks of higher than twodimensionality.

meshed network a complete interconnec-tion of all nodes in a network by means ofpoint-to-point links. They are usually a com-bination of ring and star networks.

MESI protocol a cache coherence pro-tocol for a single bus multiprocessor. Eachcache line exists in one of four states, modi-fied (M), exclusive (E), shared (S), or invalid(I). (BW)

mesopic formally, a description of lu-minances under which both human rod andcone cells are active. The mesopic luminancerange lies between the photopic and scotopicranges. Informally, describing twilight lumi-nances.See alsophotopic, scotopic.

message passing in object orientation, theexchange of messages between objects.

message switching a service-orientedclass of communication in which messagesare exchanged among terminating equipmentby traversing a set of switching nodes in astore-and-forward manner. This is analogousto an ordinary postal system. The destinationterminal need not be active at the same timeas the originator in order that the messageexchange take place.

message-passing system a multiproces-sor system that uses messages passed amongthe processors to coordinate and synchronizethe activities in the processors.

messenger cable a fully-insulated three-phase aerial cable in which three individ-ual insulated conductors are carried on in-sulated looms hung upon a bare messengerwire. Such cable is used frequently in distri-bution work.

messenger wire a grounded wire whichis used to structurally support an aerial cable.

metabolic process the method by whichcells use oxygen and produce carbon dioxideand heat.

metadyne a DC machine with more thantwo brush sets per pair of poles. The addi-tional brushes are located in the direct axisfor the armature MMF to provide most of theexcitation for higher gains.

metal halide molecule formed by the re-action of metals and halogen atoms.


metal-electrode semiconductor field-effecttransistor (MESFET) a specific type ofFET that is the dominant active (amplify-ing) device in GaAs MMICs. An FET iscomposed of three terminals called the gate,drain, and source, and a conducting “chan-nel.” In an amplifier application, the sourceis connected to ground, and DC bias is ap-plied between the drain and source causinga current to flow in the channel. The currentflow is controlled and “modulated” by the ACor DC voltage applied to the gate.

metal–insulator–metal (MIM) capacitora capacitor, which has a thin insulator layerbetween two metal electrodes. Generally,this capacitor is fabricated in semiconduc-tor process, and this insulator layer provideshigh capacitance. Two extreme behaviors ofa capacitor are that it will act as an opencircuit to low frequencies or DC (zero fre-quency), and as a short frequency at a suf-ficiently high frequency (how high is deter-mined by the capacitor value). Also called athin film capacitor.

metal-organic chemical vapor deposition(MOCVD or OMCVD) a material growthtechnique that uses metal organic moleculesin an atmospheric or low pressure growthchamber and a controlled chemical reactionon a heated substrate to grow a variety of II-VI, III-V, and group IV materials with atomiclayer control. Used to create material struc-tures for a variety of electronic and optical de-vices using quantum wells, heterostructuresand superlattices (for example, TEG, TMG,and TEA).

metal-oxide semiconductor (MOS) thebasic structure in an insulated-gate field-effect transistor. In this technology, a lay-ered capacitor, with the added property ofapplying a voltage of a proper polarity,causes the underlying semiconductor to beswitched from n-type to p-type or vice versa.MOS technology has been responsible for themainstream of integrated circuit technologyfor many years.

metal-oxide semiconductor controlledthyristor (MCT) a voltage-controlledfour-layer (pnpn) device for medium power(700 A) and medium speed (20 kHz) ap-plications, with projections to 2000 A and100 kHz. Unlike transistors, MCTs can onlybe on or off, with no intermediate operatingstates. The MCT has high current handlingcapabilities, a low forward voltage drop, andlow gate drive requirements (no large nega-tive gate current required for turn off).

metal-oxide semiconductor field-effecttransistor (MOSFET) a transistor thatuses a control electrode, the gate, to capaci-tively modulate the conductance of a surfacechannel joining two end contacts, the sourceand the drain. The gate is separated from thesemiconductor body underlying the gate bya thin gate insulator, usually silicon dioxide.The surface channel is formed at the interfacebetween the semiconductor body and the gateinsulator. Used for low power (200 A) andhigh speed (1 MHz) applications.

In power electronics applications, MOS-FETs are typically operated as switches, ineither their fully on or off states, to minimizelosses. The gate is insulated from the semi-conductor portion to enable faster switching.A gate-source voltage permits a current toflow between the drain and the source, wherecontinuous gate voltage is required to be inthe on-state. The primary disadvantage is thehigh forward voltage drop. The “metal” inMOSFET refers to the gate electrode, whichwas fabricated from a metal in early MOS-FETs, but is now typically fabricated from amaterial such as polysilicon.

metal-oxide semiconductor memorymemory in which a storage cell is con-

structed from metal-oxide semiconductor.Usually called MOS memory.

metal-oxide varistor a voltage-dependentresistor frequently used to protect electronicdevices from overvoltages due to lightningor switching surges. The resistance of the


device drops to a few ohms if the voltage ap-plied across it exceeds a calibrated value.

metallization the deposited thin metalliccoating layer on a microcircuit or semicon-ductor.

meteor burst communication VHF radiocommunication using the reflective proper-ties of the ionized trails that burning mete-orite bursts leave in the atmosphere.

method of moments (MOM) a commonprocedure used in order to solve integral anddifferential equations. The unknown func-tion is expanded on a complete (but not nec-essarily orthogonal) set of functions, oftencalled basis functions or expansion set. Thisexpansion is truncated after a finite number ofterms,N . The original functional equation isthen tested, i.e., multiplied and integrated, byanother set of functions, often called weight-ing or testing functions, hence reducing theoriginal functional equation to a matrix equa-tion suitable for numerical computation. Thecommon choice of taking the same set offunctions for the expansion and test is oftenreferred to as Galerkin version of the momentmethod.See alsointegral equation.

method of successive projections an it-erative procedure for modifying a signal sothat the modified signal has properties whichmatch an ideal objective.

metric (1) a measure of goodness againstwhich items are judged.

(2) methodologies for the measurement ofsoftware features including matters as perfor-mance and cost estimation.

metrology the process of measuring struc-tures on the wafer, such as the width of aprinted resist feature.

metropolitan-area network (MAN) acomputer communication network spanninga limited geographic area, such as city; some-times features interconnection of LANs.

Mexican-hat function a function that re-sembles the profile of a Mexican hat. Ac-cording to anatomy and physiology, the lat-eral interaction between cells in mammalianbrains has a Mexican-hat form, that is, exci-tatory between nearby cells and inhibitory atlonger range with strength falling off withdistance. According to this phenomenon,Kohonen proposes a training algorithm forself-organizing system to update not only theweights of the winner but also the weightsof its neighbors in competitive learning.Seealsoself-organizing algorithm.

MFD filter See maximally flat delayfilter.

MFIE Seemagnetic field integral equation.

MFLOP Seemegaflop.

MHD See magneto-hydrodynamicgenerator .

MHz megahertz, or millions of operationsper second.

micro cell a cell in a cellular communica-tions system having a size (or cell radius) thatis significantly smaller than the cell size of atypical cellular system. Such systems havecells with radii that are at most a few hun-dred meters and utilize base stations placedat heights that are of the order of the heightof the mobile terminal.

micro diversity a diversity technique,used in a cellular where multiple anten-nas are separated by distances equal to afew wavelengths and the diversity transmis-sion/reception is utilized to overcome theeffect of multipath fading. Compare withmacro diversity

micro shadowing shadowing due to boththe orientation with respect to transmitter andreceiver.See alsomacro shadowing.


micro-stepping a control technique —also called mini-stepping — that results ina finer positioning resolution than can be ob-tained with simple on/off control of the phasecurrents of a stepper motor. The practicalimplementation of micro-stepping requiresthe accurate and continuous control of all thephase currents of the step motor.

microbending sharp curvatures involvinglocal fiber axis displacements of a few mi-crometers and spatial wavelengths of a fewmillimeters. Microbending causes signifi-cant losses.

microcell a low-power radio network thattransmits its signal over a confined distance.

microchannel diode a high-frequencyplanar monolithic Schottky barrier diodestructure fabricated with micromachiningthat has very small parasitics and is used asa millimeter or submillimeter wave mixer orharmonic multiplier.

microchannel-plate spatial light modula-tor (MSLM) an optically addressedspatial light modulator. When input light isincident onto a photocathode on the write-in side of the MSLM, a photoelectron im-age is formed, which is multiplied to abouthundred thousand times by the microchan-nel plate. The resulting charge pattern willin turn affect the second layer, which is theLiNbO3 crystal plate having electro-optic ef-fect. The refractive index of the crystal plateis now modulated by the electric charge pat-tern. When output light passes through thecrystal plate, the phase of output light is mod-ulated by the varying refractive index of thecrystal plate. Since the LiNbO3 is birefrin-gence, phase modulation becomes polariza-tion modulation. Polarization modulation isvisualized as intensity variation using a polar-izer. The generation of electric charge and itseffect to the crystal layer are nonlinear. Theircombined effect has thresholding capabilitythat can be utilized for constructing opticallogic gates.

microchannel-plate spatial light modula-tor logic gate an optical logic gate utiliz-ing thresholding capability of a microchannel-plate spatial light modulator (MSLM).

microcode a collection of low-level oper-ations that are executed as a result of a singleinstruction being issued.

microcommand an n-bit field specifyingsome action within the control structure of aCPU, such as a gate open or closed, functionenabled or not, control path active or not, etc.

microcomputer a computer whose CPUis a microprocessor chip, and its memory andI/O interface are LSI or VLSI chips.

microcontroller an integrated circuit chipthat is designed primarily for control systemsand products. In addition to a CPU, a micro-controller typically includes memory, timingcircuits, and I/O circuitry. The reason forthis is to permit the realization of a controllerwith a minimal quantity of chips, thus achiev-ing maximal possible minituarization. Thisin turn, will reduce the volume and the costof the controller. The microcontroller is nor-mally not used for general-purpose compu-tation as is a microprocessor.

microelectromechanical system (MEMS)micrometer-scale devices fabricated as dis-

crete devices or in large arrays using inte-grated circuit fabrication techniques. Mov-able compact micromechanical or optome-chanical structures and microactivators madeusing batch processing techniques.

microelectromechanical system (MEMS)for microwave a new multidisciplinarytechnology field with enormous potential forvarious applications, including microwave.MEMS are fabricated by integrated circuitprocessing methods and commonly includesensors and actuators with physical dimen-sions of less than 1 mm on a side.


microinstruction the set of microcom-mands to be executed or not, enabled or not.Each field of a microinstruction is a micro-command.

microlithography lithography involvingthe printing of very small features, typicallyon the order of micrometers or below in size.

micromachine a fabrication process thatuses integrated circuit fabrication techniquessuch as diffusion, oxidation, wet and dryetching to realize mechanical and electricalstructures such as resonators, membranes,cavities, waveguide structures and a varietyof thermal, medical, and chemical transduc-ers.

micromemory Seecontrol memory.

microphone a device that converts acous-tical signals into electrical signals.

microprocessor a CPU realized on an LSIor VLSI chip.

microprogram a set of microcode asso-ciated with the execution of a program.

microprogramming the practice of writ-ing microcode for a set of microinstructions.

microscopy an imaging modality thatuses optical light, laser light, electrons, or an-other radiation source to illuminate a sample.The image is formed by gathering reflectedand scattered energy.

microsensor a sensor that is fabricatedusing integrated circuit and micromachiningtechnologies.

microstrip a transmission line formed bya printed conductor on top of a conductive-backed dielectric. It is often used in high-frequency, printed circuit board applications.

microstrip antenna a radiating elementconsisting of a conducting patch formed on

the surface of a dielectric slab, which in turnlies on a ground plane. Microstrip antennasare usually printed on circuit boards and fedby microstriplines etched on the same board.Also called microstrip patch antenna.

microstrip patch antenna Seemicrostripantenna.

microtiming diagram a graphical displayshowing how the waveforms vary with timebut with a time scale that has sufficient reso-lution to display clearly the delays introducedby the individual basic elements of the digitalcircuit.

microwave term used to refer to a radiosignal at a very high frequency. One broaddefinition gives the microwave frequencyrange as that from 300 MHz to 300 GHz.

microwave coplanar probe a speciallydesigned test probe for measuring devicesfrom DC to microwave frequencies usinga wafer probe station. The probe tip isconstructed using coplanar waveguide topresent a highly controlled impedance (usu-ally 50 ohms) to a device under test.

microwave engineering the engineeringof devices in the frequency range from 1 GHzto 1000 GHz corresponding to the wave-lengths from 30 cm down to 0.3 mm.

microwave transition analyzer a devicethat can combine the functionality of severaldedicated measurement instruments. It is apulsed RF measurement system that operatesin the time domain like a high frequency sam-pling oscilloscope. Using the fast Fouriertransform, the information can be convertedinto frequency domain. It can operate as asampling oscilloscope, pulsed network ana-lyzer, and a spectrum analyzer.

mid-term stability refers to system re-sponses which are shorter than long-termcfbut shorter than transientcf response, gen-erally associated with maximum excitation


limiters, load tap changers and other slow-acting devices.

Mie scattering electromagnetic theorythat describes the scattering of light byspheres.

mildly nonlinear a circuit or element inwhich the output spectrum is made up oftwo parts, the first of which is proportionalthrough gain(s), attenuation(s) and delay(s)to the input spectrum, and the second inwhich spectral shift(s), conversion(s) or gen-eration(s) takes place in an orderly and pre-dictable way. Most real-world circuits andelements are mildly nonlinear at some levelof excitation within hyperspace, and all ac-tive devices are mildly nonlinear.

Miller capacitance an excess amount ofcapacitance that appears in parallel with theinput of an inverting amplifier stage.

Miller effect the increase in the effectivegrid-cathode capacitance of a vacuum tube ora transistor due to the charge induced electro-statically on the grid by the anode through thegrid–anode capacitance.

Miller oscillator the name is usually ap-plied to crystal oscillators with one activedevice (usually a FET with the source ACgrounded) where the crystal is connected be-tween the gate and ground. The crystal isused like an inductor, another tunable paral-lel LC-circuit in the drain is used as an in-ductor as well, and the capacitance betweenthe drain and gate (Miller capacitance hencethe name) is used as a third reactance of LC-oscillator. The circuit becomes similar toHartley oscillator.

Miller’s rule a semi-empirical rule, ofgood but not exact validity, which states thatthe value of the nonlinear susceptibility of or-dern for a given material is proportional ton+ 1 products of the linear susceptibility ofthat material.

millimeter wave an electromagnetic wavein the part of the electromagnetic spectrumthat has a wavelength on the order of amillimeter. This band is centered at about300 GHz.

MIM capacitor Seemetal–insulator–metalcapacitor.

MIMD Seemultiple instruction multipledata architecture.

MIMO system Seemulti-input–multi-output system. See alsosingle-input–single-output system.

min operation an operation on two ormore variables where the resultant value isformed by taking the smallest value, or min-imum, among these variables.

min-max control a class of control algo-rithms based on worst-case design methodol-ogy in which control law is chosen in such away that it optimizes the performance underthe most unfavorable possible effect of pa-rameter variations and/or disturbances. Thedesign procedure can be viewed as a zero-sum game with control action and uncertaintyas the antagonistic players. In the case oflinear models and quadratic indices, the min-max control could be found by solving zero-sum linear-quadratic games. Minimax oper-ations may be performed on cost functionals,sensitivity functions, reachability sets, stabil-ity regions or chosen norms of model vari-ables.

mini-stepping Seemicro-stepping.

minicell a cell with a radius of 300 m to2 km, typically for pedestrian mobile users.See alsocell.

minimal orientation representationminimal orientation representation describesthe rotation of the end-effector frame withrespect to the base frame, e.g., Euler angles(there exists a set of 12 Euler angles). Mini-


mal orientation representation usually has tobe calculated through the computation of theelements of the rotation matrix i.e.,n, o, anda vectors.Seeexternal space.

minimal realization for linear station-ary finite-dimensional continuous-time dy-namical system, is a set of four matricesA,B,C,D that form state and output equa-tions in the state spaceRn with minimal di-mensionn. A minimal realization of lin-ear stationary dynamical system is alwayscontrollable and observable. The similarstatements hold true for linear stationaryfinite-dimensional discrete-time dynamicalsystems.

minimax estimate the optimum estimatefor the least favorable prior distribution.

minimum discernible signal a signalpower level equal to the noise power, usuallyexpressed in watts or decibels. Thus measur-ing the system output power with no signalapplied, and then increasing the input signalpower until a 3 dB increase is observed re-sults in the signal power being equal to thenoise power (i.e.,S/N = 1). Also called theminimum detectable signal.

minimum distance in a forward errorcontrol block code, the smallest Hammingdistance between any two code words. Ina convolutional code, the column distance atthe number of encoding intervals equal to theconstraint length.

minimum energy control of generalized2-D model given the generalized 2-Dmodel



i, j ∈ Z+ (the set of nonnegative integers)with admissible boundary conditionsxi0, i ∈Z+, x0j , j ∈ Z+ and the performance index

I (u) :=N1∑i=0

N2∑j=0

uTijQij

(N1 := N1 + n1; N2 := N2 + n2

)find a sequence of inputsuij for 0 ≤ i ≤ N1and 0≤ j ≤ N2 that transfers the model tothe desired final statexf ∈ Rn, xN1N2 = xfand minimizes the performance indexI (u)wherexij ∈ Rn is the local semistate vec-tor uij ∈ Rm is the input vector,E, Ak,Bk (k = 0,1,2) are given real matrices,Q ∈ Rm×m symmetric and positive definiteweighting matrix and (n1, n2) is the indexof the model.See alsolocal reachability ofgeneralized 2-D model.

minimum energy control of linear systemsa design problem such that for a given initialand final condition, find a control that steersdynamical system on a given time intervalfrom initial conditions to final conditions andhas minimum energy. Minimum energy con-trol problem has a solution for every control-lable linear (continuous or discrete) dynam-ical system. This solution strongly dependson system parameters and the given initialand final states.

minimum excitation limiter a controllerthat is used to limit the minimum amount offield current, or under-excitation, at a syn-chronous generator. This excitation limit isset by stability limit.

minimum free distance for any convo-lutional code, it is the minimum Hammingdistance between the all-zero path and all thepaths that diverge from and merge with theall-zero path at a given node of the trellis di-agram.

minimum mean square error (MMSE)a common estimation criterion which seeksto minimize the mean (or expected) squarederror,

E = E[eT e

]wheree represents the error.

minimum mean square estimator (MMSE)a broad class of estimators based on minimiz-ing the expected squared error criterion. Both


the Linear least squares estimator and theBayesian least squares estimator are specialcases.See alsominimum variance unbiasedestimator, linear least squares estimator,Bayesian least squares estimator,maximuma posteriori estimator, maximum likelihoodestimation.

minimum noise factor for an active cir-cuit or device, occurs when the input terminalis terminated with an impedance which pro-duces the minimum noise factor.

minimum phase system a system that hasall poles and zeroes inside the unit circle. It iscalled minimum phase because the poles andzeroes inside the unit circle cause the groupdelay, which is the derivative of the phase ofthe signal to be minimized.

minimum polynomial for a given ele-mentα of a field, and for a subfieldF , thepolynomial of smallest degree, with coeffi-cients inF havingα as a root. The set of rootsof a minimum polynomial form a conjugacyclass that is defined by the polynomial.

minimum time-to-clear See clearingtime.

minimum time-to-melt See meltingtime.

minimum variance unbiased estimatoran estimatorθ of a parameterθ is said to haveminimum variance and to be unbiased if

Eθ = θ

andE(θ − θ)2 ≤ E(θ − θ)2,

whereθ is any other estimator ofθ .

minimum-shift keying (MSK) a vari-ant of FSK in which the separation betweentransmitted frequencies is 1/(2T ), whereTis the symbol duration. In addition, the ini-tial phase associated with each bit is adjustedso that phase transitions between successivebits are continuous.

minimum-shift keying Gaussian (GMSK)a variant of MSK in which the transmit-

ted frequency makes smooth transitions be-tween the frequencies associated with the in-put bits. A GMSK signal can be generatedwhen the input to the voltage-controlled os-cillator (VCO) is a PAM signal with a Gaus-sian baseband pulse shape.

minisub in European usage, a miniaturesubstation.

Minkowski distance between two realvalued vectors(x1, x2, . . . , xn) and(y1, y2,

. . . , yn) a difference measure given by

DMinkowski =(

n∑i=1

|xi − yi |λ) 1λ

minor hysteresis loop a hysteresis loopgenerated within the major hysteresis loopwhen a magnetic material is not driven to fullpositive or negative saturation.

minority carrier a hole in an n-type or anelectron in a p-type semiconductor.

minterm a Boolean product term in whicheach variable is represented in either true orcomplement form. For example,u · v′ ·w · z′is a minterm for a four-variable function.

MIPS millions of instructions per second,a measure of the speed of a computer.

mirror optical element that reflects andmay also transmit incident light rays andbeams; used to provide feedback in laser os-cillators.

mirroring fault tolerance architecture formanaging two or more hard disks as a uniquedisk by replicating the same data on all thedisks in the mirroring system. The systemalso includes mechanism for verifying thatall disks contain the same information.

MISD Seemultiple instruction single dataarchitecture.


MISO See multi-input–single-outputsystem.

miss the event when a reference is madeto an address in a level of the memory hierar-chy that is not mapped in that level, and theaddress must be accessed from a lower levelof the memory hierarchy.

miss probability the probability of falselyannouncing the absence of a signal.

miss rate the percentage of references to acache that do not find the data word requestedin the cache, given by 1− h where h is thehit rate. Also miss ratio.

miss ratio Seemiss rate.

missile terminal guidance seeker lo-cated in the nose of a missile, a small radarwith short-range capability that scans the areaahead of the missile and guides it during theterminal phase toward a target such as a tank.

mix-and-match lithography a lithogra-phic strategy whereby different types of litho-graphic imaging tools are used to print differ-ent layers of a given device.

mixed A/D simulator a simulator that iscapable of simulating combined analog anddigital circuitry.

mixed method coordination in case ofany mixed method both direct coordinationinstruments — as in the direct method — andthe dual coordination instruments — as in theprice method — are used by the coordinatorto modify local decisions until the coordina-tion objectives are met; different combina-tions of direct and dual coordination instru-ments may be used, and so a variety of mixedmethods can be conceived.

mixed mounting technology a compo-nent mounting technology that uses boththrough-hole and surface-mounting tech-

nologies on the same packaging and inter-connecting structure.

mixer a nonlinear device containing eitherdiodes or transistors, the function of which isto combine signals of two different frequen-cies in such a way as to produce energy atother frequencies. In a typical down con-verter application, a mixer has two inputs andone output. One of the inputs is the mod-ulated carrier RF or microwave signal at afrequencyfrf , the other is a well controlledsignal from a local oscillator or VCO at a fre-quencyflo. The result of down conversion isa signal at the difference frequencyfrf −flo,which is also called the intermediate fre-quencyfif . A filter is usually connected tothe output of the mixer to allow only the de-sired IF frequency signal to be passed on forfurther processing. For example, for an RFfrequency of 10.95 GHz (=10,950 MHz) andan LO frequency of 10 GHz (=10,000 MHz),the IF frequency would be 950 MHz.

mixing amplifier Seeharmonic amplifier.

ML Seemaximum likelihood estimation.

MLR Seemultilayer resist.

MMF Seemagnetomotive force.

MMIC See monolithic microwaveintegrated circuit.

MMSE See minimum mean squareestimatoror minimum mean square error.

MMU Seememory management unit.

MMX register a register designed to holdas many as eight separate pieces of integerdata for parallel processing by a special setof MMX instructions. An MMX register canhold a single 64-bit value, two 32-bit val-ues, four 16-bit values, or eight-byte integervalues, either signed or unsigned. In imple-mentation, each MMX register is aliased to a


corresponding floating-point register. MMXtechnology is a recent addition to the IntelPentium architecture.

MNOS acronym for metal-nitride-oxide-si. A structure used in a type of nonvolatilememory device in which the oxide is suffi-ciently thin to permit electron/hole tunnelingwhile the nitride and the nitrode/oxide inter-face are used to store charge.

mobile data terminal (MDT) a computerterminal installed in a service vehicle which,through some wireless link, communicateswork order information between the crewsand their dispatch center.

mobile station that part of a radio commu-nications system, which is not permanentlylocated in a given geographical location.

mobile charge the charge due to the freeelectrons and holes.

mobile ion a charged ionic species thatis mobile in a dielectric. In metal-oxide-si(MOS) devices, Na+ is most often the sourceof the mobile ions.

mobile robot a wheeled mobile robot isa wheeled vehicle which is capable of anautonomous motion because it is equippedfor its motion, with actuators that are drivenby an on-board computer. Therefore mobilerobot does not have an external human driverand is completely autonomous.

mobility electron mobilityµn = (2LK)/(CoxW) whereCox = capacitance per unit area of the

gate-to-channel capacitor for which the ox-ide layer serves as a dielectric.L = length of the channelW = width of the channelµn = the mobility of the electrons in the

inducedn channelK = constant.

proton mobilityµp ∼= µn/2.

MOCVD See metal-organic chemicalvapor deposition.

modal analysis the decomposition of asolution to an electromagnetic analysis prob-lem into a linear combination (weighted sum)of elementary functions called modes. Theelementary functions are usually orthogonal;typical functions used include sine and cosinefunctions for problems cast in rectangular co-ordinates, Bessel functions for cylindrical co-ordinates, and spherical Bessel functions forspherical coordinates.

This technique can be used in order to findthe field produced by an arbitrary source in-side a waveguide, be it open or closed, orto find the scattering matrix relative to thediscontinuity between different waveguides.See alsoeigenfunction expansion.

modal expansion Seemodal analysis.

modal solution Seemodal analysis.

modality (1) a part of a computer’s in-struction that specifies how another part ofthe instruction is interpreted.

(2) a specific medical imaging technique,such asX-ray CT, magnetic resonance imag-ing, or ultrasound, that is used to acquire animage data set.

modality-specific a task that is specific toa single sense or movement pattern.

mode (1) one possible wave solution ofan infinite number of time–harmonic wavesolutions that exist in a waveguide or trans-mission line. Each mode is identified by acollection of numbers and is usually desig-nated in electromagnetics as either transverseelectric, transverse magnetic, or transverseelectromagnetic.

(2) electromagnetic field distributions thatmatch the boundary conditions imposed by alaser or other cavity.

See alsopropagating mode, cladding,tunneling modes, leaky modes, multi-modeoptical fiber, single-mode fiber.


mode chart a graphical illustration ofthe variation of effective refractive index (orequivalently, propagation angleθ ) with nor-malized thicknessd/λ for a slab waveguideor normalized frequencyV for an opticalfiber.

mode field radius the radius at which theelectric field in a single mode fiber falls to1/e of its value at the center of the core.

mode filter a filter that takes the mode ofthe distribution containing the various inputsignal components or, in the case of an image,the mode of the distribution of all the pixelintensity values within the neighborhood ofthe current pixel. Complications can arisefrom the sparsity of the local pixel intensitydistribution, and the mode should be that ofthe underlying rather than the actual intensitydistribution.

mode matching a particular type of modalanalysis referred to the analysis of a junc-tion between different waveguides, where themodal expansion on one side of the junctionis matched to the modal expansion on theother side of the junction.

mode-locking forcing the modes of alaser oscillator to be equally spaced in fre-quency and have a fixed phase relationship;sometimes also occurs spontaneously.Seealso longitudinal modelocking, transversemode-locking.

model based image coding (1) imagecompression using stored models of knownobjects at both encoder and decoder, the en-coder using computer vision techniques toanalyze incoming images in terms of themodels, and the decoder using computergraphics techniques to re-synthesize imagesfrom the models. The transmitted data aremodel parameters that describe at a relativelyhigh level, the motion of the model parts. Forexample, if the stored model is of a humanhead, the transmitted data may be muscleflexions.

(2) the appreciation that all image codingrelies on some model of the source, and thecategorization of methods according to thetype of model. In this scheme, definition 1 istermed “known-object coding” or “semanticcoding.”

model reference control the controlscheme in which the controlled system ismade to mimic the behavior of a referencemodel system that possesses ideal behavioralcharacteristics.

model-based predictive control Seepredictive control.

modeling the process of creating a suit-able description that emulates the perfor-mance or characteristics of the actual itembeing modeled, over some portion of the de-vice hyperspace. Modeling involves all orparts of model creation, device characteri-zation, de-embedding, parameter extraction,verification, validation, valuation and docu-mentation.See alsomathematical modeling,fuzzy modeling.

modem abbreviation for modulator-demodulator. A device containing a modula-tor and a demodulator. The modulator con-verts a binary stream into a form suitable fortransmission over an analog medium such astelephone wires or (in the case of a wirelessmodem) air. The demodulator performs thereverse operation, so two modems connectedvia an analog channel can be used to transferbinary data over the (analog) channel.

modem-FEC coding error control coding(ECC), applied to a digital signal such thatfeedforward error correction (FEC), can beused in the modem, thus detecting and oftencorrecting transmission errors.

moderator a material contained in a nu-clear reactor core which slows down neu-trons to thermal energies, primarily by neu-tron scattering.


MODFET acronym for modulationdoped FET. See high electron mobilitytransistor.

modified nodal formulation a modifica-tion of the classical nodal formulation whichallows any network to be described. Themodification consists of adding extra equa-tions and unknowns when an element not nor-mally modeled in classical nodal analysis isencountered.

modified signed-digit computing a com-puting scheme in which a number is repre-sented by modified signed-digit. This num-ber system offers carry free addition and sub-traction. Instead of 0 and 1, numbers are rep-resented by−1, 0, and 1 for the same radix2. If a number is represented by 0 and 1,we may need carry in the addition. However,since the number can be represented by threepossibilities−1, 0, and 1, the addition andsubtraction can be directly performed with-out carry following a specific trinary logictruth table for this number system.

modified z-transform a z-transform ofsignals and systems that contain nonzerodeadtimeτ in the range

0 ≤ τ < T

where T is the sampling interval. Modi-fied z-transforms are usually derived fromfundamental principles and given as separatecolumns in z-transform tables for standardfunctions.

modified-return-to-bias recording Seemagnetic recording code.

modifier an operation that modifies themembership of a fuzzy set. Examples ofmodifiers are

1.very A = µcon(A)(u) = (µA(u))2 (con-

centration);2. more or less A = µdil(A)(u) =

(µA(u)).5 (dilatation).

See alsofuzzy set, linguistic variable.

modular network a network whose over-all computation is carried out by subnetworkswhose outputs are combined in some appro-priate way. The term is most commonlyapplied to networks that partition the inputspace so that the subnetworks operate on “lo-cal” data, but is also applied to the case wherea problem can be decomposed into successivetasks, each being implemented by a suitablesubnetwork.

modularity design principle that callsfor design of small, self-contained unit thatshould lead to maintainability.

modulated filter bank a filter bank ob-tained by shifting the spectrum of a proto-type low pass filter in the frequency domainto cover the entire frequency band. Differentproperties can be obtained by using differ-ent modulation schemes and imposing otherconditions. Cosine-modulated filter banksare the most commonly used modulated filterbanks.

modulating signal the baseband sourcesignal used to encode information onto a car-rier wave by varying one or more of its char-acteristics (e.g., the amplitude, frequency, orphase of a sinusoid; or the amplitude, width,repetition rate, or position of each pulse in aperiodic pulse train).

modulation (1) variation of the amplitudeor phase of an electromagnetic wave.

(2) the process of encoding an information-carrying waveform onto a carrier wave-form, typically in preparation for transmis-sion. Seeamplitude modulation, frequencymodulation, phase modulation.

modulation efficiency ratio of the base-band bit rate to the transmission bandwidthafter modulation.

modulation index for an angle-modulatedsignal, the modulation index is the ratio of themaximum modulation deviation to the mod-ulating frequency, and represents the maxi-


mum phase deviation in the modulating sig-nal.

modulation property a property of theFourier transform in which the Fourier trans-form of a modulated signalc(t)ejwot is equalto C(w − wo), whereC(w) is the Fouriertransform ofc(t).

modulation transfer function (MTF) foran imaging system the Fourier transform ofthe system line spread function. The MTFdescribes the spatial frequency resolution ofthe system.

modulator device that varies the ampli-tude or phase of an electromagnetic wave.

modus ponens a rule of reasoning whichstates that given that two propositions,A andA ⇒ B (implication), are true, then it can beinferred thatB is also true.

modus tollens a rule of reasoning whichstates that if a propositionB is not true andgiven thatA ⇒ B, then it can be inferred thatA is also not true.

Moire pattern image caused by a combi-nation of two effects, sampling rate and re-construction filter shape; they occur in imagesignals when two conditions are met. First,when the sampling frequency is close to theNyquist Frequency for the signal (i.e., twotimes the highest frequency in the image sig-nal). Second, the cutoff frequency of the re-construction filter is located beyond one halfthe sampling frequency (e.g., a first order fil-ter). These two problems result in mirrorimages of frequency components around thesampling frequency causing banding in theimage signal. The Moire effect is seen inpractical applications due to real world prob-lems of finite filter lengths and errors in sam-pling rates.

molded case circuit breaker a low-voltage air circuit breaker that includes ther-mal and/or magnetic overcurrent sensing

which directly trips the breaker. The moldedcase circuit breaker is nearly always manu-ally closed, opened, and reset.

molecular beam a source of moleculestraveling primarily in one direction. In prac-tice, molecular beams are usually realized byexpansion of an atomic or molecular vaporinto a vacuum through a small aperture. Theresulting expanding cloud of molecules isusually made nearly unidirectional by a col-limator that blocks or otherwise removes allmolecules not propagating within a narrowrange of angles.

molecular beam epitaxy (MBE) a ma-terial growth technique that uses atomic ormolecular beams in an ultra high vacuumchamber to grow a variety of II-VI, III-V, andgroup IV materials with atomic layer con-trol. Individual molecules or atoms are ex-cited from heated sources (effusion cells) andattach to the substrate in an ordered manner.High quality growth is achieved when the sur-face diffusion coefficient is sufficiently highthat the atoms can arrange themselves coher-ently on the surface.

Used to create material structures for a va-riety of electronic and optical devices usingquantum wells, heterostructures, and super-lattices.

molecular transition coupling of energylevels in an atom by means of absorption oremission processes.

molecular vapor a material composed ofmolecules in the vapor phase.

Mollow gain gain that originates whena 2-level system is driven by a strong, nearresonant, electromagnetic field.

MOM Seemethod of moments.

moment a statistic of a random variable.For example, the first moment is called themean. In general, thenth moment is given


by

mn = E(XT n) =∫ ∞

−∞xnX(x)dx,

wherefX(x) is the probability density func-tion of X. Moments are often used to aidthe recognition of shapes. See alsoprobability densityfunction.

moment method Seemethod of moments.

momentary interruption a loss of volt-age of less than 0.1 pu for a time period of0.5 cycles to 3 seconds.

momentary monitoring the duration atsupply frequency from 30 cycles to 3 sec-onds.

momentary overvoltage an increase involtage above the system’s specified upperlimit for more than a few seconds. Generallya rather loosely-defined term.

momentum relaxation time the charac-teristic time for loss of momentum or velocitydue to scattering processes.

monaural attribute attribute of ear inputsignals (e.g., timbre, loudness) that requireonly one ear to be detected.

monitor (1) the main display device of apersonal computer. Usually uses cathode raytube (CRT) or liquid-crystal display (LCD)technology.

(2) a computer program providing basicaccess to the register contents and memorylocations, usually for debugging purposes.

monitor display Seemonitor.

monochromatic light light that has onlyone frequency component.

monochrome the representation of an im-age, analog or digital, using only one colorand black.

monochrome display adapter (MDA) amonocrome video adapter with 25 lines andsupporting 80 columns, proposed by IBM in1981.

monocular vision a vision model inwhich points in a scene are projected ontoa single image plane.

monolayer one atomic or, in the case ofmaterials such as GaAs, diatomic layer ofatoms.

monolithic microwave integrated circuit(MMIC) integrated circuits made of gal-lium arsenide (GaAs), silicon, or other semi-conducting materials where all of the compo-nents needed to make a circuit (resistors, in-ductors, capacitors, transistors, diodes, trans-mission lines) are formed onto a single waferof material using a series of process steps.

Attractive features of MMICs over com-peting hybrid (combination of two or moretechnologies) circuits are that a multitude ofnearly identical circuits can be processed si-multaneously with no assembly (soldering)using batch processing manufacturing tech-niques. A disadvantage is that circuit adjust-ment after manufacture is difficult or impos-sible. As a consequence, significantly moreeffort is required to use accurate computer-aided-design (CAD) techniques to designMMICs that will perform as desired withoutadjustment. Of course, eventually assem-bly and packaging of MMICs is performedin order to connect them into a system suchas a DBS receiver. MMICs are only cost-effective for very high volume costs, becausethe cost of the initial design is very high, asis the cost of wafer manufacture. These costscan only be recovered through high volumemanufacture. (The word “monolith” refersto a single block of stone that does not (ingeneral) permit individual variations).

monopole Seemagnetic monopole.

monopole antenna an antenna consist-ing of a straight conducting rod, wire, or


other structure oriented perpendicularly to aground plane and fed at the junction of thestructure and the ground plane.

monostatic scattering the reflection of aportion of an electromagnetic wave back inthe direction of the wave source. Monostaticscattering is measured by having the trans-mitter and receiver collocated.

Monte Carlo method a numerical tech-nique that replaces a deterministic descrip-tion of a problem with a set of random de-scriptions that have been chosen based on dis-tributions that match the underlying physicaldescription of the problem. This technique iswidely used to investigate transport and ter-minal characteristics in small semiconductorstructures.

Moore’s law the observation that the num-ber of transistors on a typical chip doublesabout every 18 months. Named for Gor-don Moore, one of the founders of FairchildSemiconductor and Intel.

MOPA Seemaster-oscillator-power-ampli-fier.

MOPS acronym for millions of operationsper second.

morphological duality a morphologicalproperty. There are three standard mean-ings of duality for morphological operators:1. Order-theoretic duality: To any propertyor concept corresponds the dual property orconcept, where the relations⊆,≤ and opera-tions∪,∩, sup, inf, etc., are replaced by theirduals⊇,≥ and∩,∪, inf , sup, etc.2. Duality under complementation (or gray-level inversion): Letψ be a morphologicaloperator, and letN be the operation trans-forming an imageI into its negativeN(I)(when I is a set,N(I) is its complement,while whenI is a gray-level numerical func-tion,N(I) has the gray-level inverted at eachpoint). Then the dual ofψ is the operatorψ∗arising from applyingψ to the negative of an

image; in other wordsψ∗ transformsI intoN(ψ(N(I))).3. Adjunction duality: The dilation and ero-sion byB, namelyδB : X 7→ X ⊕ B andεB : X 7→ XB, form an adjunction, whichmeans that for everyX, Y we have

δB(X) ≤ Y ⇐⇒ X ≤ εB(Y ),

“the dilation ofX is belowY if and only ifX is below the erosion ofY .” This relationconstitutes a bijection between the family ofdilations and that of erosions, and it can thusbe considered as a duality between them.

In the latter two cases, duality inverts theordering relation≤ between operators.Seedilation, erosion, morphological operator.

morphological filter a morphological op-eratorψ that is both:1. idempotent: applying it twice gives thesame result as applying it only once; math-ematically speaking, given an objectX, wehaveψ(ψ(X)) = ψ(X);2. increasing: it preserves the ordering re-lation ≤ between objects; given two objectsX andY , X ≤ Y impliesψ(X) ≤ ψ(Y ).Openings and closings are morphological fil-ters. Seeclosing, morphological operator,opening.

morphological operator an operationfor transforming images, which does notarise from the traditional signal processingmethodology using linear filters and Fourieranalysis but which is rather based on set-theoretical operations. Such an operator forsets (binary images) combines union, inter-section, translation, and sometimes comple-mentation. For numerical functions (gray-level images), union and intersection are re-placed by supremum and infimum (upper andlower envelope), while gray-level inversiontakes partially the role of set complementa-tion. However morphological operators forgray-level images can be visualized by apply-ing the corresponding morphological opera-tor for binary images to the umbra of a gray-level image. Seeclosing, dilation, erosion,mathematical morphology, opening, umbra.


morphological processing low-level pro-cessing technique for binarized images in-volving the shrinking or growing of local im-age regions to remove noise and reduce clut-ter.

morphological skeleton an archetypalstick figure that internally locates the centralaxis of an image.

LetB be the closed ball of unit radius cen-tered about the origin; for any radiusr ≥ 0,let rB be the homothetic ofB by factor r(thus rB is the closed ball of radiusr cen-tered about the origin). Given a non-emptybounded setX in a Euclidean space and apointp in X, p is the center of a closed ballof radiusr, which is included inX, but notcontained in a larger ball included inX, ifand only if

p ∈ XrB and∀ε > 0, p /∈ (XrB)◦εB.Here and◦ denote the erosion and openingoperations respectively. The medial axis ofX is the set of all pointsp such that there issomer ≥ 0 for which this equation holds.The morphological skeleton is defined in thesame way, except that it does not restrict itselfto the isotropic Euclidean ball: take the sameformula above, but consider in it the setB tobe any convex, bounded, and topologicallyclosed structuring element instead of the unitball.

For sets in digital space, the above ap-proach is simplified as follows: Forr =0,1,2, . . ., one takes a bounded digital struc-turing elementBr , with B0 being restrictedto the origin, andBr increasing withr; themorphological skeleton is made of all pixelsp such that there is some integerr ≥ 0 forwhich

p ∈ XBr and p /∈ (XBr)◦B1.

Seeerosion, medial axis transform, opening,structuring element.

morphological system a non-linear coun-terpart of linear shift invariant systems, basedon morphological operators which are in-variant under both spatial and gray-level

shifts. See linear shift invariant system,morphological operator.

morphology Seemathematical morpho-logy.

morphotropic phase boundary (MPB)materials that have a MPB assume a differ-ent crystalline phase depending on the com-position of the material. The MPB is sharp(a few percent in composition) and separatesthe phases of a material. It is approximatelyindependent of temperature in PZT.

MOS Seemean opinion scoreor metal-oxide semiconductor.

MOS memory Seemetal-oxide semi-conductor memory.

MOSFET See metal-oxide semiconduc-tor field effect transistor.

MOSIS acronym for metal-oxide semi-conductor implementation service. Seemetal-oxide semiconductor.

mother wavelet the wavelet from whichwavelets in different scales are obtained bythe translation and dilation operations.Seealsofather wavelet.

motherboard in a computer, the mainprinted circuit board which contains the basiccircuits and to which the other componentsof the system are attached.See alsodaughterboard.

motion analysis determination of the mo-tion of objects from sequences of images. Itplays important roles in both computer visionand image sequence processing. Examplesfor the latter include image sequence datacompression, image segmentation, interpo-lation, image matching, and tracking.

motion blur blur that is due to the motionof an object: it frequently arises when images


containing moving objects are grabbed by aninsufficiently rapid digitizer.

motion compensation generation of aprediction image, an interpolated or extrap-olated frame, from pixels, blocks, or regionsfrom known frames, displaced according toestimates of the motion between the knownframes and the target frame.

motion estimation (1) process of estimat-ing the displacement of moving objects in ascene.

(2) strictly, the estimation of movementwithin a video sequence, including cameramovement and the independent motion of ob-jects in the scene. In practice, refers to thedetermination of optical flow (the spatiotem-poral variation of intensity) which may notcorrespond to real movement (when, for ex-ample, an object moves against a backgroundof the same color). Gradient-based methodsare based on the expansion of a Taylor se-ries for displaced frame difference yieldingan “optical flow constraint equation”. Thismust be combined with other constraints toyield a variational problem whose solutionrequires the iterative use of spatial and tem-poral derivatives. Block-matching methodsrely on finding minimum error matches be-tween blocks of samples centered on thepoint of interest. Frequency-based tech-niques measure phase differences to estimatemotion.

motion measurement the measurementof velocities. Usual motion measurementtechniques such as optical flow give incom-plete and unstable results which must beregularized by comparing velocities over awhole moving object (the latter must first besegmented).

motion segmentation the decompositionof a scene into different objects according tothe variation of their velocities. Motion seg-mentation requires the measurement of ve-locities across the scene.

motion stereo a specific case in motionestimation in which the 3-D scene is station-ary and the only camera is in rigid motion.

motion vector a vector displacement rep-resenting the translation of a pixel, block, orregion between two frames of video, usu-ally determined by optical flow calculation orblock matching (see motion estimation). Forinstance, in the case of dense motion field,say, an optical flow field, a motion vector isassigned to each pixel on the image plane. Inthe case of block matching for image coding,a motion vector is assigned to each block. Inthe case of computer vision, a motion vectorsometimes refers to the velocity of a point oran object in the 3-D space. Sometimes mo-tion vectors are referred to as displacementvectors.

motional time constant a material pa-rameter that is inversely proportional to theproduct of the frequency and Q-factor for aparticular mode.

motor an electromechanical device thatconverts electrical energy from a DC or anAC source into mechanical energy, usuallyin the form of rotary motion.

motor circuit the three components ofan electrical circuit are source, load, and in-terconnecting circuit conductors. A motorcircuit is an electrical circuit designed to de-liver power to a motor. It includes the over-current protective devices, controller, discon-nect switch, circuit conductors, and the motoritself, as shown in the figure.

motor circuit protector (MCP) a listedcombination motor controller containing anadjustable instantaneous-trip circuit breakerand coordinated motor overload protec-tion. MCPs can provide short-circuit andbolted ground-fault protection via the circuitbreaker magnetic element, overload protec-


A typical electric motor circuit.

tion via the overload device, motor control,and disconnecting means all in one assembly.

motor control center an enclosure withone or more sections containing motor con-trol units that have a common power bus.

motor control circuit a circuit containingdevices such as the start/stop switches, main-line coil, main-line sealing contacts, over-load contacts, timers and timer contacts, limitswitches, antiplugging devices, and anythingelse used to control devices in the motor cir-cuit.

motor current signature analysis the useof the currents of an electric machine to pro-vide diagnostic or other information on thehealth of the machine, coupling, or load.

motor operated switch a switch operatedby a motor that is capable of being controlledfrom a remote location.

motor starter an electric controller, eithermanual or automatic, for accelerating a motorfrom rest to normal speed and for stopping themotor.

motor-generator set a set consisting ofa motor mechanically coupled to and drivingone or more generators. The set used to beemployed for AC-to-DC or DC-to-AC powerconversion or voltage level or frequency con-version. Solid-state conversion units are re-placing motor-generator sets in most appli-cations.

mouse an I/O device with a trackball usedto produce signals which are interpreted asx/y coordinates. It is used as locator device,as are the joystick, trackball and tablet. Themouse is the most commonly used locatorfor working with windowing systems for itssimilarity with the movements of the hand,metaphor or the moving hand.

A mouse may present 1, 2, or 3 buttons.The actions associated to the pressing of but-tons can lead to draw objects, select objects,activate menu, delete objects, etc. The ac-tion of pressing is called “click.” Rapidlydouble-clicking the mouse produces a differ-ent, program-dependent effect.

MOV Seemetal-oxide varistor.

move instruction a computer instructionthat transfers data from one location to an-other within a computing system. The trans-fer may be between CPU registers, betweenCPU registers and memory or I/O interfacein either direction. Some systems (suchas Motorola M68000) permit transfer by amove instruction between memory locations(memory-to-memory transfer).

moving average anN th order moving av-erage relationship between an inputx andoutputy takes the form

y[n] =N−1∑i=0

αix[n− i].

A moving average process is any processy which can be expressed as the mov-


ing average of white noisex. See alsoautoregressive.

Moving Picture Experts Group (MPEG)group that standardizes methods for mov-

ing picture compression. MPEG-1 andMPEG-2 are very flexible generic standardsfor video coding, incorporating audio andsystem-level multiplex information. Theyuse block-based motion-compensated pre-diction on 16× 16 “macroblocks,” andresidue coding with the 8× 8 discrete co-sine transform. Frames types are intraframes(coded without reference to other frames),P frames (where the prediction is generatedfrom preceding I and P frames), and B frames(where the prediction is generated bidirec-tionally from surrounding I and P frames).

MOVPE See metal-organic chemicalvapor deposition.

MOX mixed oxide fuel, containing a mix-ture of U235 and PU239 oxides as the fissilematerial.

MPE Seemaximum permissible exposure.

MPEG See Moving Picture ExpertsGroup.

MPP distanceSeemaximum a posteriori probability.

MPPC See mutually pumped phaseconjugator.

MQW Seemultiple quantum well.

MRI Seemagnetic resonance imaging.

MRVQ See mean/residual vectorquantization.

MSB in a binary word, a bit with the high-est weight associated with it.

MSC Seemagnitude squared coherence.

MSE Seemean squared error.

MSG Seemaximum stable gain.

MSI Seemedium-scale integration.

MSK Seeminimum-shift keying.

MSLM Seemicrochannel-plate spatiallight modulator.

MSLM logic gate Seemicrochannel-platespatial light modulator logic gate.

MTF Seemodulation transfer function.

MTTF Seemean time to failure.

MTTR Seemean time to repair.

MUI multi-user interference. Seemultiple access interference.

multi-carrier communications a com-munications method where the bandwidthis subdivided into several smaller frequencybands. Either the same information is trans-mitted over all subbands or different infor-mation is transmitted either simultaneouslyof successively over these bands. By doingso, the detrimental effects of frequency se-lective fading can be minimized.

multi-carrier modulation a modulationtechnique in which the channel is divided intonarrow frequency bins (subchannels), and in-put bits are multiplexed into substreams, eachof which is transmitted over a particular sub-channel. The bit rates and transmitted pow-ers associated with the substreams can be se-lected to maximize the total transmitted bitrate, subject to a total average power con-straint, and also to achieve a desired trans-mitted spectrum.

multi-input–multi-output system (MIMO)also known as multivariable (MV) systems.A system that can transform two or more in-put signals to two or more output signals.


multi-input–single-output system (MISO)a system which can transform two or more

input signals to one output signal.Seealso single-input–single-output systemandmultiple-input–multiple-output system.

multi-mode code a line code where eachsource word is represented by a code word se-lected from a set of alternatives. Code wordsare selected according to a predefined crite-rion, which may depend in part on the statis-tics of the encoded sequence, thereby causingthe same source word to be represented by anumber of different code words throughoutthe encoded sequence.

multi-mode fiber an optical fiber with arelatively large core diameter in which morethan one and usually from several hundred toseveral thousand modes may propagate. Theoptical fiber may be either a step index or agraded index fiber.

multi-mode optical fiber Seemulti-modefiber.

multi-mode oscillation oscillation inmore than a single cavity mode.

multi-phase oscillator an oscillator thatprovidesm sinusoidal voltages shifted withrespect to each other by 2π/m phase angle.It is obtained by connection ofm similar iso-lated phase-shifting circuits in a loop. Theoutputs of isolating stages provide the re-quired output voltages. These oscillators areusually used in instrumentation where light-weight auxilliary three-phase power suppliesare required. Also called m-phase oscillator.

multi-resolution analysis Seemulti-scaleanalysis.

multi-scale analysis the analysis or trans-formation of a signal using analysis ba-sis Functions or analysis filters with differ-ing time resolutions or spatial resolutions.Equivalently the basis functions have differ-ing frequency resolutions. This is in contrast

to the continuous or discrete Fourier trans-form whose analysis basis functions all have(roughly) the same frequency and time res-olution. For discrete multi-resolution analy-sis a multi-level filter bank is often used: thediscrete wavelet transform being a classic ex-ample.See alsomultiscale.

multi-user CDMA a term that has beenused to denote a CDMA system where themultiple access interference is used construc-tively in the receiver to enhance performanceand not treated merely as interfering noise asfor the conventional single-user receiver. Inprinciple, any CDMA system is multi-userCDMA, since it supports multiple users.

multi-user detection joint detection ofthe data symbols for all the users in a multi-ple access system, as opposed to single-userdetection where the data symbol for each useris detected individually.

multi-user detector a detector in a spreadspectrum multiple access system or CDMAsystem where the data bits from all the usersare detected using a joint detection algorithm.Such a detector typically has significantlybetter performance (results in lower bit errorprobabilities) in comparison to a detector thatdetects the various signals individually and,in detecting a bit from a given user, treatsall the other signals (from other users) as acomposite source of interference.

multi-user interference (MUI) Seemultiple access interference.

multi-user receiver a spread spectrum orCDMA receiver that utilizes a multi-user de-tector.

multi-variable system (MV) Seemulti-input–multi-output system.

multibus a standard system bus originallydeveloped for use in Intel’s MicrocomputerDevelopment System (MDS). This standardgives a full functional, electrical, and me-


chanical specification for a backplane busthrough which a number of circuit boardsmay be interconnected. A full range of de-vices may be involved, including computers,memory boards, I/O devices, and other pe-ripherals.

multichannel acousto-optic device acous-to-optic device with multiple independenttransducers bonded to the acousto-opticmedium to introduce multiple independentacoustic signals in the device.

multigrid an efficient numerical algo-rithm for solving large sets of linear equa-tionsAx = b, particularly for “stiff” (nearlysingular)A. The algorithm defines a hierar-chy of grids, with interpolation and decima-tion operations defined between successivegrids in the hierarchy. A system of equationsAgxg = bg is defined on each gridg: thesystems on finer grids yield higher resolutionsolutions, however coarser grids yield muchfaster convergence times. Various empiricalstrategies have been devised which dictatethe order in which different grids are usedto contribute to the final solution.See alsointerpolation, decimation, multiscale.

multigrid block matching a block match-ing technique that is carried out with a hierar-chical structure, known as a multigrid struc-ture. In the highest hierarchical level an im-age may be decomposed into large blockseach with equal size. Each block in a higherlevel is further equally decomposed into sub-blocks, forming the next level. The blockmatching is first applied to the highest level.The matching results obtained are then prop-agated to the next hierarchical level. Thatis, the matching results obtained in the high-est level are utilized as initial estimates andrefined in the second highest level. Thisprocess continues until the last hierarchicallevel. It is noted that the term multigridblock matching is sometimes intermixed withmultiresolution block matching in the litera-ture. In the strict sense they are different. Inthe former, different levels in the hierarchical

structure have the same resolution, while inthe latter, different levels have different res-olutions.

multilayer control operation or structureof the control system with the (multilayer)controller decomposed into two or more con-trol layers; usually in case of the controllerwith two control layers one uses the termstwo-layer control and two-layer controller.Typically within the multilayer control hier-archy the frequency of interventions made bycontrol layers decreases as one moves up thehierarchy.

multilayer perceptron an artificial neu-ral network consisting of an input layer, pos-sibly one or more hidden layers of neurons(perceptrons), and an output layer of neurons.Each layer receives input from the previouslayer and the outputs of the neurons feed intothe next layer.

multilayer resist (MLR) a resist schemeby which the resist is made up of more thanone layer, typically a thick conformal bot-tom layer under a thin imaging layer, possi-bly with a barrier layer in between.

multilayered medium a medium com-posed by several different layers of dielectricmaterials. Such media find wide applicationfor both microwave and optical integrated cir-cuits.

multilevel cache a cache consisting oftwo or more levels, (typically) of differentspeeds and capacities.See alsohierarchicalmemory.

multilevel code in this scheme each bit(level) of a signal is encoded with differenterror correction codes of same block length.It has inherent unequal error protection capa-bility.

multilevel control operation or structureof the control system, where the controller, orthe given control layer, is composed of sev-


eral local decision units coordinated (see co-ordination) by a supremal unit (coordinatorunit); the number of levels can be greater thantwo — in this case decision units situated atan intermediate level are supremal (coordi-nating) units for the subordinated lower levelunits and at the same time are local units asperceived by the level above; the term multi-level control is equivalent to the hierarchicalcontrol, when the latter term is used in thenarrow sense.

multilevel memory See hierarchicalmemory.

multilevel optimization decision mech-anism (or operation of such mechanism)whereby the decisions are made by solv-ing a large-scale optimization problem, par-titioned (decomposed) into several smallerproblems; in a two-level optimization caselocal decisions are influenced by the coordi-nator — in the process of iterative coordi-nation until overall satisfactory decisions areworked out; in a three-level or in a multilevelcase the coordinating unit of the higher levelcoordinates the decisions of several subordi-nate units, which themselves can be coordi-nating units for the lower levels.

multimedia a generic category of compu-ter-controlled media that combines text,sound, video, animation, and graphics intoa single presentation. An example of multi-media would be an electronic encyclopediain CD-ROM format.

multiparameter sensitivity appears as aneffort to introduce generalized functions thatrepresent the influence of all circuit elements.If the circuit function depends on more thanone component, thenF(s, x1, x2, . . . , xn) =F(s, x) and one may write that

dF(s, x)F (s, x)

= d lnF(s, x) =n∑i=1

SF(s,x)xi

dxi

xi

One can introduce a multiparameter sensitiv-ity row vector

SF(s,x)x =

[SF(s,x)x1

SF(s,x)x2

. . .SF(s,x)xn

]To characterize and compare the vectorsof this type one introduces different vectormeasures that are called sensitivity indices.Sometimes these sensitivity indices are con-sidered as multiparameter sensitivities.

multipath (1) propagation of electromag-netic waves along various paths from thetransmitter to the receiver.

(2) interference of electromagnetic wavesdue to multiple propagation paths (e.g., directwave plus ground reflection).

multipath propagation the process bywhich a radio signal propagates from thetransmitter to the receiver by way of multi-ple propagation paths. Depending on the fre-quency range employed in transmission, thephysical propagation phenomena contribut-ing to multipath propagation include reflec-tions from terrain and man-made obstacles,diffraction and refraction. Multipath propa-gation leads, similar to acoustic echoes, intomultiple replicas of the transmitted signalto be received, each having experienced po-tentially different delays, attenuations, andphase shifts.

multiple access channel a multiple trans-mitter, single receiver communication sys-tem, in which the received signal is a (possi-bly non-deterministic) function of all trans-mitted signals.See alsobroadcast channel,interference channel, two-way channel.

multiple access interference (MAI) alsomulti-user interference (MUI). Interferencebetween users in a multiple access systemstemming from nonorthogonality betweenusers, e.g., nonorthogonal code sequences inCDMA or (partially) overlapping frequencybands (FDMA) or time-slots (TDMA).


multiple dwell detector a detector typi-cally used in spread spectrum PN sequenceacquisition (i.e., PN sequence synchroniza-tion) where the detection process is carriedout in stages. In the PN code acquisitionprocess, in a spread spectrum receiver, the re-ceiver makes a hypothesis about the correctPN code phase and then tests the hypothesisby correlating the received signal with a lo-cally generated version with a phase equal tothe hypothesized phase. If the value of the re-sulting correlation is less than a given thresh-old value, then the receiver examines a newphase. If the correlation is greater than thethreshold value, then the receiver continuesthe examination of the same PN code phaseby performing a further correlation with thesame phase in a detection stage commonlyreferred to as the confirmation stage.

multiple instruction multiple data (MIMD)architecture a parallel processing sys-tem architecture where there are more thanone processors and where each processor per-forms different instructions on different datavalues.

In an optical computer implementation,this can be done with a lenslet array, a holo-gram array, or a set of beamsplitters.

multiple instruction single data (MISD)architecture a parallel processing archi-tecture where more than one processor per-forms different operations on a single streamof data, passing from one processor to an-other.

multiple quantum well (MQW) collec-tion of alternating thin layers of semicon-ductors (e.g, GaAs and AlGaAs) that resultsin strong peaks in the absorption spectrumwhich can be shifted with an applied voltage.

multiple reflection algorithm a tech-nique for the solution of nonlinear device-circuit interactions that uses a time domainrepresentation of the nonlinear device wave-forms and a frequency domain representationof the linear circuit waveforms coupled to-

gether with an ideal zero length transmissionline.

multiple scattering strong interaction,with multiple hits, of the light wave with ahighly inhomogenous medium. This pro-duces large changes both in the amplitude,phase, and state of polarization of the wave.

multiple signal classification (MUSIC) amethod in array processing for estimating pa-rameters such as direction of arrival (DOA),based on estimating the noise subspace andexploiting the fact that the true signal (withthe true parameters) belongs to the signalsubspace (the orthogonal complement to thenoise subspace).

multiple sub-Nyquist encoding (MUSE)a technique used in Japanese HDTV systems.

multiple-access time-division a methodfor allowing many users to communicate si-multaneously with a single receiver by as-signing transmissions to different time slots.The channel, which is shared among theusers, is divided into successive frames,which are subdivided into time slots. Eachuser (or packet) is assigned a specific timeslot within the current frame.

multiple-input–multiple-output systema system that transforms two or more in-put signals to two or more output signals.Also known as SISO system and single-variable (SV) system. See alsosystem,single-input–single-output system.

multiple-operand instruction a com-puter instruction that contains two or moredata elements that are used while executingthe instruction, i.e., ADD Z, X, Y is a multipleoperand instruction that specifies the valuesof X and Y are added together and stored intoZ.

multiple-stage decision making decisionmaking involving future operation of the sys-tem, as in the case of closed-loop feedback


control or limited-lookahead control, wherepossible future measurements and decisioninterventions are taken into account whenconsidering the decision taken at a giventime.

multiplex (1) to use a single unit formultiple purposes, usually by time sharingor frequency sharing.See alsomultiplexer,multiplexing.

(2) the armature winding of a commu-tated electrical machine in which multiple,identical coil windings are placed on the ro-tor. In general, the number of the “plex”describes the total number of parallel wind-ings between brush positions and, thus, alsothe multiplier on the number of parallel pathsbetween brushes that would be provided bya simplex winding. For example, a duplexwinding will have twice as many parallelelectrical paths between brushes as a sim-plex winding, a triplex winding will havethree times the number of paths, etc.Seealsosimplex, duplex, reentrancy.

multiplexer a combinational logic devicewith many input channels and usually oneoutput, connecting one and only one inputchannel at a time to the output.

multiplexer channel a computer I/Osubsystem that allows multiple slow tomedium speed devices to transfer data to orfrom the computer concurrently.See alsobyte multiplexer channel, block multiplexerchannel, selector channel.

multiplexing (1) the process of transmit-ting a large number of information units overa smaller number of channels or lines. Forexample, if we haveN independent signalsthat we want to transmit, then without usinga multiplexer we needN independent chan-nels to do so. Using a multiplexer to controlthe flow of these signals in only one channelreduces the number of wires, thus decreasingcost and increasing efficiency. Multiplexingis the superimposition of multiple signals tomake up one signal. This is done to make

Multiplexing.

the transmission of the signals efficient. Sig-nals are multiplexed at the sending end ofcommunication systems, and demultiplexedat the receiving end, in order to obtain theoriginal signals.

(2) of or being a communication systemthat can simultaneously transmit two or moremessages on the same circuit or radio chan-nel.

multiplication pattern for arrays of sim-ilar elements, it is the product of the patternof a single element and the array factor of thearray.

multiplicative acousto-optic processingacousto-optic signal processing where thelight is repeatedly modulated by a sequenceof acousto-optic devices to result in a mul-tiplication operation of the individual lightamplitude or intensity modulations.

multiplier an electronic system or com-puter software which performs the multipli-cation calculation.

multiplying D/A a D/A conversion pro-cess where the output signal is the productof a digital code multiplied times an analoginput reference signal. This allows the ana-log reference signal to be scaled by a digitalcode.

multipoint bus a bus to which multiplecomponents may be attached. The PCI busis an example of a multipoint bus. See alsopoint-to-point bus.

multiport a circuit presenting multiple ac-cess ports.


multiport memory one memory mod-ule can be accessed by two devices simul-taneously. One example is a video memory,which is interfaced to a graphics co-processoras well as the video interface of the monitor.

multiprocessor a computer system thathas more than one internal processor capa-ble of operating collectively on a computa-tion. Normally associated with those systemswhere the processors can access a commonmain memory.

There are some additional stipulations fora “genuine multiprocessor:”1. It must contain two or more processors ofapproximately comparable capabilities.2. All processors share access to a common,shared memory.3.All processors share access to common I/Ofacilities.4. The entire system is controlled by a singleoperating system (OS).

See alsomultiple instruction multipledata architecture, operating system, sharedmemory architecture.

multiprogramming a system that allowsthe processor to execute two or more pro-grams concurrently.

multipulse converter a three-phase con-verter that generates more than six pulses ofDC per cycle. Multiple converters are con-nected so that the harmonics generated byone converter are canceled by harmonics pro-duced by other converters, in order to reducethe line harmonics and improve the systempower quality.

multirate finite impulse response filtera finite impulse response (FIR) filter in whichthe sampling rate is not constant.

multirate signal processing a system inwhich there is at least one change of samplingrate. Typically an input signal is split into twoor more sub-signals each with a sampling ratelower than the input signal.

multiresolution analysis analysis methodthat decompose a signal into components atdifferent resolution level; the fine to coarsefeatures are revealed in the fine to coarse res-olution components.

multiresolution coding coding schemesthat involve a multiresolution structure.

multiscale characterized by a parameteri-zation or decomposition in scale, as opposedto (for example) time or frequency, in partic-ular, an algorithm that computes or analyzesa function or an image at multiple resolu-tions. See alsowavelet, multigrid, Laplacianpyramid, quadtree, resolution.

multispectral image an image that con-tains information from more than one rangeof frequencies. In this sense, an RGB colorimage is multispectral in that it contains red,green, and blue information. More usually,multispectral refers to frequencies of suchnumber and type that it presents difficulties ininterpretation and display. Adding infrared,ultraviolet, and radar information to an RGBimage creates a problem, since the data can-not be simply drawn on a display. The use ofmany frequencies also provides a wealth ofinformation sometimes needed to make finedistinctions between regions in an image, es-pecially in satellite imagery.See alsodatafusion, sensor fusion.

multispeed motor a motor that can beoperated at any one of two or more definitespeeds. For DC and induction motors, thespeed settings are practically independent ofthe load, although the speed may vary withload for certain types of motors. Multispeedinduction motors typically have two or moresets of windings on the stator with a differentnumber of poles, one of which is excited atany given time.

multistage depressed collector klystrona specially designed klystron in which de-creasing voltage zones cause the electronbeam to be reduced in velocity before striking


the collector element. The effect is to reducethe amount of heat that must be dissipated bythe device, improving operating efficiency.

multistage detection an iterative detec-tion strategy where increasingly more reli-able tentative decisions are made for each it-eration (stage) of the detection algorithm.

multistage interconnection network aninterconnection network built from smallswitches (often with fan-in and fan-out of 2);if 2n modules are connected on each side ofthe network, 2n stages are required.

multistage subset decoding in thisscheme decoding is done, first on the low-est partition level of the signal set and thengradually on the higher level with decoded in-formation flow from the lower to the higherlevel.

multistage vector quantization a methodfor constrained vector quantization whereseveral quantizers are cascaded in order toproduce a successively finer approximationof the input vector. Gives loss in perfor-mance, but lower complexity than optimal(single-stage) vector quantization.

multistage VQ See multistage vectorquantization.

multithreaded several instruction streamsor threads execute simultaneously.

multitone testing a measurement tech-nique whereby an audio system is charac-terized by the simultaneous application of acombination of sine waves to a device undertest.

MUSE Seemultiple sub-Nyquist encoding.

MUSIC Seemultiple signal classification.

mutual coupling electromagnetic inter-action between the elements in a phased ar-ray antenna in which the radiation from oneelement causes distortion of the current dis-tributions or aperture field distributions of theother elements. Feed network mutual cou-pling occurs when the amplitude and phaseof the excitation at a feed point are alteredby the presence of reflected waves in the net-work.

mutual exclusion a synchronization prob-lem requiring that two or more concurrent ac-tivities do not simultaneously access a sharedresource.

mutual inductance the property that ex-ists between two current-carrying conductorswhen the magnetic lines of force from onelink with those from another.

mutual information information theo-retic quantity representing the amount ofinformation given by one random variableabout another. For two discrete random vari-ablesX andY , with joint probability massfunction p(x, y) and marginal probabilitymass functionsp(x) andp(y), respectively,the mutual informationI (X;Y ) is given by

I (X;Y ) =∑x∈X

∑y∈Y

p(x, y) logp(x, y)

p(x)p(y)

whereX andY are the support sets ofX andY respectively.See alsodifferential entropy,relative entropy.

mutually pumped phase conjugator(MPPC) a phase conjugator in whichtwo phase conjugate waves are generated si-multaneously when two beams are incidentinto the phase conjugator. With a mutu-ally pumped phase conjugator, two incidentlaser beams mutually pump the photorefrac-tive medium and produce two phase conju-gate waves. There are many physical mecha-nisms that can yield mutually pumped phaseconjugation, for example, MPPC via self-oscillation in four-wave mixing, MPPC with


stimulated scattering, MPPC with ring res-onator, etc.

MV multi-variable system. Seemulti-input–multi-output system.

MVA interrupting rating the interrupt-ing rating of a device expressed in terms ofmegavolt-amperes. The conversion betweenfault volt-amps and fault current in three-phase systems isVA = 1.73 — operating

RMS line-line voltage — largest phase faultcurrent expressed in RMS symmetrical am-peres. Power circuit breakers can have sep-arate MVA and maximum current interrupt-ing ratings, with adjustments when the circuitbreaker is operated below rated voltage.

MWIR laser a laser producing light in themidwavelength (2 – 5 micron) range, usefulin chemical species identification and otherapplications.


Nn-channel MOSFET a MOSFET wherethe source and drain are composed of heav-ily doped n-type semiconductor regions on ap-type surface. Electrons form drain-sourcecurrent when the applied gate and substratepotentials invert the p-type surface betweenthem.

n-well a region of n-type semiconductorlocated at the surface of a p-type substrate(or larger P-well) usually created in order tocontain p-channel MOSFETs.

NA Seenumerical aperture.

Nakagami fading a general model of fad-ing used in the modeling of radio communi-cation channels, introduced by Nakagami in1960, which includes as a special case theRayleigh model of fading.

NaN acronym for not a number. Used inIEEE floating-point representations to des-ignate values that are not infinity or zero orwithin the bounds of the representation.

NAND gate a logic circuit that performsthe operation equivalent to the AND gate fol-lowed by the inverter. The output of a NANDgate is low only if all inputs are high.

nano prefix for metric unit that indicatesdivision by one billion.

nanocell cell of radius up to 10 m.Seealsocell.

nanoelectronics a future integrated cir-cuit technology characterized by minimumlateral feature sizes in the range 1 – 100nanometers. Also pertains to ICs utilizingquantum devices.

nanolithography Lithography involvingthe printing of ultra-small features, typicallyon the order of nanometers in size.

nanometer a metric measure of distanceequivalent to one billionth of a meter.

narrow band referring to a bandwidth of300 hertz or less.

narrow-band fading fading in which thecommunication channel amplitude can be as-sumed independent of fading. Occurs whenthe transmitted signal bandwidth is consid-erably less than the coherence bandwidth ofthe channel.

narrow-band filter narrow-band filtersare those which pass signals undistorted inone or a set of narrow frequency bands andattenuate or totally eliminate signals in theremaining frequency bands.

narrow band FM frequency modulationscheme where the ratio of peak frequency de-viation to the frequency of the modulatingsignal is smaller than 0.2.

narrow-band interference (NBI) in-terference from a source that occupies asmaller bandwidth that the desired signal.One example is a single sinusoid (or anarrow-band communication system) inter-fering with a wide-band system, possibly aspread-spectrum system.

nasal a sonorant phone that is producedwith the vocal tract closed and the velumopen. Nasals have a formant-like spectrum,but much information is carried out duringthe transition for the opening of the velum.Hence, the design of very effective auto-matic systems for the recognition of partic-ular nasals (e.g.,/m/, /n/) must take intoaccount these brief transitions.

Nash equilibrium a noncooperative equi-librium in nonzero-sum games defined by anoutcome for each player that cannot be im-


proved by altering his decision unilaterally.The roles of all players are symmetric in thesense that they make their decisions indepen-dently and none of them dominates any other.If Ji, i = 1,2, . . . , N is a cost function ofith player in N-person game anddi his strat-egy then N-tupled∗

1, d∗2, . . . , d

∗N constitutes

a Nash equilibrium solution if

J1(d∗

1, d∗2, . . . , d

∗N

) ≤ J1(d1, d

∗2, . . . , d

∗N

)J2(d∗

1, d∗2, . . . , d

∗N

) ≤ J2(d∗

1, d2, . . . , d∗N

)...

JN(d∗

1, d∗2, . . . , d

∗N

) ≤ JN(d∗

1, d∗2, . . . , dN

)for any admissibledi; i = 1,2, . . . , N . Thenonzero-sum game may admit more than oneNash equilibrium solution with outcomesdifferent in each case. Since the total order-ing among N-tuples of numbers does not ex-ist, it is usually not possible to declare oneof them as the most favorable. Nevertheless,some of strategies may be viewed as betterthan other ones if the respective outcomesare in partial order. With this notion of bet-terness, a Nash equilibrium strategy N-tupleis admissible if there exists no better one. Inthe case when the game is played, many timesin the same conditions, the Nash equilibriummay be defined for average values of the costfunctional and mixed strategies.

NASTRAN a widely used computer codefor mechanical and structural analysis, suchas for opto-mechanical and thermal analyses.

National Electrical Code (NEC) a stan-dard for electrical construction, publishedby the National Fire Protection Association(NFPA 70-1). The National Electrical Codeis often adopted by local jurisdictions andused by their electrical inspectors.

National Electrical Manufacturers Asso-ciation (NEMA) an electrical trade as-sociation that establishes standards for elec-trical equipment. In the case of electric mo-tors, NEMA establishes standard frame sizes,starting torque, starting current, and otherquantities for a given horsepower machine.

National Fire Protection Association(NFPA) sponsor and publisher of the Na-tional Electrical Code and other safety stan-dards.

National Television System Committee(NTSC) a body that recommended thestandard for colored television broadcasts inthe U.S. in 1953. NTSC video contains 525lines, a field rate of 59.94 fields/second, ahorizontal frequency of 15,734.264 Hz, andan interlaced color subcarrier frequency of3.579545 MHz. The NTSC format is also inuse by many countries other than the U.S.

natural broadening spectral broaden-ing of a transition in a laser medium dueto spontaneous decay; sometimes called life-time broadening.

natural commutation commutation ofcurrent from one switching device to anotherin a power electronics converter at the instantthe incoming voltage has a higher potentialthan that of the outgoing wave, without theaid of any commutation circuitry.See alsocommutation.

natural constraint constraint that resultsfrom the particular mechanical and geomet-ric characteristics of the task configuration.Used to define a particular situation occurringbetween the manipulator end-effector and thework environment. These constraints are de-fined in terms of linear and angular veloci-ties and forces and/or torques that are speci-fied to execute the task. Natural constraintsare intrinsically associated with the particu-lar task. For example, consider a task thatan end-effector moves along a rigid surfaceand is not free to move through that surface.In this particular situation, a natural positionconstraint exists. In addition, if the surfaceis frictionless, the end-effector is not free toapply arbitrary force tangent to the surface,and hence a natural force constraint exists.

natural frequency the frequency of anyoscillating term in the response of the linear


system which is due to the initial value ofthat system. For linear time invariant sys-tems the total response of the system canbe represented as the sum of the zero-inputresponse plus the zero state response. Thecharacteristic polynomial of a system (rep-resenting the zero input response),Q(λ) =(l − l1)(l − l2) . . . (l − ln) = 0 hasn roots;these roots are called the Natural frequencies.Natural frequencies are also known as thecharacteristic values, eigenvalues, and char-acteristic roots of the system.

natural laser a laser occurring in nature.

nautical mile (nm) 1,852 meters, exactly.Approximately 1 minute of arc on the earth’ssurface.

NBI Seenarrow-band interference.

NDR Seenegative differential resistance.

near field (1) in antennas, the electromag-netic field that is in the vicinity of an antennawhere the angular field distribution is depen-dent on the distance from the antenna.

(2) in optics, region close to a diffractingaperture where neither the form nor the sizeof a transmitted beam have changed signifi-cantly from their values at the aperture.

near infrared light in the wavelength re-gion of the infrared electromagnetic spec-trum which is adjacent to the visible spec-trum. The spectral range of the near infraredregion is typically considered to be 700 –2500 nm.

near–far effect large dissimilarities in re-ceived power from different users, therebycausing trouble in detecting the weaker users.Commonly found in DS-CDMA systems.

near–far resistant a device, e.g., a detec-tor or parameter estimator, that is insensitiveto the near–far effect.

nearest neighbor algorithm a method ofclassifying samples in which a sample is as-signed to the class of the nearest training setpattern in feature space, a special case of theK-nearest neighbor algorithm.

NEC SeeNational Electrical Code.

Necker cube a classical example of am-biguous wireframe geometrical figures. TheNecker cube can take either of two perspec-tives, as it were seen from above or from be-low.

NEESLA See nonuniformly excitedequally spaced linear array.

negation operator See complementoperator.

negative definite function a scalar func-tion V (x, t) where−V (x, t) is positive def-inite.

negative differential resistance (NDR)a condition where the slope of the current vs.voltage characteristic is negative. Negativedifferential devices include quantum devicessuch as tunnel diodes and resonant tunnelingstructures. Negative differential resistancecan also be obtained in a variety of circuittopologies.

negative photoresist a photoresist whosechemical structure allows for the areas thatare exposed to light to develop at a slowerrate than those areas not exposed to light.

negative resistance oscillator oscillatorcircuit that functions based on an active de-vice where the device is biased under certainconditions and terminated with impedancelevels such that it exhibits a negative resis-tance at a particular frequency or frequencyrange.

negative semidefinite function a scalarfunctionV (x, t), where−V (x, t) is positivesemi-definite.


negative sequence the set of balanced butreverse sequence (acb) components used insymmetrical component analysis. Normalload currents contain no negative sequencecurrent.

negative sequence overcurrent relay aprotective relay that senses and operates onnegative sequence overcurrent. Typical ap-plications include the sensing of unbalancedfaults and the protection of synchronous andinduction machines from rotor overheating.

negative transition angle the angular por-tion of the time based output signal that hasa negative slope, expressed in degrees. Thisquantity could be loosely interpreted as the“trailing edge” angle.

negative-positive-zero (NPO) an ultra-stable temperature coefficient (± 30 ppm/◦Cfrom −55◦ to 125◦C) temperature-compen-sating capacitor.

negative-sequence impedance the impe-dance offered by a circuit when negative-sequence currents alone flow through it, ex-pressed in ohms. The impedance is com-plex, with its real part being the circuit resis-tance and imaginary part, which is a functionof frequency and inductance referenced asnegative-sequence reactance, also expressedin ohms.

negative-sequence reactance inductivereactance offered by a circuit for the flow ofnegative-sequence currents alone. Expressedin ohms, the inductive reactance is a functionof frequency and the inductance of the circuitto negative-sequence current flow.See alsonegative-sequence impedance.

neighborhood in self-organizing systeman area surrounding the winner in self-

organizing competitive learning. Accordingto so-called Mexican-hat interaction betweenthe brain cells, Kohonen proposes that weightupdating should be conducted not only forthe winner but also for its neighbors.See

also Mexican-hat function, self-organizingalgorithm.

neighborhood operation an operation,such as averaging or median filtering, that isdependent upon the locality of samples in asignal, not the signal as a whole. Also calleda window operation.Seewindowing.

NEMA See National ElectricalManufacturers Association.

NEMA code letter the nameplate let-ter designation used to indicate the inputkilowatt-amp rating of a motor under locked-rotor, or starting conditions.

NEMA size a standard-size device,such as a motor controller. The NEMAsize establishes the rating of the device.See alsoNational Electrical ManufacturersAssociation.

NEMA type for induction motors NEMAestablishes five types of induction motors (A,B, C, D, and E) that have different torque-speed characteristics to account for varioustypes of loads.

nematic the type of liquid crystal in whichthe molecular chains align; such alignmentcan be controlled across the liquid crystal ifit can be constrained at the boundaries.

nematic liquid crystal one of the state ofliquid crystal materials, where the elongatedliquid crystal molecules are all oriented in thesame direction within a layer. The moleculescan be reoriented with an external electricfield, allowing use in displays and spatiallight modulators.See alsoliquid crystal onsilicon, twisted nematic.

neocognitron a biologically inspired hier-archical network developed primarily for therecognition of spatial images. The networkhas up to nine layers. The lower layers re-spond to simple features and the higher layers


respond to more complex features. Learningcan be unsupervised or supervised.

neodymium-iron-boron a high energymagnetic material composed of the threenominal elements and other additives, char-acterized by a high residual induction andhigh coercivity. NdFeB has a high magnetictemperature coefficient, which is undesirablefor high-temperature use.

neper a natural logarithm of a ratio.

nested structure an information structurein which each player has an access to theinformation acquired by all his precedents,i.e., players situated closer to the beginningof the decision process. If the difference be-tween the information available to a playerand his closer precedent involves only hisactions, then the structure is ladder-nested.This structure enables decomposition of thedecision process onto static games that in turnresults in recursive procedure for its solution.For dynamic infinite discrete-time decisionprocesses, the ladder-nested structure resultsin the classical information pattern.

nested subroutine a subroutine called byanother subroutine. The programming tech-nique of a subroutine calling another subrou-tine is called nesting.

network analyzer a test system that mea-sures RF or microwave devices in terms oftheir small signal characteristics. The resultsare typically presented as two-port parametermatrices such as s-parameters, y-parameters,or z-parameters.

network distribution system a distribu-tion system in which each load is supplied bymore than one path.

network function the ratioH(s) of theLaplace transform of the output function tothe Laplace transform of the input function.See alsotransfer function.

network interface card (NIC) the physi-cal device or circuit used to interface the net-work with a local workstation or device.

network measurement division (NMD)connector a special coaxial connec-tor type developed by Hewlett Packard fortest cables used in conjunction with an s-parameter test set.

network pruning (1) the removal froma network of interconnections and/or neuralunits that are identified (after training) as be-ing either unnecessary or unimportant. Afterpruning, the network must be retrained.

(2) the process of cutting out nodes orunwanted connections in a neural network,to save computation or computational hard-ware.See alsoartificial neural network.

network time constant minimum of theallowable transfer delay and the time requiredto fill up the buffer in the node. The NTC isindicative of how long transients last inside anetwork node. The NTC is the interval overwhich network traffic must be averaged.

network weight a scalar value represent-ing the strength of the connection betweenthe output of one neuron and the input of an-other.

neural network a parallel distributed in-formation processing structure consisting ofprocessing elements, called neurons, inter-connected via unidirectional signal channelscalled connections. Neurons can possess alocal memory and can carry out localized in-formation processing operations. Each neu-ron has a single output connection. The neu-ron output signal can be of any mathemati-cal type desired. The information processingthat goes on within each processing elementcan be defined arbitrarily, with the restrictionthat it must be completely local (it dependsonly on the current values of the input sig-nals arriving at the processing element andon values stored in the processing element’slocal memory, e.g., weights).


neural tree a tree-structured neural net-work. Such networks arise in the applicationof certain kinds of constructive algorithm.

neuro-fuzzy control system control sys-tem involving neural networks and fuzzy sys-tems, or fuzzy neural networks.See alsoadaptive fuzzy system.

neuron a nerve cell. Sensory neuronscarry information from sensory receptors inthe peripheral nervous system to the brain;motor neurons carry information from thebrain to the muscles.

neutral axis the axis near which the di-rection of the velocity of armature conduc-tors is exactly parallel to the magnetic fluxlines so that the EMF induced in the conduc-tors is zero. This axis, also referenced as themagnetic neutral axis, shifts in the directionof rotation of the machine as a generator ormotor. The amount of shift depends on thearmature current and hence on the load of themachine.

neutral a conductor which completes theelectric circuit from the load to the sourcein three-phase Y-connected and single phaseAC electric power systems, typically at ornear the potential of the earth.

neutral plane Seeneutral axis.

neutral zone in permanent magnets, aplane through which all lines of flux are par-allel to the direction of magnetization, i.e.,the plane between north and south poles.

neutrino laser laser in which the ampli-fied field consists of neutrinos rather thanelectromagnetic waves; suggested impor-tance in the early universe.

New York City blackout a particularlydisastrous failure of the power system in NewYork City in 1977.

Newton’s method a class of numeri-cal root-finding methods; that is, to solvef (x) = 0. The methods are based on iter-atively approximatingf using a low orderpolynomial and finding the roots of the poly-nomial. The most commonly used first-ordermethod iterates the following equation:

xn+1 = xn − f (xn)

f ′(xn)

wheref ′ represents the derivative off .

Newton–Euler recursive algorithm usedin robotics, to calculate joint generalizedforces as a result of two recursions alongthe kinematical structure of the robot. Firstone, a forward recursion, calculates spatialvelocities and accelerations from the base ofthe manipulator towards its tip. Second one,a backward recursion calculates the spatialforces and torques along the structure in theopposite direction. Projection of the spatialforce and torques at the joint axis results inthe generalized force at the joint. The result-ing Newton–Euler equations of motion arenot in closed form as opposite to the result ofusing Lagrange formalism to the dynamicalsystem of the robot itself.See alsoLagrangeformulation.

Newton–Raphson method a numericalmethod for finding the solution to a set of si-multaneous nonlinear equations. Variationsof this method are commonly used in circuitsimulation programs.

Neyman–Pearson detector a detectorthat minimizes the miss probability withinan upperbound constraint on the false-alarmprobability.

NFPA See National Fire ProtectionAssociation.

nibble four bits of information.

nibble-mode DRAM an arrangementwhere a dynamic RAM can return an extra


three bits, for a total of four bits (i.e., a nib-ble), with every row access.

The typical organization of a DRAM in-cludes a buffer to store a row of bits insidethe DRAM for column access. Additionaltiming signals allow repeated accesses to thebuffer without a row-access time.See alsotwo-dimensional memory organization.

NIC Seenetwork interface card.

Nichols chart in control systems, a plotshowing magnitude contours and phase con-tours of the return transfer function referredto ordinates of logarithmic loop gain and ab-scissae of loop phase angle.

night vision Seescotopic vision.

NMD connector Seenetwork measurementdivision connector.

NMI Seenonmaskable interrupt.

NMR nuclear magnetic resonance.Seemagnetic resonance imaging.

NMOS Seen-channel MOSFET.

no fetch on write strategy in the write-through cache policy where a line is notfetched from the main memory into the cacheon a cache miss if the reference is a write ref-erence. Also called non-allocate on write, asspace is not allocated in the cache on writemisses.

no load tap changer device that pro-vides for changing the tap position on atapped transformer when the transformer isde-energized. Different taps provide a differ-ent turns ratio for the transformer.

no voltage holding coil a holding coil thatkeeps the main-line contactor closed on zerovoltage conditions. DC motor controllersthat contain this feature are used in placeswhere the motor is vital to the operation ofa process. These controllers can maintain

control to the motor under momentary linepower loses, by using the CEMF of the coast-ing armature to keep power to the main-linecoil/contactor. If power to the motor con-troller is not restored within a short periodof time, the motor coasts to a speed where itcan no longer keep the main-line contactorclosed. At this point, the m-coil drops outto insure starting resistors are placed back inthe circuit.

no-load test measurement of input param-eters of an induction motor while running atnearly synchronous speed, with zero outputon the shaft. This test is used to determine themagnetizing reactance of the motor equiva-lent circuit. See alsoopen-circuit test.

no-op a computer instruction that per-forms no operation. It can be used to reservea location in memory or to put a delay be-tween other instruction execution.

no-write allocate part of a write policythat stipulates that if a copy of data being up-dated are not found in one level of the mem-ory hierarchy, space for a copy of the updateddata will not be allocated in that level. Mostfrequently used in conjunction with a write-through policy.

nodal cell cell with a radius of up to 300m. Typically an isolated cell acting as a highcapacity network node.See alsocell.

nodal system a secondary system of equa-tions using nodal voltages as variables.

node a symbol representing a physicalconnection between two electrical compo-nents in a circuit.See alsograph.

node analysis a circuit analysis techniquein which KCL is used to determine the nodevoltages in a network.

noise (1) any undesired disturbance,whether originating from the transmissionmedium or the electronics of the receiver it-


self, that gets superimposed onto the originaltransmitted signal by the time it reaches thereceiver. These disturbances tend to interferewith the information content of the originalsignal and will usually define the minimumdetectable signal level of the receiver.

(2) any undesired disturbance superim-posed onto the original input signal of anelectronic device; noise is generally cat-egorized as being either external (distur-bances superimposed onto the signal before itreaches the device) or internal (disturbancesadded to the signal by the receiving deviceitself). See alsonoise figure, noise powerratio. Some common examples of externalnoise are crosstalk and impulse noise as a re-sult of atmospheric disturbances or manmadeelectrical devices. Some examples of inter-nal noise include thermal noise, shot noise,1/f noise, and intermodulation distortion.

noise bandwidth an equivalent band-width,WN , of a system expressed as

WN =∫ inf

0 |H(ω)|2|H(ω0|2

,

whereH(ω) is the transfer function of thesystem andω0 is the central frequency.

noise circles circles of constant noise fig-ure plotted on the Smith chart. These circlescan be used to graphically impedance matcha device to achieve a desired noise figure.

noise clipping a process by which highnoise peaks are clipped or limited to elimi-nate most of their energy, thereby removingthe worst effects of impulse noise.

noise factor Seenoise figure.

noise figure an indication of the contribu-tion of that component or system to the levelof noise observed at the output. The noisefigure therefore gives an idea of the amountof noise generated within the component orsystem itself. It is usually expressed in deci-bels (dB) and is given by the ratio of the inputsignal-to-noise ratio to the output signal-to-

noise ratio of the component or system. Thisis defined as

Noise Figure (NF)1= (S/N)input

(S/N)output.

In decibels this is given by:

[NF]dB1= 10 log10

[(S/N)input

(S/N)output

]

noise figure meter a system that makesaccurate noise figure readings possible. Itmeasures the amount of the noise added by adevice. Noise figure is the ratio of the signal-to-noise ratio at the input of a device to thesignal-to-noise ratio at the output of that de-vice.

noise floor the lowest input signal powerlevel which will produce a detectable outputsignal in an electrical system. Noise floor isdetermined by the thermal noise generated inthe electrical system.

noise immunity a logic device’s abilityto tolerate input voltage fluctuation causedby noise without changing its output state.

noise power the amount of power con-tained within a noise signal. The noise powercan be found by integrating the power spec-tral density (PSD) of the noise signal overall possible frequencies. If the power spec-tral density is given byS(ω), the power,p,contained in the noise signal, is then given by

p1= 1

2π

∫ ∞

−∞S(ω)dω .

In most cases, noise signals are bandwidthlimited to between sayω1 andω2 and in thesecases the noise power will be given by

p1= 1

2π

∫ ω2

ω1

S(ω)dω .

Generally, the power spectral density is ex-pressed in watts per hertz (W/Hz) and thenoise power is expressed in watts.See alsonoise, noise spectrum.


noise power ratio (NPR) intermodula-tion distortion product generated in nonlin-ear transfer components such as high poweramplifiers (HPAs) amplifying multiple carri-ers. It is defined as a ratio of an averagedpower of white noise having a narrow notch,which represents multiple carriers, to an av-eraged power falling into a narrow bandwidthof notch.

noise rejection the ability of a feedbackcontrol system to attenuate (reduce) the am-plitude of any unwanted signal generated bythe measurement of its output variable.

noise smoothing any process by whichnoise is suppressed, following a comparisonof potential noise points with neighboring in-tensity values, as for mean filtering or medianfiltering.

noise spectrum indicates the frequencycomponents in a noise signal. For idealthermal noise (AWGN), the spectrum is flatacross all frequencies and is referred to asthe noise power spectral density expressed inwatts per hertz (W/Hz).See alsonoise, noisepower.

noise subspace in an orthogonal decom-position of a space, the orthogonal comple-ment to the signal subspace.

noise suppression any process by whichnoise is eliminated or suppressed: a moregeneral term than noise smoothing, since itincludes such processes as noise clipping andband-pass filtering.

noise temperature an alternate represen-tation of noise power. It is generally acceptedthat noise is the result of the random mo-tion of electrons within components in elec-tronic systems. The greater the temperatureof a component, the greater the random mo-tion of electrons within the component, andthis causes an increased instantaneous noisecurrent to be present within the component.With an increase in noise current, the noise

power also increases. The noise power gen-erated within a component atT Kelvin, overa bandwidth ofB hertz is given by

p = kT B

with k = 1.38× 10−23 J/K, which is Boltz-mann’s constant.

From this equation, it is clear that a com-ponent at absolute zero temperature (0 K),generates no noise power. Therefore, givena particular bandwidth,B noise power canalso be expressed as a temperature in Kelvin(K).

noise whitening a process by which noisewhose power spectrum is not white (i.e., notidentical at all frequencies) is brought to thiscondition, e.g., by means of a frequency de-pendent filter.

noiseless source coding See losslesssource coding.

noiseless source coding theorem statesthat any source can be losslessly encodedwith a code whose average number of bitsper source symbol is arbitrarily close to, butnot less than, the source entropy in bits.

noisy channel vector quantization a gen-eral term for methods in vector quantizertransmission for noisy channels.See alsochannel robust vector quantization, channeloptimized vector quantization, channelmatched VQ, redundancy-free channel coding.

noisy channel VQ Seenoisy channelvector quantization.

noisy source coding See lossy sourcecoding.

nominal voltage a number given to a sys-tem to name its classification of voltage, suchas rated values.

non-binary codes codes in which the fun-damental information units or symbols as-


sume more than two values. This is in con-trast to binary codes, for which the funda-mental information symbols are two-valuedor binary, only.See alsobinary code, blockcode, convolutional code.

non-customer-call call placed by passerbyto utility, either emergency personnel or oth-erwise, to indicate that a scenario existswhich is potentially disrupting electrical ser-vice.

non-return-to-zero inverted recordingSeemagnetic recording code.

non-return-to-zero recording Seemag-netic recording code.

non-time delay fuse Seesingle-elementfuse.

non-transposition refers to a three-phaseelectric power transmission line whose con-ductors are not transposedSeetransposition.

nonallocate on write Seeno fetch onwrite.

nonblocking cache a cache that can han-dle access by the processor even though aprevious cache miss is still unfinished.

noncausal system Seecausal system.

noncentrosymmetric medium Seecentro-symmetric medium.

noncoherent integration where only themagnitude of received signals is summed.

noncollinear geometry acousto-optical tun-able filter an acousto-optical tunable fil-ter (AOTF) device that operates similar to aBragg cell where the incident light and acous-tic wave are not collinear and the diffractedlight can be spatially separate from the inci-dent light.

noncollinear geometry AOTF Seenon-collinear geometry acousto-optical tunablefilter.

noncooperative game one of a class ofdecision processes in which decision makers(players) pursue their own interests, whichare at least partly conflicting with others’. Aconflicting situation or collision of interestsresults the situation in which a player has tomake a decision and each possible one leadsto a different outcome valued differently byall players. Depending on the number of de-cision makers involved in the process, thegame may be two-person or multi-person; de-pending on how the outcome is viewed by theplayers, the process may be a zero-sum game(only two-person) or a nonzero-sum game(two- or multi-person). If the order in whichthe decisions are made is important, a gameis dynamic, if not it is static. Dynamics ofthe game is usually related to the access ofthe decision makers to different information.Regarding the number of possible decisionsavailable to the decision makers, a game isfinite or infinite (in this case, the number ofpossible actions is usually a continuum). Dy-namic games are usually formulated in exten-sive form, which for finite games leads to afinite tree structure, while for infinite ones,involves difference (in discrete-time) or dif-ferential (in continuous-time) equations de-scribing an evolution of the underlying deci-sion process.

nondegenerate network a network thatcontains neither a circuit composed only ofcapacitors and/or independent or dependentvoltage sources nor a cutset composed onlyof inductors and/or independent or dependentcurrent sources.

nondegenerate two-wave mixing a gen-eral case of two wave mixing in which thetwo beams are of different frequencies. Intwo-wave mixing, if the frequencies of thetwo laser beams are different, the interfer-ence fringe pattern is no longer stationary. Amoving volume refractive index grating can


still be induced provided that the intensityfringe pattern does not move too fast. Theamplitude of the refractive index modulationdecreases as the speed of the fringe patternincreases. This is related to the finite timeneeded for the formation of the refractive in-dex grating in the photorefractive medium.Such a kind of two-wave mixing is referredto as nondegenerate two-wave mixing.

nondestructive readout when data areread from a particular address in a memorydevice, the contents of the memory at thataddress remain unaltered after the read oper-ation.

nondispersive medium medium in whichthe index of refraction does not vary signifi-cantly with frequency.

nonfuzzy output a function of the fir-ing degrees and the fuzzy outputs of firedfuzzy rules regardless of what defuzzificationmethod is used.

nonhomogeneous linear estimator anestimator which is a nonhomogeneous linearfunction of the data.

noninteraction of photons the nature ofphotons that they do not interact to each other.Unlike electrons governed by Coulomb’slaw, a photon cannot affect another photon. Alight beam passes through another light beamwithout any change. If photons interacted aselectrons do, the image we see from one di-rection would be distorted by other streamsof photons, and we would never understandwhat we see. The noninteraction of photonsprovides parallelism and enables us viewingin free space. On the other hand, it makescontrolling or switching light using light im-possible.

noninvasive sensor the interface deviceof an instrumentation system that measures aphysiologic variable from an organism with-out interrupting the integrity of that organ-ism. This device can be in direct contact with

the surface of the organism or it can measurethe physiologic quantity while remaining re-mote from the organism.

noninvertible system See invertiblesystem.

nonlinear dielectric property the distinctdependence of the electric permittivity of cer-tain dielectric materials on the intensity of anapplied electric field.

nonlinear distortion a change in signalproperties due to circuit transmission char-acteristics not being completely linear. Thismay result in changes in the relative ampli-tude or phase of the various frequency com-ponents of the signal or may result in inter-modulation.

nonlinear effect effect that cannot bewritten as linear functions of the drivingfields

nonlinear electro-acoustic propertynonlinear interaction between the atomic dis-placement and the electric field experiencedin certain materials that would cause mod-ulation effects resulting in the generation ofnew sideband frequencies (called Raman fre-quencies).

nonlinear electro-optic property non-linear changes in the refractive index ofcertain optically transparent materials withchange(s) in the externally applied electricfield.

nonlinear filter Seelinear filter.

nonlinear load an electrical load operat-ing in the steady state that has a current thatis not continuous or in which the impedancechanges during each cycle of the supply volt-age.

nonlinear magnetic property nonlineardependence of the magnetic susceptibility of


certain materials on the intensity of an ap-plied magnetic field.

nonlinear mean-square estimate the op-timum estimate under the mean-square per-formance criterion.

nonlinear medium medium in which theconstitutive parameters are not functions ofthe electric or magnetic field amplitudes.

nonlinear model a small signal modelthat includes internal components, essen-tially voltage controlled current sources andcapacitors, which depend on the applied volt-ages to the transistors. Nonlinear models arequite useful for understanding the behaviorof transistors.

nonlinear optics (1) the study of opti-cal phenomena that occur through the use oflight fields (e.g., laser beams) sufficiently in-tense to modify the optical properties of amaterial system. From a formal perspective,these phenomena are nonlinear in the sensethat the material polarization depends on theapplied electric field strength in a nonlinearmanner.

(2) optical processes in materials capableof producing output light with wavelengthsdifferent from that of the input light.

nonlinear response the characteristic ofcertain physical systems that some outputproperty changes in a manner more complexthan linearly in response to some applied in-put.

nonlinear Schrodinger equation thefundamental equation describing the propa-gation of short optical pulses through a non-linear medium, so-called because of a formalresemblance to the Schrodinger equation ofquantum mechanics.

nonlinear susceptibility a quantity de-scribing the nonlinear optical response of amaterial system. More precisely, the non-linear susceptibility of ordern, often des-

ignatedχn, is defined through the relationPn = cχnEn, whereE is the applied elec-tric field strength andPn is thenth order con-tribution to the polarization. The coefficientc is of order unity and differs depending onthe conventions used in defining the electricfield strength. The nonlinear susceptibilityof order n is a tensor of rankn+ 1.

nonlinear system a system that does notobey the principle of superposition.

The superposition principle states that if

y(t) = f (x(t)), y1(t) = f (x1(t))

andy2(t) = f (x2(t))

and if

x(t) = αx1(t)+ βx2(t)

theny(t) = αy1(t)+ βy2(t)

See alsolinear system.

nonlinear-optic logic gate an opticallogic gate utilizing thresholding capability ofnonlinear-optic materials.

nonlinear-optic material a material fromthe following group: GaAs, ZnS, ZnSe,CuCl, InSb, InAs, and CdS. When thesetransparent materials are illuminated with in-put light, their refractive indices depend non-linearly on the illuminating light intensity.The relation of refractive index and illumi-nating intensity resembles a threshold func-tion. The phase of output light transmittingthrough the nonlinear material will be modu-lated by the refractive index of material. Thephase modulation is translated into intensityvariation using an interferometer.

nonlinearity response of a medium thatis not directly proportional in magnitude tothe magnitude of applied fields.

nonlocal optics due to spatial dispersion,the dielectric function of a material may de-pend on the wave vector. In such a case, the


relation between the electric field and the cur-rent density will be nonlocal. The nonlocaleffect will play an important role in modelingthe interaction of electromagnetic waves withmetals when the frequency of the electromag-netic wave is around the plasma frequency ofthe metal.

nonmaskable interrupt (NMI) an ex-ternal interrupt to a CPU that cannot bemasked (disabled) by an instruction.See alsomaskable interrupt.

nonorthogonal projection a nonorthog-onal projection is a projection of a vectorconto a vectorb in a direction orthogonal tog = Qgb, whereQg is an arbitrary rotationmatrix in the plane spanned byb andc. It isdefined as

cb =(

bTQgcbTQgb

)b

nonorthogonal wavelets Seeorthogonalwavelet.

nonparabolic band refers to an energyband in which the energy dependence uponthe momentum deviates from the classicalquadratic behavior.

nonparametric estimation an estimationscheme in which no parametric description ofthe statistical model is available.

nonradiating dielectric (NRD) waveguidea rectangular dielectric waveguide bound bytwo infinite parallel metallic plates. The di-electric constant and the separation betweenthe plates is chosen such that the electromag-netic wave is above cut-off and propagatingin the dielectric, and is below cut-off outsidethe dielectric, and does not radiate out of theparallel plates.

nonrecurring engineering (NRE) coststhe foundry charges the ASIC customer.These costs include engineering time, mak-ing the masks, fabricating one lot of wafers,

and packaging and testing the prototypeparts.

nonrecursive equation See recursiveequation.

nonredundant number system the sys-tem where for each bit string there is one andonly one corresponding numerical value.

nonremovable disk Seeremovable disk.

nonrigid body motion Seerigid bodymotion.

nonsaturated region Seeohmic region.

nonseparable data Seeseparable data.

nonstate variable network variable thatis not a state variable.

nonuniform sampling See uniformsampling.

nonuniformly excited equally spaced lin-ear array (NEESLA) an antenna arrayin which all the centers of the antennas arecollinear and equally spaced, but each an-tenna can have a unique harmonic amplitudeand can have a unique phasing.

nonunity feedback the automatic controlloop configuration shown in the figure.

Nonunity feedback configuration.

nonvolatile Seenonvolatile memory.

nonvolatile memory (1) memory that re-tains its contents even when the power supplyis removed. Examples are secondary mem-ory and read-only memory.


(2) the class of computer memory that re-tains its stored information when the powersupply is cut off. It includes magnetic tape,magnetic disks, flash memory, and mosttypes of ROM.

nonvolatile random-access memory(NVRAM) SRAM or DRAM with non-volatile storage cells. Essentially each stor-age cell acts as a normal RAM cell whenpower is supplied, but when power is re-moved, an EEPROM cell is used to capturethe last state of the RAM cell and this stateis restored when power is returned.

nonzero-sum game one of a class ofgames in which the sum of the cost func-tions of the players is not constant. A num-ber of players is not limited to two as inzero-sum games, and in this sense one maydistinguish between two-person and multi-person nonzero-sum games. Since the objec-tives of the players are not fully antagonis-tic, cooperation between two or more deci-sion makers may lead to their mutual advan-tage. However, if the cooperation betweenthe players is not admissible because of in-formation constraints, lack of faith, or impos-sibility of negotiation, the game is noncoop-erative and an equilibrium may be defined ina variety of ways. The problem of solvingthe nonzero-sum game differs from that ofthe zero-sum game. The most natural solu-tion is Nash equilibrium, which is relevantto the saddle point equilibrium in zero-sumgames. Its main feature is that there is noincentive for any unilateral deviation by anyone of the players and their roles are sym-metric. If one of the players has the ability toenforce his strategy on the other ones, thena hierarchical equilibrium called von Stack-elberg equilibrium is rational. Yet anothersolution for the player is to protect himselfagainst any irrational behavior of the otherplayers and adopt min-max strategy by solv-ing a zero-sum game, although the originalis nonzero-sum.

NOR gate a logic circuit that performs theoperation equivalent to the OR gate followedby the inverter. The output of a NOR gate islow when any or all inputs are high.

norator an idealized two-terminal net-work element for which the voltage acrossit and the current through it are determinedby the network to which it is connected.

norm a vector spaceV , anorm is a real-valued functionN defined onV , satisfyingthe following requirements for everyv,w ∈V and scalarλ:1.N(0) = 0 and forv 6= 0,N(v) > 0.2.N(λv) = |λ|N(v).3. N(v + w) ≤ N(v) + N(w). The mostusual norms are theL1, L2, andL∞ norms;for a vectorx with n coordinatesx1, . . . , xnwe have

‖x‖1 =n∑k=1

|xk|,

‖x‖2 =( n∑k=1

|xk|2)1/2

,

and‖x‖∞ = nmaxk=1

|xk|;

for a functionf defined on a setE, thesenorms become

‖f ‖1 =∫E

|f (x)| dx,

‖f ‖2 =(∫

E

|f (x)|2 dx)1/2

,

and

‖f ‖∞ = min{r ≥ 0 | |f (x)|≤ r almost everywhere}.

Given a normN , the functiond defined byd(x, y) = N(x − y) is a distance functionon V . Seechessboard distance, Euclideandistance, Manhattan distance.

normal demagnetization curve the sec-ond quadrant portion of the hysteresis loopgenerated when magnetic induction (B) is


plotted against applied field (H ), which ismathematically related to the intrinsic curve;used to determine the performance of a mag-net in a magnetic circuit.

normal dispersion increase of index ofrefraction with increasing frequency; tendsto occur near the center of amplifying transi-tions or in the wings of absorbing transitions.

normal distribution See Gaussiandistribution.

normal tree a tree that contains all theindependent voltage sources, the maximumnumber of capacitors, the minimum numberof inductors, and none of the independentcurrent sources.

normal vector a vector perpendicular tothe tangent plane of a surface. Also, a vec-tor which is geometrically perpendicular toanother vector.

normalization (1) the process of refor-matting a floating point number into a stan-dard form at the completion of a floating pointarithmetic operation.

(2) the process of equalizing signal ener-gies, amplitudes, or other features prior tocomparison.

normalized number a floating-pointnumber in which the most significant digit ofthe significand (mantissa) is non-zero. In theIEEE 754 floating-point standard, the nor-malized significands are larger than or equalto 1, and smaller than 2.

normally closed contact contact of a con-tactor that is closed when the coil of the con-tactor is deenergized and opened when thecoil is energized.

normally closed, time to close a relay thatis closed when the power to its actuator is off,but has a time delay to close when power isremoved from the actuator. When power isapplied, the relay immediately opens.

normally open contact a contact that isopen under normal operating conditions andcloses when an action is initiated in its con-troller. For a contact that is part of a relay,the contact remains open when the relay isdeenergized and closes when the relay is en-ergized.

normally open, time to close a time de-lay relay that is open when the power to itsactuator is off. When power is applied tothe actuator, the relay remains open for anadjustable time delay, after which it closes.When power is removed from the actuator,the relay opens immediately.

Northeast blackout a 1965 failure ofmuch of the power grid in the northeasternstates of the USA.

Norton theorem states that the voltageacross an element that is connected to two ter-minals of a linear, bilateral network is equalto the short-circuit current between these ter-minals in the absence of the element, dividedby the admittance of the network lookingback from the terminals into the network,with all generators replaced by their internaladmittances.

NOT a Boolean operation that returns the1’s complement of the data to which it is ap-plied.

notch a disturbance of the normal voltagewaveform of duration less than 0.5 cycles,is of a polarity that is opposite to the wave-form and is hence subtracted from the normalwaveform with respect to the peak value ofthe disturbance voltage.

Noyce, Robert N. (1927–1990) Born:Denmark, Iowa, U.S.A.

Noyce is best known as one of thefounders and co-chair of Intel Corpora-tion, and a developer of the first commer-cially successful microchip, along with Gor-don Moore. Noyce began his career withWilliam Shockley’s short-lived company. He


and other engineers split with Shockley andformed Fairchild Semiconductor. Fairchildwas the first of the very successful “SiliconValley” high-tech firms. Noyce’s patent onthe microchip was challenged by Jack Kilbyand Texas Instruments. A court ruled in favorof Noyce, although it is commonly acceptedthat the original ideas were Kilby’s.

NPO Seenegative-positive-zero.

NPR Seenoise power ratio.

NRC the Nuclear Regulatory Commis-sion, a US regulatory body which overseescivilian power reactors.

NRD waveguide See nonradiatingdielectric waveguide.

NRE Seenonrecurring engineering.

NTSC SeeNational Television SystemCommittee.

NuBus an open bus specification devel-oped at MIT and used by several companies.It is a general-purpose backplane bus, de-signed for interconnecting processors, mem-ory, and I/O devices.

nuclear fuel fissile material, includingnatural uranium, enriched uranium, and someplutonium prepared for use in a nuclear reac-tor.

nuclear fuel management the process ofmanaging the degree of enrichment, the tim-ing of insertion into the core, and the place-ment and possible relocation of fuel withinthe core during the lifetime of the fuel.

nuclear magnetic resonance the phe-nomenon in which the resonant frequency ofnuclear spin is proportional to the frequencyof an applied magnetic field.Seemagneticresonance imaging.

nuclear power plant a thermal electricpower plant in which the heat for steam tur-bines is produced by nuclear fission.

nuclear reaction a reaction which causeschanges in the nucleus of an atom, thuschanging elements to another element or iso-tope, usually with the release of energy.

nuclear reactor (1) an apparatus designedto facilitate, contain, and control a nuclearchain reaction.

(2) any heat-producing array of fissile ra-dioactive materials constructed so as to pro-duce a controlled chain-reaction.

null a point on the radiation pattern thatcorresponds to zero or minimum values.

null controllability a dynamical systemthat is controllable to zero end state. Inmany cases, null controllability is equiva-lent to controllability. This is always truefor linear finite-dimensional continuous-timedynamical system. However, in discretecase, controllability may be the stronger no-tion than null controllability. In this case,the two concepts are equivalent if and onlyif rank(A) = n. For dynamical systemswith delays, these two notions are essentiallydifferent. For infinite-dimensional systemsthe relations between null controllability andcontrollability depend on the properties of thesemigroupS(t) generated by the operatorA.

null space the set of vectors that are or-thogonal to every vector of a set forming vec-tor space. For an(n, k) code, the dimensionof the null space isn− k.

null space of the Jacobian the subspaceN(J ) in Rn of joint generalized velocitiesthat do not produce any end-effector velocity,in the given manipulator posture. Nonzerosubspace for a redundant manipulator allowsto handle redundant degrees of freedom. Theexistence of nonzero subspace allows for re-dundant manipulator to generate internal mo-tions of the manipulator structure that do not


change the end-effector position and orienta-tion. As a consequence allows manipulatorreconfiguration into more dexterous posturesfor execution of a given task.

null steering a technique used to forman antenna’s radiation pattern in such a waythat there is no radiation in the direction ofan interfering signal. This technique hasbeen used with great success, for example,where CDMA systems have been overlaidwith fixed microwave communication sys-tems.

null-to-null bandwidth the width of themain lobe of a signal or system transfer func-tion in the frequency domain.

nullator an idealized two-terminal net-work element that conducts no current andyet maintains zero volts across itself. Thiselement is often used to represent a virtualconnection.

number system the representation ofnumbers as a sequence of digits with an inter-pretation rule which assigns a value to eachsequence. The conventional number systemsare fixed-radix and positional systems, wherethe digit in positioni has a weight ofri , wherer is the radix.

numerical aperture (NA) a parameter de-scribing the light-gathering capacity of an op-tical fiber. It is defined as the sine of half themaximum angle of light acceptance into thefiber

NA = √nco − ncl

wherenco is the refractive index of the core,andncl is the refractive index of the cladding.

Also used as a measure of the maximumangle of the cone of light entering or emerg-ing from an optical fiber.

numerical methods methods useful forobtaining quantitative solutions of electro-magnetic and microwave problems that havebeen expressed mathematically, including

the study of the errors and bounds on errorsin obtaining such solutions.

NUREG a contraction for Nuclear Regu-lations, which are published by the NRC.

NVRAM Seenonvolatile random-accessmemory.

nybble Seenibble.

Nyquist A/D converter A/D converterthat samples analog signals that have a max-imum frequency that is less than the Nyquistfrequency. The Nyquist frequency is definedas one-half of the sampling frequency. If asignal has frequencies above the Nyquist fre-quency, a distortion called aliasing occurs.To prevent aliasing, an antialiasing filter witha flat passband and very sharp roll-off is re-quired.

Nyquist band for pulse (or quadrature)amplitude modulation, the frequency bandwhich is the support of the Nyquist pulseshape. For a symbol rate of 1/T , the Nyquistband is the interval−1/(2T ) to 1/(2T ).

Nyquist criterion in analog-to-digitalconversion, the stipulation that the bandwidthof the sampling frequency must be greaterthan twice the bandwidth of the frequency ofthe signal being sampled.

Nyquist frequency SeeNyquist rate.

Nyquist I criterion the Nyquist I criterionfor zero inter-symbol interference states thatif h(t) is the signaling pulse and the symbolperiod isT , then we must haveh(kT ) = 1for k = 0 and 0 fork 6= 0, wherek is aninteger.

Nyquist noise Seethermal noise.

Nyquist plot a parametric frequency re-sponse plot with the real part of the transferfunction on the abscissa and the imaginarypart of the transfer function on the ordinate.


Nyquist pulse for pulse (or quadrature)amplitude modulation, the minimum band-width pulse which satisfies the Nyquist I cri-terion. Given a symbol rate 1/T , the Nyquistpulse isT −1 sin(πt/T )/(πt/T ), which hastwo-sided bandwidth 1/T .

Nyquist rate the lowest rate at which re-covery of an original signal from its sampledsignal is possible. If the highest frequencyof the analog signal isfH , the Nyquist fre-quency is the frequency of the samplesfS forproper recovery of the signal at the receivingend.

f Nyquist = fS should be> 2fH

In digital transmission systems, the analogsignal is transformed to a digital signal usingan A/D converter.

Nyquist sampling theorem fundamen-tal signal processing theorem that states thatin order to unambiguously preserve the in-formation in a continuous-time signal whensampled, the sampling frequency,fs = 1/T ,must be at least twice the highest frequencypresent in the signal.See alsoaliasing.


Oobject in object-orientation, an instanceof a class definition.

object based coding video compres-sion based on the extraction, recognition,and parameterization of objects in the scene.Unknown-object coding finds geometricalstructures representing 2D or 3D surfacesand codes these areas and their movementefficiently. Known-object coding relies onthe detection of objects for which the systemhas ahigh-level structural model. Seemodelbased image coding.

object code a file comprising an interme-diate description of a program segment.

object detection the detection of objectswithin an image. This type of process iscognate to pattern recognition. Typically,it is used to locate products ready for in-spection or to locate faults during inspection.An object usually denotes a larger character-istic in an image, such as an alignment ofpoints, a square, a disk, or a convex portionof a region, etc.; its detection involves non-local image transforms, such as theHoughtransformor the medial axis transform.Seeedge detection, Hough transform, key pointdetection, medial axis transform. See alsofeature detection.

object measurement the measurement ofobjects, with the aim of recognition or in-spection to determine whether products arewithin acceptable tolerances.

object orientation measurement of theorientation of objects, either as part of therecognition process or as part of an inspectionor image measurement process.

object recognition the process of locatingobjects and determining what types of objectsthey are, either directly or indirectly throughthe location of sub-features followed by suit-able inference procedures. Typically, infer-ence is carried out by application of Houghtransforms or association graphs.

object type the type of an object deter-mines the set of allowable operations that canbe performed on the object. This informationcan be encoded in a “tag” associated with theobject, can be found along an access pathreaching to the object, or can be determinedby the compiler that inserts “correct” instruc-tions to manipulate the object in a mannerconsistent with its type.

object-oriented analysis a method ofanalysis that examines requirements from theperspective of the classes and objects foundin the problem domain.

object-oriented design a design method-ology viewing a system as a collection of ob-jects with messages passed from object to ob-ject.

object-oriented methodology an appli-cation development methodology that usesa top-down approach based on a decompo-sition of a system in a collection of objectscommunicating via messages.

object-oriented programming object-oriented programming or task-level program-ming allows the user to command desiredsubgoals of the task directly, rather that tospecify the details of every action the robotis to take. In another words, task-level pro-gramming describes the assembly task as asequence of positional goals of the objectsrather than the motion of the robot neededto achieve these goals, and hence no explicitrobot motion is specified. Task-oriented pro-gramming is the most advanced program-ming system for robots. Not many industrialrobots are equipped with such a system.


objective function when optimizing astructure toward certain result, i.e., duringoptimization routines, the objective functionis a measure of the performance which shouldbe maximized or minimized (to be extrem-ized). See alsooptimization of microwavenetworks.

objective lens a well-corrected lens ofhigh numerical aperture, similar to a micro-scope objective, used to focus the beam oflight onto the surface of the storage medium.The objective also collects and recollimatesthe light reflected from the medium.

observability (1) the property of a systemthat ensures the ability to determine the initialstate vector by observing system outputs fora finite time interval. For linear systems, analgebraic criterion that involves system andoutput matrices can be used to test this prop-erty. See alsoobservability conditions.

(2) the ability to determine the signal valueat any node in a circuit by applying appropri-ate input states to the circuit and observingits outputs.

observability conditions a linear finite-dimensional continuous and stationary dy-namical system is observable if and only if

rank

[CT

∣∣∣AT CT ∣∣∣ (AT )2CT

|. . .|(AT)n−1

CT]

= n

It should be pointed out that observability ofstationary dynamical systems does not de-pend on the length of time interval.

observer an algorithm to estimate the statevariables from the input and output variablesof the system.

OCC Seeone-cycle control.

occlusion (1) the hiding or partial hidingof an object by another object or objects ina scene. The pattern of occlusion depends

on the viewpoint, as well as on the relativepositions of the objects.

(2) is the hiding of one object by another.In images of scenes, occluded objects are in-visible and irrelevant unless their presencecan be inferred, e.g., by motion analysis: par-tially visible objects are often said to be oc-cluded, though strictly speaking they are par-tially occluded. Occlusion and disocclusionof objects in the 3-D space add difficultiesto motion analysis and structure estimationrelated problems. For instance, when occlu-sion takes place, some motion informationwill be lost, while as disocclusion takes place,some areas that are originally occluded willbecome newly visible. Often no previous in-formation will be available regarding theseareas, thus making analysis more difficult.

occupancy of energy levels number ofatoms (or molecules, holes, electrons, etc.)per unit volume occupying an energy state.

OCR Seeoptical character recognition.

octal number system a number systemconsists of eight digits from 0 to 7; it is alsoreferred to as base 8 system.

octave a frequency ratio of two.

octave filter bank a filter bank with oc-tave spaced subbands. That is any given sub-band has twice the bandwidth of the adjacentlower frequency subband, except for the sub-band adjacent to the lowpass subband. Thediscrete wavelet transform is an octave-bandfilter bank.

octtree the 3-dimensional generalizationof thequadtree, it gives a representation of avolumetric image as a tree, where each parentnode has 8 children nodes and each leaf nodehas a label corresponding to the color of thecorresponding region. It is built recursivelyas follows. The root node corresponds to thewhole image; if all voxels are of the samecolor, then this node receives that color’s la-bel, and is a leaf; otherwise, we subdivide the


image into 8 sub-images whose dimensionsare half of those of the whole image, and eachsmaller image corresponds to a node which ischild of the root. Each of the 8 children nodesis the root of a sub-tree which is the octtreeof the corresponding sub-image; thus: if thesub-image has all its voxels of the same color,this node is a leaf to which the color’s labelis attached, otherwise we subdivide this sub-image into 8 smaller sub-images, to each ofwhich corresponds a child node of the actualnode. This recursive subdivision ends whenone reaches a sub-image of a single color,giving a leaf node with the corresponding la-bel. Seequadtree.

odd function a real-valued functionx(t)in which x(−t) = −x(−t) for all values oft . Compare witheven function.

odd order response a circuit gain or inser-tion loss versus frequency response in whichthere are an odd number of peaks in the rip-ple pattern, due to an odd number of pairedelements in the circuit. Odd order circuitsexhibit a peak for each element pair and apeak equal toLAmin (minimum attenuationloss across the band) at DC (low pass) or atwo (band pass).

odd signal Seeodd function. See alsoeven function.

odd-mode characteristic impedancecharacteristic impedance of a circuit due to

an odd-mode current or voltage excitation.Often applied in the context of a transmis-sion line coupler where the odd-mode exci-tation consists of applying equal amplitudebut opposite polarity voltages or currents ontwo conductors. The resulting impedanceunder this excitation is defined as the odd-mode characteristic impedance.

ODP Seeopen drip-proof.

ODP code Seeoptimum distance profilecode.

OEIC See optoelectronic integratedcircuit.

Oersted, Hans Christian (1777–1851)Born: Rudkobing, Langeland, Denmark

Oersted is best known as the discovererof electromagnetism. Oersted was a strongteacher and did much to bring Danish sci-ence up to world-class standards. Oerstedpredicted the magnetic effect of electric cur-rent in 1813, but was unable to prove it until1820. The publication of his results spurredthe work of Faraday and Ampere. Oerstadwent on to make other contributions in othersciences. He did not, however, return to hisstudy of electricity.

OFD code Seeoptimum free distancecode.

OFDM Seeorthogonal frequency divisionmultiplex.

off-axis illumination illumination thathas no on-axis component, i.e., that has nolight which is normally incident on the mask.Examples of off-axis illumination include an-nular and quadrupole illumination.

off-line error detection techniques of de-tecting faults by device testing (e.g., with theuse of BIST) such that the device is allowedto perform useful work while under test.

off-line testing testing process carried outwhile the tested circuit is not in use.

offset a sustained derivation or error dueto an inherent characteristic of positioningcontroller action. The difference exists at anytime between the set point and the value ofthe controlled variable.

offset short a short circuit impedancestandard with a phase offset from the refer-ence plane used in the process of calibratingvector network analyzers.


offset voltage for an ideal differential pairor op-amp, zero output corresponds to zerodifferential input. In reality, some nonzeroinput called the “offset voltage” is requiredfor the output to be zero. This means that inthe presence of an offset voltage error, a zeroinput will produce a nonzero output. Offsetvoltage is caused primarily by mismatch inthe two transistors of a differential pair.Seealsocommon centroid.

Ohm’s Law a fundamental law whichstates that the voltage across a resistance isdirectly proportional to the current flowingthrough it. The constant of proportionality isknown as the resistance.

This concept can be generalized to includethe relationship between the voltage and cur-rent in all situations, including alternatingvoltages and currents. In this case, all thequantities are measured as complex numbers,known as phasors, that are functions of fre-quency. This broadens the basic definitionof resistance, which is a real number mea-sured in ohms, to that of impedance, whichis a complex number with magnitude mea-sured in ohms and phase angle in degrees.The real part of the complex number repre-senting impedance is the resistance while theimaginary part is the reactance. Ohm’s Lawis a central concept to most electrical engi-neering theories.

Ohm, Georg Simon (1789–1854) Born:Erlangen, Germany

Ohm is best known for his discovery ofwhat we now call Ohm’s Law. Ohm held avariety of teaching posts at secondary schoolsas well as universities. In 1827 he publishedhis greatest work,Die Galvanische Kette.Along with Andre Ampere, Ohm was the firstto publish rigorously mathematical and the-oretical work on electricity. Ohm’s famouslaw states that current in a resistor is pro-portional to the applied voltage and inverselyproportional to the resistance. Ohm’s workwas initially scorned because it lacked theexperimental evidence. Worldwide acclaimchanged Ohm’s fortunes several years later.

He is honored by having his name used asthe unit of resistance, the ohm, and the unitof conductivity, the mho.

ohmic contact a heavily doped and/or lowbarrier height metal to semiconductor inter-face or contact that has a very low resistancerelative to the remainder of the device, suchthat the device performance is not signifi-cantly degraded. At lower doping levels, theohmic contact is described by Ohm’s Law,while at higher doping levels, tunneling dom-inates.

ohmic loss a term used to describe thepower dissipated due to the finite conductiv-ity of the metallic structure of an antenna,waveguide, transmission line, etc.

ohmic medium a medium in which con-ductivity is independent of the applied field.

ohmic region the voltage-controlled re-sistance region of operation of a transistor.Also referred to as the triode region, non-saturated region, and pinch-off region. Thisregion is in effect up to the point that the chan-nel of the transistor is completely depleted ofcharge carriers.

oil circuit breaker a power circuit breakerthat uses oil as an insulating and arc-clearingmedium.

oil-filled transformer a transformer inwhich the magnetic core and the windings aresubmerged in an insulating oil. In addition toserving as an insulator, the oil provides a heatexchange medium to cool the transformer.

oil-paper insulation an insulation schemeused in transformers and cables in which con-ductors are insulated with heavy paper im-pregnated with a dielectric oil.

OLAP See optical linear algebraicprocessor.

OLE Seeoptical logical etalon.


OMCVD See metal-organic chemicalvapor deposition.

omnidirectional antenna an antenna thatradiates power equally in all directions inat least one plane through its radiation cen-ter, but whose radiation pattern may vary forother such planes.

on-line error detection a real-time detec-tion capability that is performed concurrentlywith useful work (e.g., parity checkers, com-parators in duplicated systems).

on-line memory memory that is attachedto a computer system.

on-line optimization interval time inter-val over which optimization of the decisionsrequired to be made by the considered deci-sion mechanism at a particular time instantduring the control system operation is per-formed; setting of the on-line optimizationinterval is one of the essential decisions dur-ing design of a controller with the model-based predictive control mechanism.

on-line testing concurrent testing to de-tect errors while circuit is in operation.

on–off keying (OOK) a binary form ofamplitude modulation in which one of thestates of the modulated wave is the absenceof energy in the keying interval (the “off”state). The “on” state is represented by thepresence of energy in the modulated wave.

on-wafer measurements electrical mea-surements made using a wafer probe stationby directly contacting the device under testusing special test probes.

one decibel desensitization point a re-duction in a device output signal power byone decibel due to one or more additionalsignals that compress the device output.

one’s complement (1) a representation ofinteger numbers in which a data word is orga-

nized such that negative numbers all containa binary “one” in the leftmost bit while posi-tive numbers contain a “zero” in the leftmostbit, and in which the negative numbers are thebit-by-bit inverse of their positive equivalent.

(2) the operation of inverting a data wordso that all ones become zeros and vice versa.

one-cycle control (OCC) the switchedsignal is sensed to control the pulse timingsuch that the error between the control ref-erence and the average value of the switchedsignal is zero in each cycle. In a constant fre-quency implementation, the switch is turnedon by a constant frequency clock. The outputof the switch is integrated and compared toa reference; when the reference is reached,the switch is turned off and the integrator isreset. One-cycle control achieves zero errorwithin each cycle under steady-state or tran-sient conditions. In addition, this method ef-fectively rejects input perturbations, correctsswitching errors, and provides high linearity.

one-dimensional coding scheme a schemein which a run of consecutive pixels of thesame gray scale is combined together andrepresented by a single code word for trans-mission.

Scan lines are coded as “white” or “black,”which alternate along the line. The scans areassumed to begin with white and are paddedwith white if this is not the case. Generally,run-length coding is a one-dimensional cod-ing scheme.

one-dimensional correlator a correlatorwhere both input signals are one dimensional,such as temporal signals.

one-dimensional space integrating corre-lator a correlator where a single laggedproduct of two input signals is performed ata given instant and integrated spatially, suchas with a lens onto a single photodetector. Aseries of such operations are performed foreach lag value to give a temporal sequencethat is the correlation function.


one-gun color display a color CRT inwhich a single electron gun produces oneelectron beam that is controlled to producethe proper excitation of each of the three colorphosphors.

one-level memory an arrangement ofdifferent (in terms of speed, capacity, andmedium) types of memory such that the pro-grammer has a view of a single flat memoryspace.See alsohierarchical memory, virtualmemory.

one-level storage Seeone-level memory.

one-line diagram an abbreviated sche-matic representation of a power system inwhich three-phase transmission lines areshown as single lines between principal cir-cuit components and from which circuit pa-rameters are often omitted.

one-out-of-N coding a method of train-ing neural networks in which the input vec-tor and/or the output vector have only onenonzero element (usually equal to unity) foreach training example.

one-port device an electrical network inwhich only one external terminal is availablefor analysis. Antennas can be modeled asone-port devices.

one-port oscillator a device in a classof circuits in which the amplifying deviceand its nonlinearity are lumped into a sin-gle, reactance-free, controlled source. Theoscillations in this approach are most suc-cessfully found using the harmonic balancemethod (especially when the nonlinearity canbe approximated by a polynomial) or by thephase plane methods (for second-order sys-tems).

one-shot multivibrator a circuit thatis obtained from a closed-loop regenerativebistable system including two similar am-plifiers connected by coupling circuits. Inone-shot multivibrator, one of the circuits is

purely resistive, another, which is frequentlycalled “toggling” circuit, includes a reactance(capacitor or coil). The circuit is normallyin its stable state; by an external pulse it istransferred in its quasi-stable state, and thetoggling circuit helps to preserve the quasi-stable condition for the “timing-out” time.Then the circuit returns to its normal state.

one-step-ahead control the controlmethod to drive the value of thed-step-aheadoutput, whered is the inherent time delay ofthe system, to its desired level in one step.

ontogenic network a network that adaptsits topology during training through the ad-dition or deletion, as appropriate, of connec-tions and neurons, until the problem of inter-est is satisfactorily accommodated. Learningcan be either supervised or unsupervised.

OOK Seeon–off keying.

op-amp Seeoperational amplifier.

OPC Seeoptical proximity correction.

opcode a part of an assembly languageinstruction that represents an operation to beperformed by the processor. Opcode wasformed from the contraction of “operational”and “code.”

open circuit impedance the impedanceinto an N-port device when the remainingports are terminated in open circuits.

open drip-proof (ODP) pertaining to aventilated machine whose openings are con-structed to prevent drops of liquid or solidparticles falling on the machines at an angleless than 15◦ from the vertical from enteringthe machine either directly or by rolling alonga horizontal or inwardly inclined surface ofthe machine.

open kinematic chain a chain that con-sists of one sequence of links connecting twoends of the chain.


open loop gain Seeopen-loop gain.

open system architecture a layered archi-tectural design that allows subsystems and/orcomponents to be readily replaced or modi-fied; it is achieved by adherence to standard-ized interfaces between layers.

open systems interconnection (OST) modela framework for organizing networking tech-nology developed by the International Stan-dards Organization (ISO). Generally calledthe OSI model.

open waveguide a type of waveguideswhose cross section is not bounded by per-fect electric conductor, i.e., a waveguide forwhich the boundary value problem is on aninfinite domain. A few examples are opti-cal fiber, microstrips, coplanar waveguides,dielectric waveguides.

open-circuit test a transformer test con-ducted by applying nominal voltage on thelow voltage side while keeping the high volt-age side open. By measuring the power in,current, and voltage, the magnetizing reac-tance of the transformer equivalent circuitcan be determined.

open-circuit quarter-wave transmissionline transmission line of ninety degreeselectrical length where one end of the lineis terminated in an open-circuit impedance.The properties of the transmission line re-sult in the non-open circuited end to exhibita short-circuit impedance.

open-delta transformer a connectionsimilar to a delta–delta connection, exceptthat one single-phase transformer is removed.It is used to deliver three-phase power usingonly two single-phase transformers. The nor-mal capacity of the open-delta transformer isreduced to 57.7% of its delta rating.

open-loop control system a control sys-tem in which the system outputs are con-

trolled by the system inputs only, and no ac-count is taken of the actual system output.

open-loop gain the gain of an operationalamplifier with no feedback applied (with thenegative feedback loop “open”).

open-loop-feedback control predictivecontrol policy with repetitively used decisionmechanism, that at a given time instant, con-siders current state of the controlled process— or the current estimate of this state — andcomputes the values of the control inputs forthe next intervention instant or interval byperforming an open-loop optimization of theprocess operation over specified prediction(optimization) interval. The open-loop op-timization can be defined as an open-loopstochastic optimization problem or as a de-terministic optimization problem using a sin-gle forecast of the free inputs over specifiedprediction interval; other forms of open-loopoptimization problems can also be defined.Model-based-predictive control, widely usedas an industrial standard, is usually realizedin the form of the open-loop-feedback con-trol.

opening for structuring elementB, thecomposition of the erosion byB followed bythe dilation byB; it transformsX into X ◦B = (XB)⊕B. The opening byB is whatone calls an algebraic opening; this meansthat: (a) it is a morphological filter;(b) itis anti-extensive, in other words it can onlydecrease an object.Seedilation, erosion,morphological filter, structuring element.

operand specification of a storage loca-tion that provides data to or receives datafrom the operation.

operand address the location of an el-ement of data that will be processed by thecomputer.

operand address register the internalCPU register that points to the memory loca-


tion that contains the data element that willbe processed by the computer.

operating point Seeset point.

operating system a set of programs thatmanages the operations of a computer. Itoversees the interaction between the hard-ware and the software and provides a set ofservices to system users.

operating temperature the ambient casetemperature to which a device or circuit is ex-posed during operation under all supply biasand RF signal conditions, at which it mustmeet all specified requirements unless other-wise called out. The operating temperaturerange is the minimum to maximum operatingtemperatures of the device.

operation specification of one or a setof computations on the specified sourceoperands placing the results in the specifieddestination operands.

operational amplifier a high-gain DC-coupled amplifier with a differential inputand single-ended output. In nearly all am-plifier applications, the op-amp is used withnegative feedback (“closed-loop”), so thatthe closed-loop gain of the amplifier dependsprimarily on the feedback network compo-nents, and not on the op-amp itself. It iswidely used as a basic building block in elec-tronic designs. Abbreviated as “op-amp.”

operational control control or decisionmaking activity that requires, or in some in-stances may require, a human operator (dis-patcher) to approve — or modify — auto-matically computed decisions before they areactually implemented; such a situation is typ-ical in production management, in control ofsupply and environmental systems, at upperlayers of process control systems.

operational impedance a representationin which the impedance of a system is ex-pressed as a function of the Heaviside op-

eratorp = d/dt or the Laplace operators = jω. In the modeling of synchronousmachines, the Park’s transformed stator fluxlinkages per second are often expressed interms of impedancesXq(p), and Xd(p),termed the quadrature- and direct-axis opera-tional impedances, respectively. Using these,the dynamics of the rotor windings are rep-resented within the operational impedances,and therefore the rotor of a synchronous ma-chine can be considered as either a distributedor lumped parameter system.

operational space Seeexternal space.

operational space control SeeCartesian-based control.

operator SeeCanny operator, Laplacianoperator, Marr-Hildreth operator, morpho-logical operator, Sobel operator.

opposition effect Seeweak localizationof light.

optical adder an optical device capableof performing the function of arithmetic ad-dition using binary signals. It can be con-structed using a series of cascaded full adderswhere carry has to ripple through each fulladder from least significant bit to most sig-nificant bit. It can also be constructed usingcascaded layers. Each layer consists of a se-ries of half adders. Carry also has to ripplethrough the cascaded layers. It is also knownas digital adder.

optical addition adding operation usinglight. Two incoherent light beams have in-tensitiesA andB, respectively. When twobeams are combined, i.e., illuminating thesame area, the resultant intensity isA+ B.

optical amplifier amplifier of electromag-netic waves at optical frequencies, usually bythe process of simulated emission or nonlin-ear optics.Seelaser amplifier.


optical beam beam of electromagneticpower at optical frequencies.

optical bistability the property of certainnonlinear optical system to possess two pos-sible output states for a given input state. Inone typical example, the bistable optical de-vice has the form of a third-order nonlinearoptical material placed inside of an opticalcavity, and the device can display two pos-sible different transmitted intensities for cer-tain values of the input intensity.

optical bus an optical channel used fortransmitting a signal from a source to one ormore detectors. A bus allows only the sameinterchange of information to take place atdifferent detectors. A source is connected tomany detectors.

optical cavity resonant structure for main-taining oscillation modes at optical frequen-cies.

optical character recognition (OCR) aprocess in which optically scanned charactersare recognized automatically by machine. Itis widely utilized in document storage, pro-cessing and management.

optical circulator a four-ports optical de-vice that can be used to monitor or sampleincident light (input port) as well as reflectedlight (output port) with the two other unidi-rectional coupling ports.

optical communications communicationof information at optical carrier frequencies.There are two general categories:

1. free-space, such as with lasers and op-tical modulators, and

2. guided-wave, such as over optical fiberand using guided lightwave devices.

optical computer (1) a general purposedigital computer that uses photons as an elec-tronic computer uses electrons. The technol-ogy is still immature. The nonlinear opera-tions that must be performed by a processor

in computer are difficult to implement opti-cally.

(2) an analog optical processor capable ofperforming computations such as correlation,image subtraction, edge enhancement, andmatrix-vector multiplication that are usuallyperformed by an electronic digital computer.Such a computer is usually very specific andinflexible, but well adapted to its tasks.

optical computing the use of optics toaid in any type of mathematical computation.Categories of optical computing include.

1.analog processing such as using opticalFourier transforms for spectral analysis andcorrelation,

2. optical switching and interconnectionusing nonlinear, e.g., bistable, optical de-vices, and

3. use of optical devices in digital com-puters.

optical demultiplexer a device that di-rects an input optical signal to an output portdepending on its wavelength.

optical disk a disk on which data arestored optically. Data are written by alter-ing the reflectivity of the surface, and readby measuring the surface reflection of a lightsource. Storage is organized in the same wayas on a magnetic disk, but higher storage den-sity can be achieved.

optical disk track the region on a com-pact disk in which pits or other features are lo-cated to store digital data. CD-ROM pit sizesare approximately 0.5 microns in width andfrom 0.8 to 3.6 microns in length, dependingon the number of ones and zeros representedby the length. The track-to-track spacing is1.6 microns.

optical energy energy of an electromag-netic wave oscillating at optical frequencies.See alsoenergy, electromagnet.

optical expert system an expert systemthat utilizes optical devices for performing


logic operations. An expert system mostlyrequires logic gates to perform inferences.In an optical expert system, not only discreteoptical logic gates are used, but the paral-lelism and capability of performing matrix-vector multiplication of optical processor areexploited as well. Sequential reasoning in anelectronic expert system is replaced by paral-lel reasoning in an optical expert system, soits speed can be increased substantially. Aswith the optical computer, it is still immature.

optical fiber a single optical transmis-sion fiber usually comprised of a cylindricalcore (5–100 mm diameter) in which the lightis guided of higher index of refraction sur-rounded concentrically by a cladding (125–250 mm diameter) with a lower index of re-fraction. More properly defined as an opticalwaveguide. Some optical fibers may havemultiple concentric cores and/or claddings.Optical fibers made be of all glass, all plas-tic, or a combination of glass core and plasticcladding construction. Optical glass fibersmay be silica- or fluoride-based glass.

optical fiber signal distortion a change inthe temporal shape of an optical signal trans-mitted through an optical fiber caused by acombination of wavelength effects (disper-sion) and multimode and polarization effects.The wavelength effects include material dis-persion, profile dispersion, and waveguidedispersion. The multi-mode effects causedistortion by the differential time delays be-tween the various modes propagating in amultimode fiber. The polarization effectcauses distortion by the differential time de-lay between the two polarizations of a singlemode.

optical flow the 2-D field of apparent ve-locities of pixels on an image plane; i.e., theraw motion information arising from the dis-placement of points in the visual (optical)field. Let each image pointp have an inten-sity I and a velocityv = (vx, vy), which areboth functions ofp and timet , where the ve-locity represents the image plane projection

of the point’s 3-dimensional velocity. Underthe assumption that the point’s intensity doesnot change along its trajectory, it follows that

(∂I/∂x)vx + (∂I/∂y)vy + (∂I/∂z)vz

+∂I/∂t = ∇I · v + ∂I/∂t = 0.

See alsoaperture problem.

optical flux Seeoptical flow.

optical Fourier transform the imple-mentation of the Fourier transform usingthe transform properties of a lens, spatiallight modulators, such as acousto-optic mod-ulators, to generate the optical input in-formation, and photodetector arrays to de-tect the optical field representing the Fouriertransform. See alsoFourier transform,two-dimensional Fourier transform.

optical full adder an optical device form-ing part of an adder and able to receive threeinputs, augend, addend, and carry from theprevious stage, and deliver two outputs, sumand carry. Several logic gates are required toprovide the sum and the carry.

optical gain increase in the amplitude ofan optical signal with propagation distance inan optical amplifier.

optical guided-wave device an optical de-vice that transmits or modifies light while itis confined in a thin-film optical waveguide.

optical half adder an optical device form-ing part of an adder and able to receive twoinputs, augend and addend, and deliver twooutputs, sum and carry. The sum is the XORfunction of two inputs, and the carry is theAND function of two inputs. An optical halfadder consists of an optical XOR gate and anoptical AND gate. It is also known as one-digit adder.

optical inference engine See opticalexpert system.


optical integrated circuit guided waveoptics on a single substrate. If also incor-porating active and electrical devices, alsoknown as opto-electronic integrated circuit(OEIC).

optical interconnect an optical communi-cation system in an electronic computer thatconsists of three primary parts:

1. sources,

2. optical paths with switches or spatiallight modulators, and

3. detectors.

An optical signal from a source of an arrayof sources is transmitted to a detector of an ar-ray of detectors in optical interconnects. Thetransmitted optical signal from the source isconverted from an electronic signal. The de-tector in turn reverts the optical signal intoan electronic signal. The merits of opticalinterconnects include large bandwidth, highspeed of light propagation (the velocity ofelectric signals propagating in a wire dependson the capacitance per unit length), no in-terferences, high interconnection density andparallelism, and dynamic reconfiguration.

optical invariant a parameter that re-mains constant throughout an optical system.A consequence of the optical invariant is thatthe product of numerical aperture and mag-nification in the system is constant.

optical isolator a unidirectional opticaldevice that only permits the transmission oflight in the forward direction. Any reflectedlight from the output port is blocked by thedevice from returning to the input port withvery high extinction ratio.

optical Kerr effect SeeKerr effect.

optical linear algebraic processor (OLAP)an optical processor that performs specific

matrix algebraic operations as fundamentalbuilding blocks for optical computation andsignal processing.

optical lithography lithography methodthat uses light to print a pattern in a photo-sensitive material. Also called photolithog-raphy.

optical logic logic operations, usually bi-nary, that are performed optically.

optical logic gate an optical device forperforming Boolean logic operations. Thebasic idea is that since a computer is built byBoolean logic gates, if we can make opticallogic gates, then we can eventually build acomplete optical computer. Since light doesnot affect light, it cannot directly performnonlinear operations represented by sixteenBoolean logic gates. A thresholding device isrequired if intensities of two input beams aresimply added in the logic gate. For example,to perform AND operation, 00-0, 01-0, 10-0,11-1, the threshold is set at 2. Only when bothbeams have high intensities, the output is 1.Otherwise, the output is 0. To perform ORoperation, 00-0, 01-1, 10-1, 11-1, the thresh-old is set at 1. The gate can be constructedby directing two input light beams onto thesame detector. After passing through an elec-tronic thresholding device, the resulting elec-tric signal from the detector shows the logicoutput. To implement optical logic gate with-out using electronic circuit, optical threshold-ing devices such as MSLM or nonlinear opticmaterials can be used.

optical logical etalon (OLE) pulsed non-linear Fabry–Perot etalon that requires twowavelengths (λ1 = signal, λ2 = clock).

optical maser early equivalent name forlaser.

optical matrix–matrix multiplication anoperation performing matrix–matrix multi-plication using optical devices. For two-by-two matricesA and B, the matrix–matrixmultiplication produces two-by-two matrixC, whose elements are

lc11 = a11b11 + a12b21


c12 = a11b12 + a12b22

c21 = a21b11 + a22b21

c22 = a21b12 + a22b22

An optical implementation is as follows. Ma-trix elementsb12, b22, b11, andb21 are repre-sented by four vertical lines on a spatial lightmodulator. MatrixA is represented by a two-by-two source array. Two images of matrixA are generated by an optical means, for ex-ample, two lenslets. An image of A coverstwo lines ofb12 andb22, another image ofAcovers linesb11 andb21. Using two cylindri-cal lenses, light passing through the spatiallight modulator is focused into four pointsat the four corners of a square representingc11, c12, c21, and c22. Various optical ar-rangements for implementing matrix–matrixmultiplication have been proposed. Opticalmatrix–matrix multiplication is needed forsolving algebraic problems, which could befaster than an electronic computer because ofparallel processing.

optical matrix-vector multiplication anoperation performing matrix–vector multi-plication using optical devices. For matrixW and vectorX, the matrix–vector multi-plication produces vectorY , whose elementsare

y1 = w11x1 + w12x2 + w13x3

y2 = w21x1 + w22x2 + w23x3

y3 = w31x1 + w32x2 + w33x3

A simple coherent optical processor to im-plement this operation is as follows. Vectorelementsx1, x2, andx3 are represented by thetransmittance of three vertical lines on a firstspatial light modulator. A collimated coher-ent light beam passes through the first spatiallight modulator. The modulated light thenpasses through a second spatial light modu-lator displaying three-by-three squares withtransmittances ofw11 to w33. Two spatiallight modulators are aligned such that the ver-tical linex1 covers three squares ofw11,w21,w31, linex2 coversw12,w22,w32, and linex3coversw13, w23, w33. The light passing the

second spatial light modulator is focused by acylindrical lens to integrate light in horizontaldirection. In other words,w11x1,w12x2, andw12x3 are summed up at a point on the fo-cal line of cylindrical lens. A detector placedat this point will producey1. Two other de-tectors will providey2 and y3 in a similarway. Various optical arrangements are pos-sible to implement matrix–vector multiplica-tion, including correlators. Optical matrix–vector multiplication is important for cross-bar switch and neural networks. It can alsobe used for finding eigenvalues and eigenvec-tors, solving linear equations, and computingthe discrete Fourier transform.

optical modulator device or system thatmodulates the amplitude or phase of an opti-cal signal.

optical multiplication multiplying oper-ation using light. A light beam has intensity-A. When the beam passes through a trans-parency with transmittanceB, the intensityof transmitting light beam isAB.

optical neural network optical processorimplementing neural network models and al-gorithms. A neural network is an informationprocessing system that mimics the structureof the human brain. Two features of neu-ral networks are recognition capability andlearning capability. In recognizing process,the neural net is formulated as

zj = f

∑ji

xi + θj

wherezj is the output of thej th neuron,wjiis the interconnection weight between theithinput neuron and thej th neuron,xi is the in-put coming from theith input neuron,θj isthe bias in thej th neuron, andf is a non-linear transfer function. Notice thatzj andxi are binary. Nonlinear transfer functionfcould simply be a threshold function. Theformation of interconnection weight matrix iscalled learning process. It appears that in rec-ognizing process, the neural network has to


perform a matrix–vector multiplication and anonlinear operation. The matrix–vector mul-tiplication can be easily performed using anoptical system. However, the nonlinear oper-ation can not be performed by optical means.It may be performed using electronic cir-cuits. For two-dimensional neural process-ing, tensor–matrix multiplication is needed,which can also be realized by optical means.

optical parametric oscillator a nonlinearoptical device that can produce a frequency-tunable output when pumped by a fixed-frequency laser beam. The device consistsof a second-order nonlinear optical crystalplaced inside of an optical resonator as wellas additional components for precise controlof the output characteristics. When pumpedby a laser beam at the pump frequencyωp, itproduces two output frequencies, one at thesignal frequencyωs and one at the so-calledidler frequencyωi , whereωp = ωs + ωi .

optical path optical elements in the pathof the laser beam in an optical drive. The pathbegins at the laser itself and contains a colli-mating lens, beam shaping optics, beam split-ters, polarization-sensitive elements, pho-todetectors, and an objective lens.

optical path length the distance an in-cident photon travels within a material be-fore it emerges and impinges on a detector.The path may be directly through the mate-rial, in which case the pathlength will equalthe thickness of the light absorbing mate-rial, or may involve scattering due to hetero-geneous structures, causing the pathlengthto exceed the material’s thickness (effectivepathlength).

optical phase conjugation an optical pro-cess in which a time reversed replica of the in-cident wave is generated. The time-reversedreplica is identical to the incident wave, ex-cept the direction of propagation. Opticalphase conjugation can be achieved by usinga deformable mirror whose mirror surfacematches exactly the wavefront of the incident

wave. For example, consider the reflection ofa spherical wave from a spherical mirror withan identical radius of curvature. The reflectedwave is a time-reversed replica of the incidentwave. Optical phase conjugation can also beobtained by using optical four-wave mixing,stimulated Brillouin scattering (SBS) in non-linear media. In optical phase conjugationvia four-wave mixing, the nonlinear mediumis pumped by a pair of counterpropagatingbeams. When a signal beam is incident intothe medium, a phase conjugate beam is gen-erated which is a time-reversed replica of theincident beam.

optical potential transformer (OPT) apotential transformer that uses a voltage-sensitive optical device, typically in conjunc-tion with optical fibers, to avoid the need forthe heavy insulation required of electromag-netic or capacitive potential transformers.

optical processing See optical signalprocessing.

optical proximity correction (OPC) amethod of selectively changing the shapes ofpatterns on the mask in order to more ex-actly obtain the desired printed patterns onthe wafer.

optical proximity effect proximity effectthat occurs during optical lithography.

optical pumping excitation of an atom ormolecule resulting from absorption of opti-cal frequency electromagnetic radiation; theelectromagnetically assisted accumulation ofpopulation into or out of one or more statesof a quantum mechanical system. In prac-tice, this generally involves selective absorp-tion of the electromagnetic field to populatean excited state, followed by a less selectivedecay into more than one ground state. Forexample, a system having ground state spinsublevels can be optically pumped by circu-larly polarized light into a single ground statespin sublevel. In a multilevel system, moreselective transfer of population from one state


to another can be achieved by adiabatic pas-sage.

optical rectification the second-ordernonlinear optical process in which a materialdevelops a static electric field in response toand proportional to the square of the strengthof an applied optical field.

optical repeater optoelectric device thatreceives a signal and amplifies it and retrans-mits it. In digital systems, the signal is re-generated.

optical representation of binary numbersthe representation of binary numbers 0 and

1 using light. Since an optical detector issensitive to light intensity, it is a very logicalchoice to represent binary numbers 0 and 1with dark and bright states or low and highintensity, respectively. However, some diffi-culty would occur when 1 has to be the out-put from 0, because no energy can be gen-erated for 1 from 0 without pumping light.Binary numbers 0 and 1 can be representedby a coded pattern instead of the intensity ofa single spot. They can also be representedby two orthogonal polarizations, although thefinal states should be converted to intensity,which is the only parameter that can be de-tected by a detector.

optical resonator electromagnetic cavitydesigned to have low loss at optical frequen-cies.

optical signal processing the use of lightwith optical devices to process signals andimages. These systems exploit the highbandwidths and inherent parallelism offeredby optical components.

optical soliton a pulse that propagateswithout change in shape through a disper-sive nonlinear optical medium as a conse-quence of an exact balance between disper-sive and nonlinear effects. The propagationof such a pulse is described by the nonlinearSchrodinger equation.

optical switching a process in which oneoptical beam is controlled by another opticalwave or by an electro-optical signal.

optical tensor–matrix multiplication anoperation performing tensor–matrix multi-plication using optical devices. For tensorW and matrixX, the tensor–matrix multipli-cation produces matrixY , whose elementsare

y11 = w1111x11 + w1112x12 + w1211x21

+ w1212x22

y12 = w1121x11 + w1122x12

+ w1221x21 + w1222x22

y21 = w2111x11 + w2112x12 + w2211x21

+ w2212x22

y22 = w2121x11 + w2122x12

+ w2221x12 + w2222x22

An incoherent optical processor can imple-ment this operation as follows. The two-by-two matrix (

x11 x12x21 x22

)represented by four LEDs is imaged by alenslet onto a part of the tensor that is a two-by-two matrix(

w1111w1112w1211w1212

)represented by a spatial light modulator.Light passing through the spatial light mod-ulator is integrated by another lenslet toprovide y11. Thus, two sets of two-by-two lenslet arrays with four LEDs and aspatial light modulator displaying all ten-sor elements as four matrices are sufficientto perform this tensor–matrix multiplica-tion. Various optical arrangements can im-plement tensor–matrix multiplication. Opti-cal tensor–matrix multiplication is importantfor two-dimensional neural networks.

optical time domain reflectometry (OTDR)device used to locate faults or determine at-tenuation in a length of optical fiber. The


technique relies on launching a pulse of lightinto the fiber and measuring the backwardscattered power with time, which is then re-lated to distance along the fiber.

optical transfer function the normal-ized Fourier transform of the incoherent pointspread function, i.e.,

H(ωx, ωy) =∫∞−∞

∫∞−∞ h(x, y)exp[−i(ωxx + ωyy)]dxdy∫∞

−∞∫∞−∞ h(x, y)dxdy

where h(x, y) is a point spread functionwhich represents the image intensity re-sponse to a point source of light.Seealso modulation transfer function, impulseresponse, point spread function.

optical time-domain reflectometer opticalfiber an instrument capable of launchingoptical pulses into an optical fiber transmis-sion link or network, detecting the backscat-tered optical pulses, and displaying thesebackscattered pulses as a function of dis-tance. Used to measure optical fiber length,attenuation, connector and splice losses, andto detect faults in optical fiber networks.

optically inhomogeneous medium a me-dium whose refractive index randomly varieseither in space, in time, or in both. It producesscattering of the light transmitted or reflectedat it.

optimal decision the best decision, fromthe point of view of given objectives andavailable information, that could be takenby the considered decision (control) unit;the term optimal decision is also used in abroader sense — to denote the best deci-sion that can be worked out by the consid-ered decision mechanism, although this de-cision mechanism may itself be suboptimal.An example is model-based optimization asa decision mechanism at the upper layer of atwo-layer controller for the steady-state pro-cess — the results of this optimization willbe referred to as the optimal decisions.

optimal sensitivity minimal, in the senseof the H∞ norm, value of the sensitivityfunction found using H infinity design (H∞)techniques. For single-input–single-outputsystems, theH∞ norm is simply the peakmagnitude of the gain of the sensitivity func-tion. In the multi-input–multi-output case, itis expressed by the maximal singular value ofthe sensitivity matrix function. To meet thedesign objectives, the optimal sensitivity isfound for the sensitivity function multipliedby left and right weighting functions. It en-ables one to specify minimum bandwidth fre-quency, allowable tracking error at selectedfrequencies, the shape of sensitivity functionover selected frequency range, and maximumallowed peak magnitude.

optimization determining the valuesof the set of free parameters that min-imizes or maximizes an objective func-tion. The minimization or maximizationmay be subject to additional constraints.Seealsogradient descent, graph search, geneticalgorithm, minimax estimate, relaxationlabeling, simulated annealing.

optimization of microwave networksthe procedure used in the design of a mi-crowave system in order to maximize or min-imize some performance index. May entailthe selection of a component, of a particularstructure, or a technique.See alsoobjectivefunction.

optimizing control layer operation orstructure of the control layer of a multilayercontroller, usually situated above the directcontrol layer, where the objective of the de-cision mechanism is to minimize (or maxi-mize) a given performance function associ-ated with the control system operation; typ-ical example of optimizing control is the setpoint control of an industrial process con-cerned with optimal operation of this pro-cess in the steady state conditions.See alsosteady-state control.


optimum combining a technique forweighted addition of the output of severalcommunication channels where the weightsare chosen to maximize the ratio between thedesired signal and the sum of the interfer-ence and noise in combined output. Usedin adaptive antenna systems.See alsoanglediversity, antenna diversity.

optimum coupling choice of output cou-pling mirror reflectivity that yields the largestoutput power from a laser oscillator.

optimum distance profile (ODP) codea convolutional code with superior distanceprofile given the code rate, memory, and al-phabet size. Optimum distance profile codesare suitable if, e.g., sequential decoding isused.

optimum free distance (OFD) code aconvolutional code with the largest possiblefree distance given the code rate, memory,and alphabet size.

optode a fiber optic sensor used to deter-mine the concentration of a particular chem-ical species present in the sensor’s environ-ment by utilizing spectroscopic changes in asensing element placed at the end of the op-tical fiber.

optoelectronic integrated circuit (OEIC)opto-electronic device combining optical

and electrical devices on a common substrate.This includes combinations of semiconduc-tor lasers, modulators, photodetectors, andelectrical processing circuitry.

optoelectronics the interaction of lightand electrons in which information in an elec-trical signal is transferred to an optical beamor vice versa, e.g., as occurs in optical fibercommunications components.

optrode Seeoptode.

opus sweep an amplifier large signal sta-bility test in which a signal within the pass-

band is held constant in both frequency andpower such that compression is achieved. Asecond signal operated at a much smallerlevel is then injected and swept across theentire frequency band, observing the gain ofthis small signal through the compressed am-plifier. Sharp gain peaks in this small signalresponse are indications of large signal sta-bility problems. The test is dependent on biasconditions, large signal frequency, large sig-nal compression, and amplifier source andload impedance.

OR the Boolean operator that implementsthe disjunction of two predicates. The truetable for∨ ≡ X ORY is

X Y X ∨ YF F F

F T T

T F T

T T T

n-ary ORs can be obtained as disjunction ofbinary ORs.

OR gate a logic circuit that performs theOR operation. The output of the gate is highif one or all of its inputs are high.

oral cavity the human cavity wherespeech is produced and where the articula-tions of the speech organs makes it possibleto produce different sounds.

Orange book SeeIEEE Color Books.

ordinary refractive index the refractiveindex that is invariant with light direction,affecting one particular optical polarizationcomponent.

ordinary wave polarization component ofa light wave that is affected by the ordinaryrefractive index.

organic LED Seeorganic light emittingdiode.


organic light emitting diode a groupof recently developed organic material thatemits light in response to electrical input. Al-though lower in efficiency, they have greatermanufacturing flexibility than semiconduc-tor LED.

orientation for an orthonormal coordinateframe attached to an object, the direction ofthe three axes of the frame relative to the baseorthonormal coordinate frame.

orthogonal for two signals or functions,the condition that their inner product is zero.For example, two real continuous functionss1(t) ands2(t) are orthogonal if∫

s1(t)s2(t) dt = 0

orthogonal code division multiple accessscheme a code division multiple accessscheme (CDMA), i.e., a spread spectrum sys-tem with multiple users, where the PN se-quences utilized by the different users are or-thogonal sequences. In such a system, themultiple access interference is zero. The or-thogonal sequences can be obtained from therows of a Hadamard matrix, for example.

orthogonal CDMA Seeorthogonal codedivision multiple access scheme.

orthogonal filter bank a biorthogonal fil-ter bank whose polyphase transfer functionsof the analysis and synthesis filters satisfyF(z) = HT (z−1).

orthogonal frequency division multiplex(OFDM) a frequency division multiplexscheme where the subcarriers for each of thefrequency divided bands (subbands) are or-thononal to each other. This allows for spec-tral overlap between successive subbands.

orthogonal hopping sequences a set offrequency hopping sequences (hopping pat-terns) for which no two sequences have thesame value for a given position in the se-

quence. The set of sequences can be rep-resented as the rows of a matrix. In such acase for a given column of the matrix, theelements are distinct.

orthogonal matrix a matrixA whose in-verse isAT ; that is,ATA = I , where I isthe identity matrix. More generally,A is aunitary matrix if its inverse isAH (that is, thecomplex-conjugate transpose); consequentlya real unitary matrix is an orthogonal matrix.Although not recommended, the phrase “or-thogonal matrix” is sometimes used to referto a unitary matrix.

orthogonal projection for a vectorcontoa vectorb, the orthogonal projection is

cb =(

bT cbT b

)b

orthogonal set of functions for real signalsetfm(t) over[t1, t2] such that∫ t2

t1

fm(τ)fn(τ )dτ ={

0 m 6= n

1 m = n

orthogonal functions are necessary for trans-forms such as the FFT, or the DCT.

orthogonal transform a transform whosebasis functions are orthogonal. See orthog-onalset of functions.Thetransformmatrixofadiscrete orthogonal transform is an orthogo-nal matrix. Sometimes orthogonal transformis used to refer to a unitary transform. Or-thogonal real transforms exhibit the propertyof energy conservation.

orthogonal wavelet wavelet functionsthat form orthogonal basis by translation anddilation of a mother wavelet.

orthographic projection a form of pro-jection in which the rays forming an imageare modeled as moving along parallel pathson their way to the image plane: usually thepaths are taken to be orthogonal to the imageplane. Orthographic projection suppresses


information on depth in the scene. A limit-ing case of perspective projection.

orthonormal functions an orthogonalsignal setfm(t) on [t1, t2] such that∫ t2

t1

|fm(t)| dt = 1

for all m.

orthonormal transform Seeorthogonaltransform.

oscillation condition (1) condition thatmust be satisfied for an active circuit or de-vice to exhibit a stable frequency oscillation.For a stable oscillation to be achieved, a con-dition is placed on the impedance presentedto the active device, the impedance exhib-ited by the active device, and the derivativeof these impedances with respect to both fre-quency and voltage.

(2) requirement that the electromagneticfield in a laser oscillator be self-sustaining in-cluding all loss and gain elements; also sub-conditions on self-consistency of amplitude,phase, polarization, and frequency.

oscillation frequency frequency pro-duced by a laser oscillator, including effectsof the optical cavity, active medium, and anyother elements in the cavity.

oscillation threshold condition underwhich unsaturated gain is equal to loss.

oscillator (1) a circuit that generates arepetitive series of pulses at a certain fre-quency.

(2) an amplifier that has an output for zeroinput, i.e., it is providing an output signaleven though there is no signal applied at theinput.

oscillatory transient a rapid changein frequency (not power frequency) duringsteady state that is bipolar for voltage or cur-rent.

oscillograph a continuous recording ofthe waveforms of an electric power line, for-merly made with a cathode-ray tube but cur-rently with a digital signal recorder, kept up-dated to record abnormalities during switch-ing operations and fault conditions.

OSI model Seeopen systems interconnec-tion model.

OTDR Seeoptical time domain reflecto-metry.

OTDR optical fiber Seeoptical timedomain reflectometry, optical fiber.

OTP one-time programmable. Seeprogrammable read-only memory.

out-of-order issue the situation in whichinstructions are sent to be executed not nec-essarily in the order that they appear in theprogram. An instruction is issued as soon asany data dependencies with other instructionsare resolved.

out-of-step an abnormal condition whengenerators in a power system cannot operatein synchronism.

out-of-step relay a protective relay thatsenses that a synchronous generator haspulled out of step, and is operating at a fre-quency different than the system frequency.

outage (1) the percentage of time or areafor which a communication system does notprovide acceptable quality.

(2) loss of power from all or part of apower system.

outage inferencing the act of identifyingthe probable location of an outage based oninformation received from customer troublecalls and power monitoring units.

outlier a statistically unlikely event inwhich an observation is very far (by severalstandard deviations) removed from the mean.


Also refers to points which are far removedfrom fitted lines and curves. In experimentalcircumstances, “outlier” frequently refers toa corrupt or invalid datum.

output backoff Seebackoff.

output block when the processor writesa large amount of data to its output, it writesone block, a certain amount of data, at a time.That amount is referred to as an output block.

output buffer when the processor writesto its output device, it must make sure that de-vice will be able to accept the data. One com-mon technique is that the processor writes tocertain memory addresses, the output buffer,where the output device can access it.

output dependency the situation whentwo sequential instructions in a programwrite to the same location. To obtain the de-sired result, the second instruction must writeto the location after the first instruction. Alsoknown as write-after-write hazard.

output device a device that presents theresults sent to the user, and typical output de-vices are the screen and the printer. (Accord-ing to some interpretations, it also includesthe stable storage where the processor savesdata.)

output filter a lowpass filter used to at-tenuate the switching ripple at the output ofswitching circuits to a tolerable level.

output impedance the ratio of the drop involtage to the current drawn is known as theoutput impedance of the electric source andis measured in units of ohms.

output neuron layer a neuron (layer ofneurons) that produces the network output(outputs). In feedforward networks, the setof weights connected directly to the outputneurons is often also referred to as the outputlayer.

output power (1) the difference in thepower available under perfectly matchedconditions and the reflected power taking theoutput return loss into account, expressed inwatts.

(2) in lasers, the useful output from a laseroscillator.

output return loss negative ratio of thereflected power to the incident power at theoutput port of a device, referenced to theload or system impedance, expressed in deci-bels. The negative sign results from the term“loss.” Thus an output return loss of 20 dBresults when 1/100th of the incident power isreflected.

output routine low-level software thathandles communication with output devices.Handles the formatting of data as well aseventual protocols and timings with the out-put devices.

output swing for a semiconductor device,the difference between the output high volt-age and the output low voltage.

output vector a vector formed by the out-put variables of a network.

outstar configuration consists of neu-rons driving a set of outputs through synap-tic weights. An outstar neuron produces adesired excitation pattern to other neuronswhenever it fires.See alsooutstar training,instar configuration.

outstar rule a learning rule that incorpo-rates both Hebbian learning and weight de-cay. A weight is strengthened if its input sig-nal and the activation of the neuron receivingthe signal are both strong. If not, its valuedecays. A similar rule exists for instars, buthas found less application.

outstar training neuron training wherethe weights are updated according to

wi(t + 1) = wi(t)+ ν (yi − wi(t))


where ν is the training rate starting fromabout 1 and gradually reduced to 0 duringthe training.See alsoinstar configuration.

over-compounded a compound DC gen-erator in which the terminal voltage increasesas load current increases. Extra turns areadded in the series winding to generate theadditional voltage after compensating for thearmature voltage drop and the armature re-sistance drop.

overcurrent (1) current in a circuit thatexceeds a preset limit.

(2) motor current magnitude in the normalcircuit path exceeding the full-load current.

overcurrent protection (1) the act of pro-tecting electrical and electronic devices orcircuits from a dangerous amount of inputcurrent.

(2) the effect of a device operative on ex-cessive current.

overcurrent relay a protective relay thatoperates when fed a current larger than itsminimum pick-up value.

overdamped more damping than a crit-ically damped system. For a characteristicequation of the form:

s2 + 2ζωns + ω2n

the system is overdamped ifζ < 1.0; theroots of the characteristic equation are com-plex conjugate pairs.

overexcited the condition where the fieldwinding current is greater than a rated valuethat produces the rated MMF in the armature.With a synchronous or a DC machine, theexcitation current is a direct current in thefield windings.

If a machine is overexcited, the excessMMF must be counterbalanced in the arma-ture. In the case of a DC motor, the overex-citation is counterbalanced by the increase ofarmature current, which is translated by theincrease of both torque and speed. In the case

of synchronous motor, a leading componentof the armature current is present and the ma-chine operates at a leading power factor.

overflow a data condition in arithmeticoperations of signed numbers where the mag-nitude of a result exceeds the number of bitsassigned to represent the magnitude. The re-sult changes the sign bit, thus, making theresult incorrect.

overlap angle Seecommutation angle.

overlapped execution processing severalinstructions during the same clock pulses.

overlay (1) in wireless communications,refers to a modulated signal (communicationscheme) that occupies an RF channel that isalready occupied by other signals. The clas-sical example is that of a spread spectrumsignal that is (spectrally) placed on top of aset of narrow band channels that carry nar-row band signals. The wideband spread spec-trum signal has a low power spectral density;hence the overlay scheme can coexist withthe narrow band signals occupying the fre-quency band.

(2) a vector describing the positional ac-curacy with which a new lithographic patternhas been printed on top of an existing patternon the wafer, measured at any point on thewafer. Also called registration.

overload a situation that results in elec-trical equipment carrying more than its ratedcurrent. Placing too much electrical load ona generator or too much mechanical load ona motor would cause an overload.

overload heater a term used to describethe thermal sensors that detect motor over-load currents. Usually located on the motorstarter, the heaters cause the overload relayto operate.

overload protection a protective devicewhich opens the circuit to a piece of electri-


cal equipment or power line in the event ofcurrent exceeding the upper design limit.

overload relay a device designed to de-tect and interrupt motor overload conditions.Motor overload relays may be actuated bythermal (temperature), magnetic (current), orelectronic (voltage and current) sensors.

overmoded the condition of a waveguideat a frequency where two or more modes areabove cutoff (propagating).

overmodulation technique used in pulse-width-modulated (PWM) switching schemesto obtain a higher output voltage. Overmod-ulation occurs when the control (or modu-lation) signal magnitude exceeds the magni-tude of the triangle carrier signal that it isbeing compared to and the number of pulsesbegin to drop out. This does introduce greateroutput voltage distortion.

oversampling sampling a continuous-time signal at more than the Nyquist fre-quency.

oversampling converter A/D converterthat samples frequencies at a rate muchhigher than the Nyquist frequency. Typicaloversampling rates are 32 and 64 times thesampling rate that would be required with theNyquist converters.

overshoot the amount by which an outputvalue momentarily exceeds the ideal outputvalue for an underdamped system.

overstress failure failure mechanismsdue to a single occurrence of a stress eventwhen the intrinsic strength of an element ofthe product is exceeded.

overvoltage a voltage having a valuelarger than the nominal voltage for a dura-tion greater than 1 minute.

overvoltage relay a protective relay thatoperates on overvoltage.

over-excitation limiter Seemaximumexcitation limiter.

oxidation for a semiconductor manufac-turing process, the process of growing silicondioxide on a silicon wafer subjected to ele-vated temperature in an oxygen-containingenvironment.

oxide charge the charge in the oxide,which may be grown in, or may be introducedby charge injection from, external sources orby ionizing radiation. Its presence in the ox-ide of a MOS device causes a shift in theflatband voltage and the threshold voltage.

oxide trap a defect, impurity, or disor-dered bond in the oxide that can trap an elec-tron or a hole.


PP commonly used symbol for power inwatts or milliwatts.

PDC common symbol for DC power inwatts.

Pinput common symbol for power input toa device in watts.

Pload common symbol for power deliveredto the load.

Pref common symbol for power referencelevel in watts or milliwatts.

p-channel MOSFET a MOSFET wherethe source and drain are composed of heav-ily doped p-type semiconductor regions ina n-type surface. Holes form drain-sourcecurrent when the applied gate and substratepotentials invert the n-type surface betweenthem.

P-I-N photodiode a photodiode (detec-tor) in which a layer of intrinsic (undoped)material is added between the p-n junction.This has the effect of increasing the amountof incident optical power absorbed in the de-vice and hence the efficiency in convertingoptical power into electrical current.

p-n junction (1) a junction between re-gions of the same bulk material that differ inthe concentration of dopants, n-type on oneside and p-type on the other.

(2) metallurgical interface of two regionsin a semiconductor where one region con-tains impurity elements that create equiva-lent positive charge carriers (p-type) and theother semiconductor region contains impu-rities that create negative charge carriers (n-type).

(3) a physical region where n-type and p-type materials are in contact (Seedoping).The diode is based on a single p-n junction.

P-well a region of p-type semiconductorlocated at the surface of a n-type substrate(or larger N-well) usually created in order tocontain n-channel MOSFETs.

P1dB acronym for 1 dB compressionpower. This gives a measure of the maxi-mum signal power level that can be processedwithout causing significant signal distortionor saturation effects. Technically, this refersto the power level at the input or the output ofa component or system at which the satura-tion of active devices like transistors causesthe gain to be compressed by 1 dB from thelinear gain.

PAC learning a supervised learningframework in which training examplesx arerandomly and independently drawn from afixed, but unknown, probability distributionon the set of all examples. Each exampleis labeled with the valuef (x) of the targetfunction to be learned. A PAC (probably,approximately correct) learning algorithm isone which, on the basis of a finite numberof examples, is able, with high probability,to learn a close approximation to the targetfunction.

package in MMIC technology, die orchips have to ultimately be packaged to beuseful. An example of a package is the T07“can.” The MMIC chip is connected withinthe can with bond wires connecting from padson the chip to lead pins on the package. Thepackage protects the chip from the environ-ment and allows easy connection of the chipwith other components needed to assemblean entire system, such as a DBS TV receiver.

packed decimal a data format for the effi-cient storage and manipulation of real num-bers, similar to BCD, with digits stored indecimal form, two per byte.


packet a unit of data which is sent overa network. A packet comprises a payloadcontaining some data, and either a header ora trailer containing control information.

packet reservation multiple access(PRMA) a user transmitting a packet ina particular slot will have the correspondingslot reserved in the next frame. The basestation will broadcast the state of the slots(reserved and free for contention) at the be-ginning of each frame.

packet switching means of switching dataamong the ports (inputs and outputs) of aswitch such that the data is transferred in unitsof variable size.

packet-switched bus Seesplit transaction.

pad (1) a device (network) that impedancematches and/or attenuates. Typically used torefer to a coax attenuator.

(2) a concrete foundation, usually pre-fabricated and used to support power trans-formers in underground residential distribu-tion work.

pad-mount transformer a heavily-enclosed distribution transformer mounted atgrade level upon a concrete slab or pad.

Pade approximation a pole-zero approx-imation of deadtime based on the expansion

e−sτ =1 − θ + 1

2!θ2 − 1

3!θ3 + . . .+ 1

n!θn + . . .

1 + θ + 12!θ2 − 1

3!θ3 + . . .+ 1n!θn + . . .

whereθ = sτ/2. When the infinite seriesin both numerator and denominator are trun-cated tom terms, the approximation containsm zeros andm poles and is known as themthorder Pade approximation to deadtime. Thisformula allows pole-zero methods like rootlocus or pole placement which cannot nor-mally deal with system containing deadtime,to be applied to such systems.

page Seevirtual memory.

page fault event that occurs when the pro-cessor requests a page that is currently not inmain memory. When the processor tries toaccess an instruction or data element that ison a page that is not currently in main mem-ory, a page fault occurs. The system mustretrieve the page from secondary storage be-fore execution can continue.

page frame a contiguous block of mem-ory locations used to hold a page.See alsovirtual memory.

page miss penalty when a page miss oc-curs, the processor will manage the load ofthe requested page as well as the potentialreplacement of another page. The time in-volved, which is entirely devoted to the pagemiss, is referred to as the page miss penalty.

page offset the page offset is the indexof a byte or a word within a page, and iscalculated as the physical as well as virtualaddress modulus of the page size.

page printing a printing technique wherethe information to be printed on a page iselectronically composed and stored beforeshipping to the printer. The printer then printsthe full page nonstop. Printing speed is usu-ally given in units of pages per minute (ppm).

page replacement at a page miss, whena page will be loaded into the main memory,the main memory might have no space leftfor that page. To provide space for that newpage, the processor will have to choose a pageto replace.

page table a mechanism for the transla-tion of addresses from logical to physical in aprocessor equipped with virtual memory ca-pability. Each row of the page table containsa reference to a logical block of addressesand a reference to a corresponding block ofphysical storage. Every memory reference istranslated within the CPU before storage is


accessed. The page table may itself be storedin standard memory or may be stored within aspecial type of memory known as associativememory. A page table stored in associativememory is known as a translation lookasidebuffer.

page-fault-frequency replacement a re-placement algorithm for pages in main mem-ory. This is the reciprocal of the time betweensuccessive page faults. Replacement is ac-cording to whether the page-fault frequencyis above or below some threshold.

page-mode DRAM a technique that usesa buffer like a static RAM; by changing thecolumn address, random bits can be accessedin the buffer until the next row access or re-fresh time occurs.

This organization is typically used inDRAM for column access. Additional tim-ing signals allow repeated accesses to thebuffer without a row-access time.See alsotwo-dimensional memory organization.

page-printing printer a human-readableoutput device used for producing documentsin a written form. The printer stores a wholepage in memory before printing it (e.g., alaser printer).

paged segmentation the combination ofpaging and segmentation in which segmentsare divided into equal sized pages. Allowsindividual pages of a segment rather than thewhole segment to be transferred into and outof the main memory.

paged-segment a segment partitionedinto an integral number of pages.

paging the process of transferring pagesbetween main memory and secondary mem-ory.

paging channel a channel in a wirelesscommunication system used to send pagingmessages. Paging messages are typicallyused to set up telephone calls and also to

send control messages to terminals that areinvolved in a communication session.

Pagourek–Witsenhausen paradox thebest known result concerning comparison ofperformance sensitivity between open-loopand closed-loop nominally optimal systems.The result provides the somewhat deceptiveanswer to a question of what is the devia-tion of the performance index for a closed-loop and open-loop implementations of nom-inally optimal control systems due to param-eter deviations from its nominal value (forwhich the optimal control has been calcu-lated). The answer considered either obvi-ous or paradoxical can be expressed as fol-lows: the infinitesimal sensitivity of the per-formance index expressed by its first varia-tion caused by a variation of the parametervector is the same whether an open-loop ora closed-loop implementation of the nomi-nally optimal control is used. The result pri-mary found for linear-quadratic problem hasbeen generalized for nonlinear time-varyingsufficiently smooth optimal control problemswith free terminal state.

PAL Seephase alternate lineor program-mable array logic.

PAM Seepulse amplitude modulation.

PAM system Seeprism | air | metalsystem.

panelboard an assembly of one or morepanel units containing power buses, auto-matic overcurrent protective devices, that isplaced in a cabinet or cutout box located inor flush on a wall. The assembly can onlybe accessed from the front and may containswitches for operation of light, heat, or powercircuits. Seeswitchboard.

pantograph an apparatus for applyingsliding contacts to the power lines above anelectric railroad locomotive.


paper tape strips of paper capable of stor-ing or recording information, most often inthe form of punched holes representing thevalues. Now obsolete.

Papoulis’ generalization a sampling the-ory applicable to many cases wherein sig-nal samples are obtained either nonuniformlyand/or indirectly.

parabolic index profile quadratic trans-verse variation of the index of refraction;leads to analytic solutions of the paraxialequations for rays and beams.

parabolic reflector a reflecting surfacedefined by a paraboloid of revolution or sec-tion of a paraboloid of revolution.

paraelectric the nonpolar phase intowhich the ferroelectric transforms aboveTc,frequently called the paraelectric phase.

parallel adder a logic circuit that addstwo binary numbers by adding pairs of digitsstarting with the least significant digits. Anycarry generated is added with the next pairof digits. The term “parallel” is misleading,since all the digits of each numbers are notadded simultaneously.

parallel architecture a computer sys-tem architecture made up of multiple CPUs.When the number of parallel processors issmall, the system is known as a multipro-cessing system; when the number of CPUsis large, the system is known as a massivelyparallel system.

parallel bus a data communication pathbetween parts of the system that has one linefor each bit of data being transmitted.

parallel computing computing per-formed on computers that have more then oneCPU operating simultaneously.

parallel computing system a systemwhose parts are simultaneously running ondifferent processors.

parallel data transfer the data transferproceeds simultaneously over a number ofpaths, or a bus with a width of multiple bits,so that multiple bits are transferred every cy-cle. A technique to increase the bandwidthover that of serial data transfer.

parallel feed Seecorporate feed.

parallel I/O Seeparallel input/output.

parallel I/O interface I/O interface con-sisting of multiple lines to allow for the si-multaneous transfer of several bits. Com-monly used for high-speed devices, e.g., disk,tape, etc.See alsoserial I/O interface.

parallel input/output generic class of in-put/output (I/O) operations that use multiplelines to connect the controller and the periph-eral. Multiple bits are transferred simultane-ously at any time over the data bus.

parallel interference cancellation a mul-tiple access interference cancellation strategyfor CDMA based on multistage detection. Inthe first stage, tentative decisions are madefor each user in parallel. For each user, anestimate of the resulting multiple access in-terference is made and substracted from thereceived signal. In the succeeding stage, amore reliable tentative decision can then bemade for each user. This proceeds succes-sively for a predetermined number of stages.

parallel manipulator manipulator thatconsists of a base platform, one moving plat-form and various legs. Each leg is a kine-matic chain of the serial type, whose end linksare the two platforms. Parallel manipulatorscontain unactuated joints, which makes theiranalysis more complex than those of serialtype. A paradigm of parallel manipulatorsis the flight simulator consisting of six legsactuated by hydraulic pistons.


parallel paths the number of separatepaths through the armature winding that ex-ist between the brushes of a DC machine.In a DC machine’s armature, the conductorsand coils are placed in their slots and con-nected to the commutator using either thelap winding method or the wave windingmethod. The number of conductors that areconnected in parallel depend on the numberof poles the machine has, and whether thewinding connections are lap or wave. For thelap wound armature, the number of parallelpaths is found by multiplying the number ofpoles by the number of revolutions it takes tofill all the slots of the armature. The numberof revolutions it takes to fill the slots is knownas the machine’s “plex” value. In a simplexwound armature, the “plex” value is 1, duplexhas a “plex” value of 2, triplex has a “plex”value of 3, and so on. For the wave woundarmature, the number of parallel paths is twotimes the “plex” value. This same conceptcan also be applied to AC machinery.

parallel plate waveguide a type ofwaveguide formed by two parallel metallicplates separated by a certain distance. It sup-ports propagation of several types of modes:the TEM mode and the TE and TM modefamilies.

parallel port a data port in which a col-lection of data bits are transmitted simulta-neously.

parallel processing (1) an environmentin which a program is divided into multiplethreads of control, each of which is capableof running simultaneously, at the same timeinstant.

(2) processing carried out by a numberof processing elements working in parallel,thereby speeding up the rate at which opera-tions on large data sets such as images can beachieved. Often used in the design of real-time systems.

parallel transmission the transmission ofmultiple bits in parallel.

parallel-to-serial conversion a processwhereby data, whose bits are simultaneouslytransferred in parallel, is translated to datawhose bits are serially transferred one at atime. During the translation process, sometiming information may be included (such asstart and stop bits) or is implicitly assumed.

parallel-transfer disk a disk in which it ispossible to simultaneously read from or writeto multiple disk surfaces. Advantageous inproviding high data transfer rates.

parallelism the possibility to simultane-ously execute different parts of a system ondifferent processors.

Although parallelism can be found inmany electronic computing systems, paral-lelism is the inherent property of an opticalsystem. For example, a lens, the simplest op-tical system, forms the whole image at onceand not point-by-point or part-by-part. If theimage consists of one million points, the lensprocesses one million data in parallel.

parallelogram mechanism a manipula-tor that has a kinematic chain of the serialtype and part of its kinematic chain forms aclosed kinematic chain.

paramagnetic materials with permeabil-ity slightly greater than unity. Sodium, potas-sium, and oxygen are examples.

parameter coding also called vocoding.In parameter coding, the signal is analyzedwith respect to a model of the vocal mech-anism and the parameters of the model aretransmitted.

parameter estimation the procedure ofestimation of model parameters based on themodel’s response to certain test inputs.

parameter matrix anN × N matrix ofcomplex linear parameters which describethe behavioral relationships between theNports of a circuit or network. The most com-monly utilized parameter matrices are con-


ductance (G), impedance (Z), admittance(Y ), (H ), scattering (S), chain (ABCD) andscattering chain (8), all of which can be read-ily transformed from one to another throughsimple algebraic manipulations. Any set ofthese parameters are referenced to given portsource and/or load impedance, at fixed inputfrequencies and fixed power levels.

parameter space a domain formed by allpossible values of the given parameters.

parametric amplification a nonlinearoptical process in which a signal wave offrequencyωs and a higher-frequency pumpwave of frequencyωp propagate througha second-order nonlinear optical material,leading to the amplification of the signalwave and the generation of an idler waveof frequencyωp − ωs . See alsoopticalparametric oscillator.

parametric coding refers to the class ofsignal compression methods that are basedon a criterion where parameters, or features,of the signal are extracted and coded. Con-trasts waveform coding techniques, since thereproduced waveform can be (analytically)quite different from the input, but will stillbe a subjectively (in terms of vision or hear-ing) good, even indistinguishable, replica ofthe input signal.

parametric fixed form control rule con-trol rule given in a predeclared form, with anumber of parameters to be tuned; tuning ofthose parameters can be performed either off-line, prior to control rule implementation, oron-line. A classical example of a parametricfixed form control rule is PID industrial con-troller; another example is a neural networkbased control rule with a number — usuallylarge — of parameters of the network to betuned.

parametricoscillator Seeoptical parametric oscillator.

parasitic capacitance the generally unde-sirable and not-designed-for capacitance be-tween two conductors in proximity of oneanother.

parasitic inductance the generally unde-sirable and not-designed-for inductance as-sociated with a conductor, or path of currenton a conductor.

parasitic reactance the generally unde-sirable and not-designed-for reactance asso-ciated with one or more conductors in a cir-cuit.

parasitic resistance the generally unde-sirable and not-designed-for resistance asso-ciated with a conductor, or path of current ona conductor.

paraxial approximation neglect of thesecond derivative of the nearly plane-waveamplitude in the direction of propaga-tion; makes possible analytic solutions fordiffracting beams.

paraxial optics formalism for optics inwhich the paraxial approximation is em-ployed.

paraxial ray ray propagating so nearlyparallel to thezaxis that its length can be con-sidered equal to thez propagation distance.

paraxial ray equation set of second-order differential equations in the propaga-tion distance for the trajectory of a light raypropagating almost parallel to a fixed axis.

paraxial wave solution of the scalar waveequation in the paraxial approximation.

PARCOR coefficients Seepartial-correla-tion parameters.

parity property of a binary sequence thatdetermines if the number of 1’s in the se-quence is either odd or even.


parity bit an extra bit included in a bi-nary sequence to make the total number of1’s (including itself) either odd or even. Forinstance, for the following binary sequence101, one would insert a parity bitP(odd)=1to make the total number of 1’s odd; a paritybit P(even)=0would be inserted to make thisnumber even.See alsoerror detecting code.

parity check matrix a matrix whose rowsare orthogonal to the rows in the generatormatrix of a linear forward error control blockcode. A nonzero result of element-wise fi-nite field multiplication of the demodulatedword by this matrix indicates the presence ofsymbol errors in the demodulated word.

Is generated from the parity check poly-nomial of any linear(n, k) code and has di-mension of(n − k × n). It is used by thedecoder for error detection by checking theparity bits.

parity detection circuit a parity checklogic incorporated within the processor to fa-cilitate the detection of internal parity errors(reading data from caches, internal buffers,external data, and address parity errors).

parity-check code a binary linear blockcode.

Park’s transformation a change of vari-ables represented by a linear matrix multi-plication used in the analysis of electric ma-chines.Seerotor reference frame.

parking on a bus, a priority scheme thatallows a bus master to gain control of the buswithout arbitration.

parse tree the tree that is used for parsingstrings of a given language.

Parseval’s equation See Parseval’stheorem.

Parseval’s theorem a relationship thatstates that the integral of the square of themagnitude of a periodic function is the sum

of the square of the magnitude of each har-monic component.

Rigorously, suppose that two continuoustime signalsf1(t)andf2(t)have correspond-ing Fourier transformsF1(ω) and F2(ω),and thatF2(ω) is the complex conjugate ofF2(ω). Then Parseval’s theorem states that∫ ∞

−∞f1(t)f2(t)dt =1

2π

∫ ∞

−∞F1(ω)F2(ω)dω .

If f1(t) = f2(t), then the left-hand side ofthe above equation provides an expression ofthe energy of a signal, which can be relatedto its Fourier transform as follows:∫ ∞

−∞f (t)2dt = 1

2π

∫ ∞

−∞|F(ω)|2dω .

parsing the process of detecting whethera given string belongs to a given language,typically represented by grammars.

partial coherence the ratio of the sineof the maximum half-angle of illuminationstriking the mask to the numerical apertureof the objective lens. Also called the degreeof coherence, coherence factor, or the pupilfilling function, this term is usually given thesymbols.

partial element equivalent method anintegral equation technique in which the elec-tromagnetic problem is reduced to a lumpedcircuit problem by defining some regions inspace associated with a node in the lumpedcircuit. This method takes electric field inter-actions in the original problem into accountby finding (through the integral equations)either a capacitor to ground at infinity anda summation of current controlled currentsources or a capacitor connected in seriesto a summation of voltage controlled volt-age sources. The magnetic field interactionsare taken into account by finding an induc-tance in series with a summation of currentcontrolled voltage sources which is placedbetween nodes.


partial fraction expansion the methodof partial fraction expansion consists of tak-ing a function that is the ratio of polynomialsand expanding it as a linear combination ofsimpler terms of the same type. This toolis useful in inverting Fourier, Laplace or z-transforms and in analyzing linear time in-variant systems described by linear constant-coefficient differential or difference equa-tions.

partially coherent illumination a typeof illumination resulting from a finite sizesource of light that illuminates the mask withlight from a limited, nonzero range of direc-tions.

partial restoration refers to the situationwhere not all customers who are part of alarger outage can have power restored.

Utilities typically make efforts to restorepower to all customers, but situational fac-tors may dictate that the restoration processproceed in stages.

partial-correlation parameters the pa-rameters that are obtained when solving theautocorrelation equations for the problem oflinear prediction.

partially decorrelating noise whiteningthe process of transforming the matched-filtered sufficient statistic of a CDMA signalinto a corresponding sufficient statistic whichis affected by uncorrelated (white) noise andis partially decorrelated, i.e., the resultingsufficient statistic for aK-user system is de-scribed by

x = Fd + n ,

wherex is a lengthK column vector of suffi-cient statistics,F is aK ×K lower left trian-gular matrix,d is a lengthK column vectorof data symbols, andn is a lengthK columnvector of AWGN noise samples.

participation factor the ratio of thechange in the power output of a generatorversus the total change in power demand.The participation factor of each generator in

a power system is found in the solution of theeconomic dispatch problem.

partition noise electrical noise gener-ated within a vacuum tube when the electronstream hits obstacles and is divided.

partition table a table that enables thegrouping of states into equivalent sets.

partitioned process controlled processthat is considered as consisting of severalsub-processes that can be interconnected;process partitioning is an essential step incontrol problem decomposition — for exam-ple, before decentralized or hierarchical con-trol can be introduced. The term “partitionedsystem” is used when one is referring to thepartitioned control system, but is also oftenused to denote just the partitioned controlledprocess.

passband (1) the range of frequencies, orfrequency band, for which a filter passes thefrequency components of an input signal.

(2) the frequency difference between thehigher and lower band edges, expressed inradians/second. The band edges are usuallydefined as the highest and lowest frequencieswithin a contiguous band of interest at whichthe loss equals the maximum attenuation lossacross the band.Compare withstopband.

passband edge the frequency at whichthe signal becomes significantly attenuated;typically the frequency at which the signalis attenuated at 3 dB from the maximum re-sponse.

passband ripple difference between themaximum attenuation loss and the minimumattenuation loss across the band, expressedin decibels. This parameter is also knownas the loss attenuation in ratio form. Theband edges are usually defined as the highestand lowest frequencies within a contiguousband of interest at which the loss equals themaximum attenuation loss across the band.


passivation the process in which an in-sulting dielectric layer is formed over thesurface of the die. Passivation is normallyachieved by thermal oxidation of the siliconand a thin layer of silicon dioxide is obtainedin this manner. Other passivation dielectriccoatings may also be applied, such as siliconglass.

passive backplane in printed circuitboards, a circuit board in which other boardsare plugged, which contains no active circuitelements to control signal quality.

passive filter a filter circuit that uses onlypassive components, i.e., resistors, inductors,and capacitors. These circuits are useful athigher frequencies and as prototypes for lad-der filters that are active.

passive magnetic bearing a magneticbearing that does not require input energy forstable support during operation. Generallyimplemented using permanent magnets.

passive network an electronic circuitmade up of passive elements. Passive ele-ments are capacitors, resistors and inductors,and have no gain characteristics.

passive optical network a network wherethe optical fiber plays the role of a broad-band, passive interconnect, and the functionsof switching and control are made using elec-tronics.

passive redundancy a circuit redundancytechnique which assures fault masking byerror correcting codes or N-modular redun-dancy with voting.

passivity naturally associated with powerdissipation. It can be defined for linear aswell as nonlinear systems. A formal defini-tion of passivity requires a representation ofsystems by an operator mapping signals tosignals. The signal space under considera-tion is assumed to be extendedL2 space with

a scalar product defined by

〈x|y〉 =∫ ∞

0x(s)y(s)ds

A system with inputu and outputy is passiveif

〈x|y〉 ≥ 0

The system is input strictly passive if thereexistsε > 0 such that

〈x|y〉 ≥ ε‖u‖2

and output strictly passive if there existsε >0 such that

〈x|y〉 ≥ ε‖y‖2

path the space curve that the manipulatorend-effector moves along from the initial lo-cation (position and orientation) to the finallocation is called the path. Notice that thepath is a pure geometric description of mo-tion. Path can be specified in the joint spaceor in the operational space.

path-delay testing any one of several pos-sible techniques to verify that signal transi-tions created by one clock event will travelthrough a particular logic/path in a subcir-cuit, IC component, or system and will reachtheir final steady-state values before a subse-quent clock event.

path-set the set of all edges in a path.

patrolling in overhead power lines, ac-tion taken when the location of a fault is notknown. The crew will typically follow over-head spans until the location of the outage isfound.

pattern a plot of the distribution of ra-diated power. Typically, the pattern consistsof a main lobe (major lobe), in which mostof the radiated power is confined, and a num-ber of progressively weaker sidelobes (minorlobes).


pattern classification assignment of apattern (typically a vector of measurementsor a feature derived from measurements) to aclass.

pattern matching (1) the detection, inan image, of a subimage in which pixel val-ues are structured according to a predefinedschema.

(2) a process by which patterns may berecognized, potentially and sometimes inpractice by template matching, but otherwiseby inference following the location of fea-tures. Typically, inference is carried out byapplication of Hough transforms or associa-tion graphs.

pattern recognition the feature extrac-tion, clustering, and classification processesassociated with assigning meaning to mea-surements. There are statistical, syntacticand structural methods for pattern recogni-tion, and neural network methods are oftenconsidered as a subset of pattern recognition.

pattern sensitive integrated circuits inwhich an error may result from a certain datapattern being encountered.

pattern synthesis the process of design-ing an antenna such that its radiation charac-teristics meet desired specifications.

patterning the processes of lithography(producing a pattern that covers portions ofthe substrate with resist) followed by etching(selective removal of material not covered byresist) or otherwise transferring the patterninto the substrate.

pause instruction an assembly languageinstruction whose execution causes a mo-mentary pause in program execution.

Pawlak’s information system a systemmodel denotedS can be viewed as a pairS = (U,A), whereU = {x1, . . . , xn} is anonempty finite set of objects. The elementsofU may be interpreted, for example, as con-

cepts, events, goals, political parties, individ-uals, states, etc. The setU is called the uni-verse. The elements of the setA, denotedaj ,j = 1, . . . , m, are called the attributes. Theattributes are vector-valued functions. Forexample,

aj : U → {−1,0,1} .

An example of a simple information systemis shown in the following table.

a1 a2 a3x1 -1 0 0x2 0 0 1x3 0 1 1x4 1 1 1

The first component ofa1 being−1 maymean thatx1 is opposed to the issuea1, whilex4 supporta1, etc. Such data can be collectedfrom newspapers, surveys, or experts.

payload the portion of a packet that isneither the header nor the trailer and is eitheruser- or protocol-specific data.

PC See program counter, personalcomputer.

PC-relative addressing an addressingmechanism for machine instructions in whichthe address of the target location is given bythe contents of the program counter and anoffset held as a constant in the instructionadded together. Allows the target location tobe specified as a number of locations fromthe current (or next) instruction. Generallyonly used for branch instructions.

PCA Seeprincipal component analysis.

PCB See printed circuit board. poly-chlorinatedbiphenyls.

PCBA acronym for printed circuit boardassembly.

PCM Seepulse-code modulation.


PCN See personal communicationsnetwork.

PCS See personal communicationsservices.

PDC Seepersonal digital cellular.

PDF Seeprobability density function.

PDF-optimized scalar quantization SeeLloyd–Max scalar quantization.

PDMA Seepolarization division multipleaccess.

PDP Seepiecewise deterministic process.

PDU Seeprotocol data unit.

PE Seeprocessing element.

peak accumulation mode an operat-ing feature of digital storage oscilloscope inwhich the maximum and minimum excur-sions of the waveform are displayed for agiven point in time.

peak detector an electronic circuit whichoutputs a DC voltage which indicates thepeak amplitude of an alternating waveform.

peak let-through current the maximumvalue of the available short-circuit currentthat is let through a current-limiting fuse.Seealso current limiting fuse.

peak output power the instantaneouspeak power delivered to a load, expressedin watts. It is simply the maximum powerat any instant in time taken during the to-tal time period being considered. This termis often used when a device or applicationis subjected to time-varying signals such aspulse modulation.

peak power the maximum value of thesquare of the absolute value of a signal overthe duration of the signal; i.e., for the signal

x(t), maxt |x(t)|2. The maximum instanta-neous power of a signal.

peak signal to noise ratio (PSNR) objec-tive distortion measure of the difference be-tween a discrete time signalxi , defined over[1 . . . n], and a degraded, restored or other-wise processed version of the signalxi , de-fined as

PSNR = 10log10x2maxn∑n

i=1(xi − xi )2dB

PSNR differs from the standard signal tonoise ratio by using the square of the sig-nal’s maximum value rather than its variancein the numerator. In image coding, this valueis often assumed to be 255, for 8-bit images.

peaking generator (1) a utility generatingunit, typically driven by a gas-fired turbine,available to rapidly come on line when thesystem demand reaches its highest levels.

(2) a generator used by a plant to reducethe peak demand drawn in a given period oftime (typically during a one-month interval).

peaking unit a generator used only to sup-ply peak periods of electric power demand.

Pearl Street Station the first investor-owned electric utility plant, started by Edisonin 1882 in New York City. It provided low-voltage, DC electric service to 85 customerswith an electric load of 400 lamps.

PEB Seepost-exposure bake.

PEC Seeperfect electric conductor.

pel a picture element that has been en-coded as black or white, with no gray scalein between.

pellicle a thin, transparent membraneplaced above and/or below a photomask toprotect the photomask from particulate con-tamination. Particles on the pellicle are sig-nificantly out of focus, and thus have a muchreduced chance of impacting image quality.


penalty function Seecost function.

penstock a water tube that feeds the tur-bine. It is used when the slope is too steepfor using an open canal.

pentode vacuum tube with five activeelectrodes: cathode, control grid, screengrid, suppressor grid, plate.

per-unit system a dimensionless systemfor expressing each quantity in terms of afractional part of a “base” value, often thenominal or rated value of the system. Typicalelectrical calculations require four base quan-tities (voltage, current, impedance, and ap-parent power), any two of which may be cho-sen arbitrarily. The per-unit system greatlysimplifies calculations in electrical systemscontaining transformers with non-unity turnsratios, making the voltage differences trans-parent.

percent impedance the per-unit impedanceexpressed as a percentage on a certain MVAand voltage base.

percent system a variation of the per-unitsystem in which the ratios expressing systemquantities are expressed as a percentage ofthe base quantity.

perceptron one of the earliest neural algo-rithms demonstrating recognition and learn-ing ability. Since the output of neuron is bi-nary, the neuron can classify an input intotwo classes,A andB. The perceptron modelcontaining a single neuron is expressed asfollows: z = 1, if

∑wixi > T andz = 0,

if∑wixi < T wherez is the output,xi is

an input,wi is an element of the intercon-nection weight matrix, andT is a general-ized threshold. If the interconnection weightmatrix is known, an input vector(xi) can beclassified intoA orB according to the resultof z. On the other hand, the interconnectionweight matrix can be formed if a set of in-put vectors(xi) and their desired outputszare known. The process of forming the inter-

connection weight matrix from known input–output pairs is called learning. The percep-tron learning is as follows. First, the weightwi and the thresholdT are set to small ran-dom values. Then the outputz is calculatedusing a set of known input. The calculatedzis compared with the desired outputt . Thechange of weight is then calculated as fol-lows: δwi = η(t − z)xi whereη is the gainterm that controls the learning rate, which isbetween 0 and 1. The learning rule deter-mines that the corrected weight is

wi = wi(old)+ δwi

The process is repeated until(t−z) = 0. Thewhole learning process is completed after allinput–output pairs have been tested with thenetwork. If the combination of inputs are lin-early separable, after a finite number of steps,the iteration is completed with the correct in-terconnection weight matrix. Various opticalsystems have been proposed to implement theperceptron, including correlators.

The original perceptron was a feedfor-ward network of linear threshold units withtwo layers of weights, only one layer ofwhich (the output layer) was trainable, theother layer having fixed values.

perceptron convergence procedure a su-pervised learning technique developed forthe original perceptron. If the outputyi ofunit i in the output layer is in error, its in-put weights,wij , are adjusted according toDwij = h(ti − yi)xj , whereti is the targetoutput for uniti andh is a positive constant(often taken to be unity);xj is the output ofunit j in the previous layer which is multi-plied by weightwij and then fed to uniti inthe output layer.

perceptual coding involves the coding ofthe contextual information of the image fea-tures by observing the minimum perceptuallevels of the human observer.

perfect code a t-error correcting forwarderror control block code in which the num-ber of nonzero syndromes exactly equals the


number of error patterns oft or fewer errors.Hamming codes and Golay codes are the onlylinear nontrivial perfect codes.

perfect electric conductor (PEC) a con-ductor that has infinite conductivity or zeroresistivity.

perfect reconstruction the conditionstates that the output of a filter bank is a de-layed version of the corresponding input. Inother words, there are no aliasing and (phaseand magnitude) distortions for the output ofthis filter bank.

perfect shuffle interconnects that con-nect sources 1,2,3,4,5,6,7,8 to detectors1,5,2,6,3,7,4,8, respectively. The op-eration divides 1,2,3,4,5,6,7,8 into twoequal parts 1,2,3,4 and 5,6,7,8 and theninterleaves them. The size of the array mustbe 2n. The array returns to its original or-der aftern operations. When the option toexchange pairs of neighboring elements isadded to a perfect shuffle network, any arbi-trary permutation of the elements is achiev-able.

performability the probability that a sys-tem is performing at or above some level ofperformance,L, at the instant of timet .

periodic convolution a type of convo-lution that involves two periodic sequences.Its calculation is slightly different from thatof discrete linear convolution in that it onlytakes summations of the products within oneperiod instead of taking summations for allpossible products. Circular convolution isan operation applied to two finite-length se-quences. In order to make its result equalto that of the linear convolution, the two se-quences have to be zero-padded appropri-ately. These two zero-padded sequences canthen serve as periods to formulate two peri-odic sequences. At this time, the result ofthe periodic convolution is equal to circularconvolution.

periodic coordination coordination pro-cess occurring when a sequence of controldecisions is required over a given time in-terval and these decisions are directly madeby the local units but the operation of theseunits is periodically adjusted by the coordina-tor; an example is operation of an industrialprocess controller with several independentregulators, which at specified time instantsare provided with new parameters or additivecompensation signals so as to achieve betterresults from the overall point of view, for ex-ample, to achieve better transient responsesto the changes of the free inputs.

periodic signal a continuous time signalf (t) with periodT ≥ 0 such that

f (t) = f (t + T ) for all − ∞ < t < ∞ .

The smallest suchT is its fundamentalpe-riod. For example, the signal

f (t) = A sin(ωt + φ)

is periodic with fundamental period 2π/ω.Let T be a positive integer. A discrete

time signalf [k] is periodic with periodT if

f [k] = f [k + nT ] for all integersk.

The smallest suchT is its fundamental pe-riod. For example,f [k] = cos[π2 k] is pe-riodic with period 4. On the other hand,f [k] = cos(k) is not periodic, as a positiveintegerT does not exist to satisfy the defini-tion.

If a signal is not periodic, then it isaperi-odic.

periodic structure structure consisting ofsuccessive identically similar sections, sim-ilarly oriented, the electrical properties ofeach section not being uniform throughout.Note that the periodicity is in space and not intime. The analysis of infinite periodic struc-ture is significantly simplified by the Flo-quet’s theorem.See alsoFloquet’s theorem.

periodic waveform phrase used to de-scribe a waveform that repeats itself in a uni-form, periodic manner. Mathematically, for


the case of a continuous-time waveform, thischaracteristic is often expressed asx(t) =x(t ± kT ), which implies that the waveformdescribed by the functionx(t) takes on thesame value for any increment of thekT ,wherek is any integer and the characteris-tic valueT > 0, describes the fundamentalperiod ofx(t). For the case of the discrete-time waveform, we writex(n) = x(t ± kN),which implies that the waveformx(n) takeson the same value for any increment of sam-ple numberkN , wherek is any integer and thecharacteristic value integerN > 0 describesthe fundamental period ofx(n).

peripheral an ancillary device to a com-puter that generally provides input/output ca-pabilities.

peripheral adapter a device used to con-nect a peripheral device to the main com-puter; sometimes called an I/O card, I/O con-troller, or peripheral controller.

peripheral control unit a device used toconnect a peripheral device to the main com-puter; sometimes called an I/O card, I/O con-troller, or peripheral controller.

peripheral controller See peripheralcontrol unit.

peripheral device a physical mechanismattached to a computer that can be used tostore output from the computer, provide inputto the computer, or do both.See alsoI/Odevice.

peripheral processor a computer thatcontrols I/O communications and data trans-fers to peripheral devices. It is capable ofexecuting programs much like a main com-puter.See alsoI/O channel.

peripheral transfer a data exchange be-tween a peripheral device and the main com-puter.

peripheral unit a physical mechanismattached to a computer that can be used tostore output from the computer, provide in-put to the computer, or do both.See alsoI/Odevice.

permalloy a family of ferromagnetic al-loys consisting of iron, nickel, and molyb-denum that saturate at moderate flux densitylevels and have a low coercive force.

permanent fault a fault that remains inexistence indefinitely if no corrective actionsare taken.

permanent magnet (PM) a magnet thatproduces an external magnetic field by virtueof the alignment of domains inside the ma-terial and retains its magnetism after beingsubjected to demagnetizing fields.

permanent magnet AC motor a genericterm used to describe both permanent magnetsynchronous motors and brushless DC mo-tors.

permanent magnet brushless DC machinea machine that is similar in structure to a per-manent magnet synchronous machine, con-taining armature windings on the stator andpermanent magnets on the rotor. The per-manent magnet brushless DC machine, how-ever, is characterized by a trapezoidal fluxdensity distribution in the airgap instead ofthe sinusoidal distribution of the synchronousmachine. In operation, a DC voltage is ap-plied sequentially to the stator coils to createa rotating field that pulls the rotor with it. Tocorrectly operate, the brushless DC machinerequires sensors to determine the rotor posi-tion so that the proper stator phases may beexcited.

permanent magnet DC machine a DCmachine in which the field excitation in thestator is provided by permanent magnets in-stead of electromagnets.


permanent magnet DC motor Seepermanent magnet DC machine.

permanent magnet machine a machinethat uses permanent magnets to establish thefield. In DC machines, the permanent mag-nets are placed on the stator, while on ACsynchronous machines they are placed on therotor.

permanent magnet stepper motor astepper motor that has a permanent magnetassembly on the rotor.

permanent magnet synchronous machinea polyphase AC motor with rotor mounted

permanent magnets and sinusoidal distribu-tion of stator phase windings. The field wind-ings in the rotor are replaced by permanentmagnets to provide the field excitation inthese machines.

permanent split-capacitor (PSC) motora induction motor that operates from a single-phase supply. The motor contains two phasewindings in quadrature; however, one ofthem has a capacitor in series with it to cre-ate a phase shift between the winding cur-rents. Both windings and the capacitor op-erate continuously so the machine acts like atwo-phase machine when running at its op-erating speed, producing less vibration andnoise than a single-phase motor. Since the ca-pacitor runs continuously, it is sized smallerthan the capacitor used in a capacitor-startinduction motor (CSIM). Thus, the PSC mo-tor produces a lower starting torque than theCSIM.

permeability tensor relationship betweenthe magnetic field vector and the magneticflux density vector in a medium with no hys-teresis; flux density divided by the magneticfield in scalar media. Permeability indicatesthe ease with which a magnetic material canbe magnetized. An electromagnet with ahigher permeable core material will producea stronger magnetic field than one with a

lower permeable core material. Permeabilityis analogous to conductance, when describ-ing electron flow through a material.See alsoreluctance.

permeameter making use of Hall effectgaussmeters, search coils, and flux meters,the permeameter, or hysteresigraph, recordsthe major hysteresis loop of a material, fromwhich its basic material properties can be de-termined: residual induction, coercivity, en-ergy product, saturation flux density, and re-coil permeability.

permeance the magnetic analog for con-ductance, indicating the ease with whichmagnetic flux will follow a certain path,which can be approximated by calculationsbased purely on magnetic circuit geometry.

permeance coefficient the slope of theload line for a magnetic circuit, determinedsolely by physical geometry of the magnetand permeable materials around it; the ratioof magnetic induction (B) and applied field(H ) at the operating point.

permission Seeaccess right.

permittivity Seeelectric permittivity.

persistent current a current circulating ina closed structure without applied potential.Examples are the supercurrent in a supercon-ducting magnet and the current in a closedmesoscopic ring in a magnetic field.

persistent spectral hole burning spectralhole burning with a long lifetime, usually onthe scale of seconds or longer.

personal computer a general term fora microcomputer used for such purposes asword processing, email, financial manage-ment, and game playing. While, the acronym“PC” is applied to microcomputers used bothfor business and personal use, the term “per-sonal computer” is more usually applied tothose used by individuals or families at home.


personal communications network (PCN)a telecommunications network designed to

provide services to a person rather than a geo-graphic location. The network may comprisea range of different technologies from end toend and contains within it intelligence to en-able the communication to be directed to theappropriate terminal or device carried withthe person.

personal communications services (PCS)a mobile telephone service with an essentialurban and suburban coverage characterizedby low cost pocket terminals, communica-tions at a price comparable to a cable tele-phone, and distribution of the services andproducts to the general public. This defini-tion is independent of the technology used.

PCS is the proposed next generation ofwireless network services, providing voicecommunication services similar to today’scellular services, only with smaller cells,lower power, and cheaper rates.

personal digital cellular (PDC) one typeof digital cellular phone system. PDC is800 MHz/1.5 GHz band, FDD, TDMA sys-tem, and it handles 1/4p shift QPSK modu-lated signal with 32 kbit/s. This cellular sys-tem was developed and operated in Japan.

personal handy phone system (PHS) adigital microcell system designed in Japanthat operates in the 1.9 GHz band. PHS pro-vides cordless telephone or telepoint servicessimilar to other digital cordless technologies,such as CT2. The spectrum allocation forPHS is 1895–1918 MHz. There are 77 ra-dio channels, each 300 kHz wide and dividedinto 2× 4 time slots. This provides a total of308 duplex traffic channels, with duplexingvia TDD. The carrier bit rate is 384 kb/s, andPHS terminals have a peak transmit power of80 mW.

perspective distortion a type of objectdistortion that results from projecting 3-Dshapes onto 2-D image planes by conver-gence of rays towards a center of projection:

See alsoperspective projection. This type ofdistortion is also called foreshortening.

perspective inversion a property of per-spective projection in which planar (typicallysilhouetted) objects that are not perpendicu-lar to the axis of projection will appear, in theabsence of additional information, to have ei-ther of two possible orientations in space.

perspective projection projection of a3-D object on a plane, termed “projectionplane,” as if it were imaged by an ideal pin-hole camera located in a point termed “centerof perspective.”

perspective transformation a matrixtransformation that represents the perspec-tive projection of objects in 3-D scenes into2-D images or the projection of one 2-D im-age into another 2-D image. Perspectivetransformations are conveniently carried outusing 4× 4 matrices representing homoge-neous co-ordinate transformations.

perturbations of controllable systemsthe set of controllable linear stationarycontinuous-time finite-dimensional dynami-cal systems is dense and open in the set ofall linear stationary continuous-time finite-dimensional dynamical systems with thesame state and input dimensions.

PET Seepositron emission tomography.

Petersen coil another term for a groundfault neutralizer.

petticoat another name for shed, a featureof an insulator.

PFN Seepulse forming network.

PGA Seefield-programmable gate array.

phantom an artificial target, sometimesdesigned to mimic the size, shape, and atten-uation characteristics of actual tissue, that is


used to test and calibrate imaging hardwareand software.

phase (1) a notion used extensively in in-terpreting complex quantities such as Fourierseries, Fourier transforms etc. Given a com-plex numberc = x+ iy = r cosφ+ ir sinφ,thenr represents the magnitude ofc andφ thephase.

(2) a horizontal translation parameter ofthe signal. Given a sinusoidal signals(t) =Asin(2πf t + φ), thenf represents the fre-quency (in Hz) ofs andφ the phase.

phase alternate line (PAL) a color televi-sion system that inverts the (R-Y) color signalon alternate lines. The color burst is locatedon the back porch of the composite video sig-nal and the phase alternates every line foridentification of the (R-Y) difference signalphase. The color demodulator uses the shiftin the color burst phase and the (R-Y) phasereversals to eliminate any hue changes causedby phase errors. The PAL color demodula-tion method eliminates the tint control butdoes cause desaturated colors.

The vertical scan parameters for the PALtelevision system are based on the Europeanpower line frequency of 50 Hz. The differentPAL standards in existence in different coun-tries allow for a slightly wider video band-width. The primary PAL system parameters(with the allowed bandwidth variance) are

Vertical Fields Frequency 50 HzNumber of Interlaced Fields 2Vertical Frame Frequency 25 HzNumber of Horizontal Lines/Frame 625Horizontal Scan Frequency 15,625 HzVideo Bandwidth 5 to 6 MHzColor Subcarrier Frequency 4.433618 MHzSound Modulation FMRF Channel Bandwidth 7 to 8 MHz

[* Alternative definitions for the 3 televi-sion systems that you may enjoy]

[NTSC - Never Twice the Same Color, SE-CAM - System Essentially Contrary to theAmerican Method, PAL - Perfect At Last]

phase angle the angle in the complexplane of a complex valuex = a + jb cal-

culated as

6 x = arctanb

a.

phase angle meter meter used to measurethe phase angle difference between two ACquantities. In power systems, typical metersuse perpendicular moving coils to measurethe phase angle between an AC current andan AC voltage. More accurate devices typi-cally measure the time interval between zerocrossings of the two input signals.

phase breaking the process by which thequantum mechanical phase, which is relatedto energy, of a particle is destroyed. Themost common example is the inelastic scat-tering due to electron–electron and electron–phonon interactions.

phase coded modulation (PCM) a formof angle modulation. The modulated wave,s(t), is given by:

s(t) = Ac cos[2πfct + kpm(t)],wherem(t) is the message signal,fc is thecarrier frequency, andkp is the phase sensi-tivity of the modulator.

phase comparator often referred to asa phase detector; a three-port device thatproduces an analog output proportional tothe phase difference between its two inputs.Since both inputs are periodic, the relativeoutput voltage (or current) as a function ofinput phase difference (i.e., the transfer func-tion) is also periodic; the shape of the trans-fer function (sometimes called the “outputcharacteristic” of the phase detector) dependsupon the particular technique used to accom-plish the phase detection. These includesinusoidal, triangula, and sawtooth shapefactors. Analog/digital implementation, re-quired linearity, and range of input phase dif-ference are primary factors in determining asuitable output characteristic for a specificphase comparator application.


phase comparison relay a phase com-parison relay is a protective relay used ontransmission lines which operates by com-paring phase angles of signals generated atopposite ends of the line. They employ adedicated communications channel to makethe comparison. The signals compared aretypically corresponding phase currents or se-quence currents.

phase conjugate mirror See phaseconjugator.

phase conjugation a technique for pro-viding dispersion compensation that relies onspectral inversion of the optical signal us-ing a nonlinear process known as four-wave-mixing.

phase conjugator an optical set-up orsystem that can generate the time-reversedreplica of the incident wave. Phase conjuga-tors play an important role in many opticalsystems that require the transmission of op-tical waves through scattering media such asatmosphere, optical fibers. Photorefractivecrystals such as BaTiO3, SBN, BSO are byfar the most efficient media for the generationof phase conjugate waves.

phase constant a constant, which is gen-erally complex, that is important in the studyof electromagnetic waves. The phase con-stant is equal to the frequency of excitationof the wave times 2 times pi times the squareroot of the product of the permeability andthe permittivity of the medium that the waveis traveling in. Also called propagation con-stant or wavenumber.

phase control a method for controllingthe amount of power delivered to a load byvarying the delay angle. This controls thedelay between the instant when the voltageacross the power semiconductor goes posi-tive and the actual start in conduction of thedevice.

phase delay the difference in the abso-lute angles between a point on a wavefront atthe device output and the corresponding pointon the incident input wavefront, expressed inseconds or degrees. The delay can exceed360 degrees.

phase detector gain the ratio of the DCoutput voltage of the phase detector to theinput phase difference. This is usually ex-pressed in units of volts per radian.

phase deviation the peak difference be-tween the instantaneous angle of the modu-lated wave and the angle of the carrier. Inthe case of a sinusoidal modulating function,the value of the phase deviation, expressed inradians, is equal to the modulation index.

phase discriminator a device for detect-ing the phase deviation of a phase modulatedsignal.

phase distortion problem that occurswhen the phase shift in the output signal of anamplifier is not proportional to the frequency.

phase error the difference in phase be-tween two sinusoidal waveforms having thesame frequency.

phase grating an optical grating charac-terized by a spatially periodic variation in therefractive index of a medium.

phase interrupting collision a collisionthat interrupts the phase of the wavefunctionsof lasing atoms and consequently broadentheir emission spectra.

phase locked loop a device containing avoltage controlled oscillator, a synchronousdetector and a low-pass filter, arranged in afeedback loop so that the oscillator outputcan be made to lock onto a weak incomingperiodic signal such as a radio wave. Also ofuse in frequency synchronization.


phase matching the condition that thephase of the nonlinear polarization bears aspatially fixed relation to that of the opticalfield generated by the polarization. This con-dition is a requirement for high efficiency innonlinear optical generation processes.

phase modulation a type of angle modu-lation whereby information is encoded ontoa carrier wave by modifying its phase angleas a function of time in proportion to the in-telligence signal amplitude.

phase modulator a device that alters thephase of a signal.

phase noise frequency variation in a car-rier signal that appears as energy at frequen-cies other than the carrier frequency.

phase only binary filter transmission orreflection phase plate in which neighboringregions differ in phase shift by pi radians.

phase parameter complex parameter rep-resenting corrections to the gain and phase ofa Gaussian beam.

phase plane a two-dimensional statespace.

phase plate transparent medium that in-troduces different phase shifts to differenttransverse regions of an optical wave for thepurpose of introducing or reducing phase oramplitude structure on the wave; often hav-ing only two phase shift values differing byπ .

phase portrait many different trajectoriesof a second-order dynamical system plottedin the phase plane.

phase response the way in which a systemalters the phase of an input sinusoid.

phase ripple the variation in phase re-sponse across the operating bandwidth of anoptical or electrical device.

phase sensitive detection Seesynchronousdetection.

phase sensitive measurement measure-ment in which the phase of an AC signal istracked in a feedback loop in order to improvedetection sensitivity. The most common ex-amples are phase-locked loops in control sys-tems and lock-in amplifier measurements inelectronics. In the latter, a small AC sig-nal is added to the bias voltage supply, andthis signal is then detected in any measuredquantity with an amplifier whose phase canbe varied to “lock onto” that of the initial ACsignal. This effectively mixes the two ACsignals, and their difference (at DC) is used tocharacterize the measurement. The effectivebandwidth is determined by the bandwidth atthe DC level.

phase sequence describes the rotationalorientation of the voltage phasors in a 3-phase electrical power system. A positivephase sequence, designated by the nomen-clature ABC, indicates a 3-phase connectionin which the B phase voltage lags the A phasevoltage by 120 degrees, and the C phase volt-age leads the A phase voltage by 120 degrees.A negative phase sequence, designated byACB, reverses this relationship so that theB phase leads the A phase, and the C phaselags the A phase.See alsophase sequenceindicator.

phase sequence indicator device usedto detect the phase sequence of a 3-phaseelectrical power system. See alsophasesequence.

phase shift a time displacement of a wave-form with respect to another waveform of thesame frequency.

phase shifter a device that changes thephase angle between two buses in a powersystem. Conventional phase shifters are spe-cial autotransformers in with each phase volt-age in connected in series with a variablecomponent of voltage from another phase.


By adjusting the variable component, thephase angle can be changed. Newer phaseshifters are built with power electronic de-vices. Phase shifters are often used in an-tenna arrays.

phase spectrum the phase angle of theFourier transform6 F(ω),−∞ < ω < ∞of a signalf (t). For example, the phasespectrum of a rectangular pulse of unit widthis given in the figure. See alsoFouriertransform.

Phase spectrum.

phase velocity the velocity, at a givenfrequency, and for a given mode, of anequiphase surface in the direction of prop-agation. In other words, it is the velocity atwhich an observer travelling in the propaga-tion direction should move in order to seethe phase unchanged. It can be greater thanthe speed of light, since no transportation ofenergy is actually involved.

phase-controlled converter converter inwhich the power devices are turned off at thenatural crossing of zero voltage in AC to DCconversion applications.

phase-controlled oscillator a voltagecontrolled oscillator whose output frequencyis determined by some type of phase dif-ference detecting circuitry.See alsophaseerror.

phase-locked loop (PLL) (1) a circuit forsynchronizing a variable local oscillator withthe phase of a feedback portion of that oscil-lator so as to operate at a constant phase anglerelative to the reference signal source. Maybe categorized as 1st, 2nd, 3rd, and so on,order loops according to the number of inte-grators in the so-called loop filter. The num-ber of integrators is one less than the order ofthe loop. This loop is used for demodulators,frequency synthesizers, and timing recoverycircuits.

(2) a closed-loop feedback control systemwhere the feedback signal is a phase variationof the signal frequency rather than a voltageor a current.

phase-only filter an optical mask imposesa pattern of phase variations over an imagepassing through it; used frequently in imagecorrelators.

phase-matching conditions in waveg-uides formed by two or more different me-dia, the phase-matching condition states thephase velocity along the propagation direc-tion in the different media should be the same.

phase-shift keying (PSK) an informa-tion encoding method that uses changes inthe phase of the signal carrier.

phase-shifting mask a mask that containsa spatial variation not only in intensity trans-mittance but phase transmittance as well.

phase-to-phase fault a fault with twotransmission lines being short circuited.

phased array Seephased array antenna.

phased array antenna an antenna com-posed of an aperture of individual radiatingelements. Beam scanning is implemented bychanging the phases of the signals at the an-tenna elements with the weights remainingfixed as the beam is steered.


phasor a complex number representingthe amplitude and phase of a sinusoidal func-tion.

phoneme the smallest units in phonemics.Phonemes are produced by different mannersof articulation (e.g., plosives, fricatives, vow-els, liquids, nasals). The automatic speechrecognition on large lexicons is often basedon the recognition of units like phonemes inorder to break the complexity of the globalrecognition process, from speech frames towords and sentences.

phonemics the study of sound units in theframework of descriptive linguistics. Basi-cally, unlike phonetics, the sounds are stud-ied by taking into account the language andnot only observable features in the signal.

phones the smallest units in phonetics,where the emphasis is placed on observable,measurable characteristics of the speech sig-nal.

phonetic knowledge knowledge of theacoustic structure of phones. It is of funda-mental importance for the design of effectiveautomatic speech recognition systems.

phonetics the study of the acoustic soundswhere the emphasis is placed on observable,measurable characteristics of the speech sig-nal.

phonon a quantized packet of energy as-sociated with material lattice vibrations thathave been excited by an incident photon.

phonon maser Seeacoustic laser.

phosphorescence emission of light from along-lived electronically excited state. Phos-phorescent emission is a quantum mechani-cally forbidden transition between electroniclevels of different spin states.

photochemistry chemical reactions thatoccur as a result of electronically exciting atleast one of the reactant molecules with light.

photodetector device capable of produc-ing or modifying an electrical signal in pro-portion to the amount of light falling on theactive area of the device.

photoelastic effect mechanical strain in asolid causes a change in the index of refrac-tion that can affect the phase of a light wavetraveling in the strain medium.

photoelectric effect the phenomenonwhereby light of sufficiently short wave-length falls on the emitter electrode of a pho-tocell and causes electrons to be emitted fromthe electrode.

photogalvanic effect Seephotovoltaiceffect.

photolithography Seeoptical lithography.

photoluminescence the process by whichlight is emitted from solids, atoms, gases, af-ter excitation by an additional light source.The input light excites electrons to higherenergy states, and as they relax they emitlight (through electron-hole recombination)whose frequency is characteristic of the sta-tistical properties of the carriers.

photomask Seemask.

photon a minimum energy quantum oflight energy proportional to the frequency ofthe radiation.

photon echo (1) an optical field emitted bya macroscopic polarization that has been gen-erated by reversing the dephasing process in amaterial having an inhomogeneously broad-ened spectrum.

(2) complex output pulses that are gen-erated when two intense input pulses interact


with the same semiclassically described lasermedium.

photon lifetime Seecavity lifetime.

photon noise fundamental noise due tothe quantum nature of light; a statistical vari-ation in optical intensity due to measure-ments of discrete number of photons.

photopic formally, a description of lumi-nances under which human cone cells are ac-tive. Informally, describing daylight lumi-nances.

photorefractive beam fanning a pho-torefractive phenomenon in which a beam ofcoherent light is scattered into a fanned pat-tern by a photorefractive crystal (e.g., bariumtitanate, lithium niobate). When a laser beampasses through a photorefractive crystal, sig-nificant scattering often occurs. The scat-tered light appears to be asymmetrical withrespect to the beam except for propagationalong the c-axis. For laser beams of moder-ate power, the scattered light appears to de-velop slowly in time and eventually reaches asteady-state scattering pattern. This is knownas photorefractive beam fanning. The beamfanning originates from an initial scatteringdue to crystal imperfections. The initiallyscattered light is amplified due to the physi-cal overlap and the energy coupling betweenthe incident beam and the scattered beam.Beam fanning often occurs in highly efficientphotorefractive media, even if the material isnear-perfect. Photorefractive beam fanningplays an important role in the initiation ofmany phase conjugators and resonators, eventhrough the fanning itself can be a source ofnoise in many experimental measurements.

photorefractive crystal crystalline solidsthat exhibit photorefractive effect. Thephotorefractive effect is observed in manyelectro-optic crystals, including BaTiO3,KNbO3, LiNbO3, Sr1−xBaxxNb2O6 (SBN),Ba2−xSrxK1−yNay Nb5O15 (KNSBN),Bi12SiO20 (BSO), Bi12GeO20 (BGO), GaAs,

InP, CdTe, etc. These crystals are referredto as photorefractive crystals. They are byfar the most efficient medium for the gen-eration of phase conjugation and real-timeholography using relatively low intensity lev-els (e.g., 1 W/cm2). In addition to the effi-cient holographic response, beam couplingknown as two-wave mixing also occurs nat-urally in photorefractive crystals.

photorefractive effect in the broad sense,any optical process in which the refractive in-dex of a material system is modified by an op-tical field. More precisely, a particular non-linear optical process in which loosely boundcharges are redistributed within a crystallinematerial in response to an incident light field,thereby producing a static electric field, andmodifying the refractive index of the materialby the linear electro-optic effect. Changes inrefractive index upon exposure to a light pat-tern in certain materials such as BSO, lithiumniobate, and barium titanate. A manifesta-tion of the electro-optic effect on a micro-scopic scale due to spatial charge transport.

photorefractive grating a refractive in-dex grating often produced by using the illu-mination of two coherent laser beams, havingsufficient photon energy, in photorefractivecrystals.

photoresist a photosensitive material thatforms a three-dimensional relief image by ex-posure to light and allows the transfer of theimage into the underlying substrate (for ex-ample, by resisting an etch).

photoresist contrast a measure of the re-solving power of a photoresist, the photore-sist contrast is defined in one of two ways.The measured contrast is the slope of the stan-dard H-D curve as the thickness of resist ap-proaches zero. The theoretical contrast is themaximum slope of a plot of log-developmentrate versus log-exposure energy (the theoret-ical H-D curve). The photoresist contrast isusually given the symbolg.


photovoltaic effect a photoelectric phe-nomenon in certain photorefractive crystals,for example, LiNbO3, BaTiO3, LiTaO3, inwhich the illumination of light leads to thegeneration of electric current along certaindirection in the crystals. This leads to the ac-cumulation of charges on the surfaces of thecrystals, causing an open circuit voltage.

photovoltaics conversion of insolationinto DC electricity by means of solid statep-n junction diodes.

PHPS Seepersonal handy phone system.

PHS Seepersonal handy phone system.

physical address Seereal address.

physical medium the communicationchannel over which signals are transmitted.Broadcast media, in which all stations re-ceive each transmission, are primarily usedin local-area networks. Common media areoptical fibers, coaxial cable, twisted copper-wire pairs, and airwaves.

physical model a mathematical modelbased on device physics. Physical modelsgenerally have a mid-range to high modelingvaluation coefficient.

physical optics a high-frequency tech-nique to approximate electromagnetic scat-tering from an object by representing thesmooth, electrically large parts of the objectby equivalent currents at the object’s surface.

physical page number the page-frameaddress (in main memory) of a page fromvirtual memory.

physical placement of logic (PPL) a de-sign entry method between full custom andstandard cell. It begins at the block diagramlevel where the detailed block specificationand the corresponding layout are done simul-taneously. PPL is targeted at the occasionaldesigner.

physical sensor an interface device at theinput of an instrumentation system that quan-titatively measures a physical quantity suchas pressure or temperature.

physical theory of diffraction a correc-tion to physical optics that includes diffractedfields due to edges and corners.

physical vapor deposition a process inwhich a conductive or insulating film is de-posited on a wafer surface without the as-sistance of chemical reaction. Examples arevacuum evaporation and sputtering.

pickup current the specified value that, ifexceeded, causes the relay to act on its con-tact and cause a circuit breaker action. It isthe threshold current for system protection,and a magnitude above this is considered afault or abnormal condition.

pickup voltage Seepickup current.

picocell cell with radius of a few meters.See alsocell.

picture description language a languagein which parts of scenes are labelled andtheir relative positions are described in a spe-cial symbolic form. Typical relationshipsbetween objects are “inside,” “adjacent to,”“underneath,” and so on. Such a languagecan be parsed and interpreted symbolicallyto build up a meaningful understanding ofthe picture.

picture–carrier the carrier frequency of1.25 MHz above the lower frequency limitof a standard NTSC television signal (lumi-nance carrier in color TV).

PID control Seeproportional-integral-derivative control.

piecewise deterministic process (PDP) aclass of Markov stochastic processes that fol-low deterministic dynamics between randomjumps. A generator of piecewise determin-


istic process (PDP) is an integro-differentialoperator. Roughly speaking PDPs may becharacterized by three functions:

1. deterministic dynamics between jumpsdriven by a differential equation,

2. jump intensity,

3. jump magnitude probabilistic distribu-tion.

PDPs can be used to describe some queue-ing, storage, and renewal processes. More-over, they may also be applied in modelingcontrol processes in the presence of failures,abrupt changes of working conditions, jump-ing deterioration of quality of working plants,as well as inventory and maintenance sys-tems. The important class of PDPs is pro-duced by linear systems with Markov jumps.

Pierce gun an electron gun with whichthe focusing electrode is tilted 67.5◦ from theaxis of the electron beam and the acceleratoranode is also tilted forward to produce a rec-tilinear flow of an electron beam.

Pierce oscillator usually, an FET crys-tal oscillator where the crystal is connectedbetween gate and drain. The crystal is usedlike an inductor, the source of the FET is ACgrounded, and the gate-source capacitance ofthe device and drain-source capacitance areused as parts of the LC-oscillator similar toColpitts oscillator.

piezoelectric pertaining to a material thatpossesses a noncentrosymetric crystal struc-ture that will generate charge on the applica-tion of a mechanical stress. As in the case ofa pyroelectric materials, this can be detectedas either a potential difference or as a chargeflowing in an external circuit.

piezoelectric transducer device that con-verts electric signals to ultrasonic waves, andvice versa, by means of the piezoelectric ef-fect in solids.

pig slang for line hoseSeeline hose.

pig tail a type of hot stick that can slideover an overhead conductor.

piggyback board Seedaughter board.

pigtail short electrical conductor used toconnect the brushes of an electrical motor orgenerator to the external electrical connec-tions of the machine.

pilot carrier a means of providing a car-rier at the receiver, which matches the re-ceived carrier in phase and frequency. In thismethod, which is employed in suppressedcarrier modulation systems, a carrier of verysmall amplitude is inserted into the modu-lated signal prior to transmission, extractedand amplified in the receiver, and then em-ployed as a matching carrier for coherent de-tection.

pilot sequence a spreading code sequenceused in a CDMA cellular network such as IS-95 to facilitate the PN code synchronizationand coherent demodulation of the receivedsignal. In the most typical case, the pilotsequence is utilized in the forward channel,i.e., transmission by the base station.

pin the electronic connection that allowsconnection between an integrated circuit orcircuit board and some socket into which itis plugged.

PIN diode a diode with a large intrinsicregion sandwiched between p- and n-dopedsemiconducting regions.

pin insulator an electric insulator whichis concentric with a hollow, threaded holeso that it can be screwed onto a steel pinmounted on a utility pole or crossarm.

pinch effect the collapse of a hollow con-ductor due to the magnetic effects of verylarge currents. Sometimes observed in cableshields which have been struck by lightning.


pinch resistance the resistance of a fullydepleted channel of a junction field effecttransistor (JFET).

pinch-off region Seeohmic region.

pinch-off voltage the gate-to-source volt-age at which the channel current is reducedto a very small predetermined level specifiedin milliamperes per millimeter. The effectis caused by carrier depletion of the channelarea.

pincushion distortion the geometric dis-tortion present in an image in which both hor-izontal and vertical lines appear to collapsetoward the display center. For a CRT im-age system, pincushion distortion is a resultof the interaction of the long face plate ra-dius of curvature and the short radius fromthe deflection center to the face plate. Theinteraction causes the top, bottom, left, andright raster center points to be closer to thecenter of the CRT face plate deflection thanthe top left, top right, bottom left, and bottomright raster corners. Consequently, a squaregrid has the appearance of a pin cushion.

pipe cable a paper-insulated high-voltageelectric power transmission cable laid withina rigid steel pipe containing pressurized in-sulating oil.

pipe flush systems that fetch streams ofdata or instructions sometimes have to inter-rupt the stream when an unusual event occurs.When this happens, the pipeline containingthe stream of instructions or data must beemptied before execution can continue; thisis called flushing the pipe.

pipeline chaining a design approach usedin computers whereby the output stream ofone arithmetic pipeline is fed directly intoanother arithmetic pipeline. Used in vectorcomputers to improve their performance.

pipeline interlock a hardware mecha-nism to prevent instructions from proceeding

through a pipeline when a data dependencyor other conflict exists.

pipeline processor a processor that exe-cutes more than one instruction at a time, inpipelined fashion.

pipelined bus Seesplit transaction.

pipelined cache a cache memory with alatency of several clock cycles that supportsone new access every cycle. A new accesscan be started even before finishing a previ-ous one. The access to the cache is dividedinto several stages whose operation can beoverlapped. For instance, the cache can bepipelined to speed up write accesses: tagsand data are stored in independently address-able modules so that the next tag comparisoncan be overlapped with the current write ac-cess. Read accesses are performed in a singlecycle (tag and data read at the same time).

pipelining a technique to increase through-put. A long task is divided into components,and each component is distributed to one pro-cessor. A new task can begin even though theformer tasks have not been completed. In thepipelined operation, different components ofdifferent tasks are executed at the same timeby different processors. Pipelining leads toan increase in the system latency, i.e., the timeelapsed between the starting of a task and thecompletion of the task.

pitch commonly used by physicists andmusicians, defined with reference to the fre-quency. Given two signals with frequenciesf1 andf2, the difference in pitch is definedby 1200 log2

f2f1

. See alsocoil pitch.

pitch angle an angle between a tangent toa helix and another tangent to a cylinder thatcontains the helix and is perpendicular to thecylinder axis at a common tangential pointon the helix.


pitch factor in an electric machine, the ra-tio of the fractional pitch in electrical degreesto the full pitch, also in electrical degrees.

pivoting when applying Gaussian elim-ination to solve a set of simultaneous linearequations, the natural solution order is some-times varied. The process of varying the nat-ural solution order is termed pivoting. Piv-oting is used to avoid fill-in and to maintainthe accuracy of a solution.

pixel contraction of “picture element:”each sample of a digital image, i.e., asquare or rectangular area of size1x ×1y of constant intensity, located at position(k1x, l1y) of the image plane. Also calledpel.

pixel adjacency the property of pixelsbeing next to each other. The adjacencyof pixels is ambiguous and is defined sev-eral ways. Pixels with four-adjacency orfour-connected pixels share an edge. Pixelswith eight-adjacency or eight-connected pix-els share an edge or a corner.See alsochaincode, connectivity, pixel.

pixel density a parameter that specifieshow closely the pixel elements are spaced ona given display, usually a color display.

PLA Seeprogrammable logic array.

placement a placement routine deter-mines an optimal position on the chip for a setof cells in a way that the total occupied areaand the total estimated length of connectionsare minimized.

planar array in addition to placing ele-ments along a line to form a linear array, onecan position them on a plane to form a pla-nar array. For example, a circular array is aspecial form of planar array, where the ele-ments are placed along a circle that is usuallylocated on a horizontal plane. Planar arraysprovide additional variables that can be usedto control and shape the array’s beam pattern.

The main beam of the array can be steered to-wards any point in its half space.

planar doped barrier a material growthtechnique that allows control of doping on asingle atomic layer, rather than the more con-ventional doping over a thicker region. Thistechnique is used for bandstructure control inheterojunction transistors and is also used inthe fabrication of planar doped barrier detec-tor diodes.

planar magnetron a permanent magnetarrangement consisting of a steel base onwhich an outer ring of one polarity surroundsan inner ring or island of the opposite polarity,creating a magnetic flux “tunnel” for trappingelectrons in the sputtering process.

planar mirror an optically flat mirror,the flatness generally specified as a fractionalnumber of optical wavelengths.

planar–optic interconnect an intercon-nect in which optical signals travel alongzigzag path in a transparent substrate witha source and a detector array, and other op-tical elements such as mirrors, gratings, andholograms. All optical elements on the sub-strate can be fabricated by lithographic tech-niques. The merits are compact packaging,easy alignments, large heat removal capac-ity, and suitability for integration with opto-electronic devices.

plane earth loss the propagation loss ex-perienced when isotropic transmit and re-ceive antennas are operated in the vicinity ofa flat reflecting plane. Frequently used as anapproximation for propagation over Earth’ssurface.

plane wave (1) an electromagnetic wavein which each wavefront (surface of constantphase) forms a plane of infinite extent andpropagates in a direction perpendicular to thewavefront. A uniform plane wave has thesame amplitude over an entire wavefront; a


nonuniform plane wave has varying ampli-tude.

(2) a wave having as equiphase surfacea plane. It is commonly represented by thefunctional dependence expj (ωt−k·r ) whereωis the angular frequency,t is the time,r isthe position vector, andk is the wave vec-tor. Sometimes monochromaticity is also im-plied.

plane-parallel resonator a laser res-onator in which the mirrors are flat and par-allel to each other.

planned outage an interruption in serviceto consumers which has been caused by a se-quence of events which were pre-planned bythe utility, for example to perform mainte-nance or construction.

plasma the fourth state of matter com-prised of positive ions and negative elec-trons of equal and sufficiently high densityto nearly cancel out any applied electric field.Not to be confused with blood plasma.

plasma frequency a critical frequencyin a plasma (e.g., ionosphere) below whichradio waves cannot propagate through theplasma. Usually, in the ionosphere for anf > fp the wave gets refracted towardsEarth.

plasmon a polariton in a plasma medium.

plastic a term for a flexible roll-on poly-mer insulation layer.

plate the positive electrode or anode in avacuum tube.

platinotron a strap feed starting radial M-type backward wave oscillator.

platter Seedisk platter.

plausibility function a function that givesa measure of plausibility or belief.

PLC Seeprogrammable logic controller,power line conditioner.

PLD Seeprogrammable logic device.

PLL Seephase-locked loop.

plosives sounds produced when the vocaltract, completely shut off at a certain point,is subsequently reopened so as a small ex-plosion occurs. The “silence” that precedesthe explosion can either be voiced or voice-less. In the first case we have voiced plosives(/b/, /d/, /g/), while in the second case wehave voiceless plosives (/p/, /t/, /k/).

plossl an eyepiece consisting of four lenselements, a pair of cemented doublets facingeach other in a symmetrical configuration.

Plotkin’s upper bound for any (n, k)linear block code, the minimum distance isbounded asymptotically as, forn → ∞

(1 − k

n

)≥ 2

d

n

plug fuse a small fuse with a threaded basedesigned to be installed in a mated screw-typereceptacle. Plug fuses are rated 125 V and aretypically applied on branch circuits in ratingsof 15 A, 20 A, and 30 A. Type S plug fusesare designed to prevent overfusing.

plug-in board a printed-circuit board thatcan be plugged into another board or moduleto provide some functionality.

plugging a procedure to bring a three-phase motor to an abrupt stop by reversingthe direction of the rotating magnetic field inthe airgap. The reversal is accomplished byreversing two of the phase connections to themotor.

Plumbicon tube a type of camera tubefirst developed in the 1960s by Philips.


PLZT See lead lanthanum zirconiumtitanate.

PM Seepermanent magnet.

PMA Seepost-metal anneal.

PMA system Seeprism | air | metalsystem.

PMN generic name for electrostrictivematerials of the lead (Pb) magnesium niobatefamily.

PMOS acronym for P-channel MOSFET.

PMU Seepower monitoring unit.

PN Seepseudo-noise.

PN code Seepseudo-noisecode tracking.

PN code tracking Seepseudo-noise codetracking.

PN sequence pseudo-noise sequence.Seespreading sequence.

Pockels coefficients denotes the strengthof the linear effect in electro-optic materials.Named after the pioneer researcher F. Pock-els.

Pockels effect change in refractive indexin noncentrosymmetric materials upon appli-cation of an electric field. Same as the linearelectro-optic effect. Named after the pioneerresearcher F. Pockels.See alsoelectro-opticeffect.

Pockels readout optical modulator(PROM) an optically addressed spatiallight modulator device using the electro-opticeffect in BSO.

Poincare sphere a conceptual aid inwhich the state of polarization is representedby a point on a sphere. For example,points along the equator correspond to linearpolarization (horizontal, vertical, or slant),

the poles correspond to circular polarization(right-hand for upper pole and left-hand forlower pole), and all other points correspondto elliptical polarization.

point feature a small feature that can beregarded as centered at a point, so that inter-feature distances can be measured, as part ofa process of inspection, or prior to a processof inference of the presence of an object fromits features. Commonly used point featuresinclude corners and small holes, or fiducialmarks (e.g., on printed circuit boards).Seealsosalient feature, key point detection.

point operation an image processing op-eration in which individual pixels are mappedto new values irrespective of the values of anyneighboring pixels.

point process a type of image processingin which the enhancement at any pixel de-pends only on the gray level at that pixel. Thisis in contrast to an operation whose value de-pends on the location of the pixel or on thevalues of neighboring pixels.See alsograylevel, neighborhood operation.

point source a light source so small that itssize and shape cannot be determined from thecharacteristics of the light emanating from it.The light emitted has a spherical wave frontand is spatially coherent.

point spread function (PSF) multidi-mensional impulse response. Output of amultidimensional system when a point func-tion (δ−function) is input.

point-to-point bus a bus that connects andprovides communication capability betweentwo, and only two, components. It shouldbe noted that buses support communicationswithin the CPU, as well as external to theCPU. Most internal buses are point-to-pointbuses. In contrast, see also multipoint bus.

point-to-point motion in the point-to-point motion, the manipulator has to move


from an initial to a final joint configurationin a given timetf .

pointer in programming languages, avariable that contains the address of anothervariable.

poison any material or process which ab-sorbs neutrons and thus dampens a nuclearfission reaction, e.g., control rods.

Poisson counting process SeePoissonprocess.

Poisson distribution a probability distri-bution widely used in system modeling. Anon-negative integer-valued random variableX is Poisson-distributed ifProb(X = k) =e−aak/k!, wherea is a positive parametersometimes called the intensity of the distri-bution. For example, the number of ‘purely’random events occurring over a time intervalof t often follows a Poisson distribution withparametera proportional tot .

Poisson process a random point processdenoting the occurrence of a sequence ofevents at discrete points in time. The timedifference between the different events is arandom variable. For a given time interval oflengthT , the number of events (points in theprocess) is a random variable with Poissondistribution given by the following probabil-

ity law P [N = k] = (λT )ke−λTk! , whereN is

the number of events that occur in the intervalof lengthT , andλ is the expected number ofoccurrences of events per unit time. The timeinterval between any two events is a randomvariable with exponential distribution givenby F(τ) = 1 − e−λτ , τ > 0.

Poisson’s equation a partial differentialequation that expresses the relation betweenthe scalar potentialV and the charge densityat any pointρ. Mathematically described by∇2V = ρ

ε, where∇2 is the Laplacian,ρ

is the forcing function,V is the equation’ssolution, andε is the dielectric constant ofthe medium.

polariton term originally used to desig-nate the polarization field particles analogousto photons. It is currently used to denote thecoupling of the electromagnetic field with thepolar excitations in solids.

polarity the notation used in the assign-ment of voltages. In DC generators, the po-larity of the armature voltage can be reversedby either reversing its field current or by ro-tating the generator in reverse direction.

polarization (1) the shape traced out bythe tip of the electric field vector as a functionof time and the sense in which it is shaped.

(2) a description of the form of the tem-poral variation of the electric field vector ofa light field. In general, the polarization statecan be described by the ellipse that tip of theelectric field vector traces each optical cycle.Commonly encountered limiting forms arelinear polarization and circular polarization.

(3) the response of material systems, anapplied light field by developing a time vary-ing dipole moment. The response is de-scribed quantitatively in terms of the dipolemoment per unit volume, which is known asthe polarization vector.

(4) description of the direction and motionof the electric field vector of a wave. Planewaves may be linearly or elliptically (includ-ing circularly) polarized.

polarization controller a device that al-ters only the polarization state of the incidentlight.

polarization distortion distortion in thetemporal shape of an optical signal transmit-ted through an optical fiber caused by the dif-ferential time delay between the two polar-izations of each propagating mode. Usuallyof significance only in a single mode fiber.

polarization diversity the use of at leasttwo antennas with different polarization char-acteristics in a depolarizing medium so as toproduce communication channels with sub-stantially uncorrelated fading characteristics.


polarization division multiple access(PDMA) a multiple access technique whereuser channels are separated in the polar-ization domain of the transmitted signal(horizontal–vertical for linear polarizationand left–right for circular polarization).

polarization ellipse the most generalform of polarization specification in whichthe polarization of an electromagnetic waveis completely specified by the orientation andaxial ratio of the polarization ellipse and bythe sense of rotation (right-hand or left-hand).

polarization logic gate an optical logicgate using polarization modulating devices.Polarization modulating devices includeMOSLM (magneto-optic spatial light mod-ulator), LCTV (liquid crystal television),LCLV (liquid crystal light valve), MSLM(microchannel-plate spatial light modulator).These devices are very useful in performingXOR and XNOR operations. The basic op-eration is as follows. Two devices representinputs A and B, respectively. When a beamof light passes through the first device, its po-larization is rotated according to input A. Thebeam then passes through the second deviceresulting that its polarization is either rotatedfurther away or rotated back according to in-put B. To perform other logic operations, anadditional thresholding device is necessarywhether it is based on optical addition or mul-tiplication.

polarization loss also called polarizationmismatch. It occurs when the polarization ofthe incident wave is different than the polar-ization of the receiving antenna.

polarization maintaining fiber a “singlemode” fiber that supports two (linearly po-larized) orthogonal modes.

polarization mismatch Seepolarizationloss.

polarization vector that part of the fluxdensity attributable to orientation of bound

charges inside a dielectric by an applied elec-tric field. It is also the electric dipole momentper unit volume inside the dielectric and isequal to the product of the electric suscepti-bility and the applied electric field intensity.

polarization-sensitive device device thatexhibits behavior dependent on the polariza-tion of the incident electromagnetic wave.A polarizing filter exhibits transmission asa function of the polarization of the light in-cident on it.

polarized capacitor an electrolytic ca-pacitor in which the dielectric film is formedon only one metal electrode. The impedanceto the flow of current is then greater in onedirection than in the other. Reversed polaritycan damage the part if excessive current flowoccurs.

pole (1) the roots of the denominatorD(s) = 0, for irreducible rational func-tion X(s) = N(s)

D(s)at which pointsX(s) be-

come unbounded. Rational functions im-portant to signal processing applications in-clude Laplace transforms; the pole locationsof Laplace transforms provide valuable in-formation on system stability and other be-haviors.

(2) the values of a complex function whichcause the value of the function to equal in-finity (positive or negative). The poles areall natural frequencies of vibration, or reso-nances of the circuit described by the equa-tion, but occur at infinite (finite if loss ispresent) attenuation. They are properties ofthe function itself, and are not influenced byany other elements in the system (immune toload pulling).

(3) one end of a magnet or electromagnetin electrical machines, created by the flux ofthe machine.

pole line any power line which is carriedoverhead on utility poles.


pole of 2-D transfer matrix a pair of com-plex numbers(p1, p2)

T (z1, z2) = N(z1, z2)

d(z1, z2)

N(z1, z2) ∈ Rp×m [z1, z2]

that are the root of the 2-D polynomiald(z1, z2), i.e., d(p1, p2) = 0, whereRp×m [z1, z2] is the set ofp×m polynomialmatrices inz1 andz2 with real coefficients.

pole pitch the angular distance (normallyin electrical degrees) between the axes of twopoles in an electrical machine.

pole top pin a steel pin onto which a pininsulator is screwed.

pole-coefficient sensitivity when a coef-ficient dk of the polynomialD(s) = d0 +d1s+ d2s

2 + . . . is a variable parameter thenthe roots of the polynomial are functions ofthis parameter. If these roots are simple, onecan use the pole-coefficient sensitivity

Sdk (pi) = ∂pi

∂dk/dk= −dk pki

D′(pi)The pole-coefficient sensitivity is frequentlyused in evaluation of active circuit stability.

pole-top transformer generally a distri-bution transformer which is mounted atop autility pole near the customer.

pole-zero plot a graphical representationof a Laplace transform is known as a pole-zero plot. Except for a scale factor, the nu-merator and denominator polynomials in arational Laplace transformX(s) = N(s)

D(s)can

be specified by their roots. Thus, marking thelocation of the roots ofN(s) andD(s) in thes-plane provides a powerful way of visualiz-ing a Laplace transform. Pole-zero plots areused extensively in signal processing designand analysis, e.g., to determine the stabilityof a system function.

polled interrupt a mechanism in whichthe CPU identifies an interrupting device

by polling each device.See alsovectoredinterrupt.

polling sequencing through a group of pe-ripheral devices and checking the status ofeach. This is typically done to determinewhich device(s) are ready to transfer data.

polychlorinated biphenyls chemicalcompounds added to insulating oils to im-prove stability and insulating capability.

polygon detection the detection of poly-gon shapes, often from corner signals or fromstraight edges present in an image. Poly-gon detection is important when locating ma-chined parts in images, e.g., prior to robot as-sembly tasks.See alsorectangle detection,square detection.

polygonalization a method of represent-ing a contour by a set of connected straightline segments; for closed curves, these seg-ments form a polygon.

polymer membrane type of structureused to mechanically modulate light.

polynomial warp polynomial warping isa type of commonly used image processingoperation designed to modify an image ge-ometrically. This type of operation is use-ful when an image is subject to some un-known physical spatial distortion, and thenmeasured over a rectangular array. The ob-jective is then to perform a spatial correctionwarp to produce a corrected image array.

polyphase filter bank a filter bank wherethe filters are implemented in terms of theirpolyphase representation. The decimatorscan be moved ahead of the polyphase filters inthe analysis section, thereby reducing com-putation.

polyphase representation a sequencerepresentation. A sequenceH(z) can be rep-resented by the summation ofM sequences


Ek(z) orRk(z) as

H(z) =M−1∑k=0

z−kEk(zM) Type I polyphase.

or

H(z) =M−1∑k=0

z−(M−1−k)Rk(zM)

Type II polyphase.

polyphase system electrical system thathas more than one phase, which are separatedby angles of 360◦/n, wheren is the numberof phases. For example, three phase systemsare polyphase systems where the three phasesare separated by 120 electrical degrees fromeach other. A six-phase system is a polyphasesystem where each successive phase is sepa-rated by 60 electrical degrees from the other.

polysalicide a variation in the formationof polysilicide where the metal is depositedafter the polysilicon and reacted with the sil-icon during a subsequent thermal annealingcycle.

polysilicon a polycrystalline or amor-phous form of silicon deposited on the sur-face of a wafer during integrated circuit fabri-cation. In modern technologies, it most oftenforms the MOSFET gate, the bipolar emitter,the plates of a capacitor or a resistor.

pop instruction an instruction that re-trieves contents from the top of the stack andplaces the contents in a specified register.

popcorning a plastic package crack or de-lamination that is caused by the phase changeand expansion of internally condensed mois-ture in the package during reflow soldering,which results in stress the plastic packagecannot withstand.

population dynamics a variety of mod-els used to describe evolution, growth, ki-netics, and dynamics of diverse popula-tions. Population dynamics models may

be stochastic or deterministic, discrete orcontinuous, differential or integral, andcover a number of mathematical tools fromordinary differential and difference equa-tions, through partial differential, integro-differential, functional and integral equationsto particle systems, cellular automata, neuralnetworks and genetic algorithms. The type ofthe model depends on the type of population,the objective of modeling, available data,knowledge of phenomena, etc. The mostoften modeled and analyzed populations in-clude human and animal populations for de-mographic and epidemiologic purposes, cellpopulations including cancer, blood, bonemarrow, eukaryotic cells, virus, bacteria,fungi, genomes, and biomolecules. Con-trol problems in population dynamics maybe formulated in terms of optimal treatmentprotocols (for example cell-cycle-specificcontrol), vaccination, harvesting strategies,modulation of growth and so on. The sim-plest models are found by clustering dis-tributed in reality systems into lumped com-partments. It leads to compartmental mod-els of population dynamics. Linear modelscould be obtain under hypotheses of Malthu-sian (exponential) growth of the population.If, however, such a model is used to describethe evolution of the population under con-trol (for example treatment by drugs, vacci-nation, harvesting), the model is no longerlinear, and the simplest class of control mod-els that could be used is given by bilinear con-trol systems. Since real populations nevergrow unboundedly, more realistic models aregiven by nonlinear models with saturationeffects. The simplest nonlinear differentialmodels represent logistic, Pearl–Verlhurst,Michaelis–Menton, and Gompertz dynam-ics.

population inversion usually the densityof atoms or molecules in the higher state of alaser transition minus the density in the lowerstate.


population of energy level number ordensity of atoms or molecules in an energystate.

population trapping optical pumpinginto a noninteracting state or states of a quan-tum mechanical system.

porous silicon noncrystalline silicongrown so as to produce porous, sponge-likestructures that seem to enhance optical emis-sion.

port (1) a terminal pair.(2) a place of connection between one

electronic device and another.(3) a point in a computer system where

external devices can be connected.

port protection device device in line withmodem that intercepts computer communi-cation attempts and requires further authen-tication.

portability the possibility to move soft-ware to another hardware/operating systemenvironment without changes.

position error the final steady differencebetween a step function setpoint and the pro-cess output in a unity feedback control sys-tem. Thus it is the asymptotic error in posi-tion that arises in a closed loop system thatis commanded to maintain a constant posi-tion. See alsoacceleration errorandvelocityerror.

position error constant a gainKp fromwhich the position errorep is readily deter-mined. It is a concept that is useful in the de-sign of unity feedback control systems, sinceit transforms a constraint on the final errorto a constraint on the gain of the open loopsystem. The relevant equations are

ep = 1

1 +Kpand Kp = lim

s → ∞ q(s)

where q(s) is the transfer function modelof the open loop system, including the con-

troller and the process in cascade.See alsoacceleration error constantandvelocity errorconstant.

position sensor a device used to detectthe position of the rotor with respect to thestator. The most commonly used positionsensors for electric motors are Hall effect de-vices, encoders, and resolvers with resolver-to-digital converters.

position servo a servo where mechani-cal shaft position is the controlled parameter.See alsoservo.

positioner a mechanical device used tomove an antenna or target to a desired posi-tion for measurement purposes. Positionerscan be single or multi-axis and are usuallycontrolled by computers and automated mea-surement equipment.

positive 2-D Roesser model the 2-DRoesser model[xhi+1,jxvi,j+1

]=[A1 A2A3 A4

][xhijxvij

]+[B1B2

]uij

yij = C

[xhijxvij

]+Duij

i, j ∈ Z+ (the set of nonnegative integers) iscalled positive if for all boundary conditionsxh0j ∈ Rn1+ , j ∈ Z+, andxvi0 ∈ Rn2+ , i ∈ Z+and all inputs sequencesuij ∈ Rm+, i, j ∈Z+ we havexhij ∈ R

n1+ , xvij ∈ Rn2+ for

i, j ∈ Z+ andyij ∈ Rp+ for i, j ∈ Rp+ whereRn+ is the set ofn-dimensional vectors withnonnegative components andxhij andxvij arehorizontal and vertical state vectors, respec-tively. The 2-D Roesser model is positive ifand only if all entries of the matricesA1,A2,A3, A4, B1, B2, C, andD are nonnegative.

positive definite function a scalar func-tionV (x)with continuous partial derivativeswith respect to the components of the vectorx where

1.V (0) = 0.


2.V (x) > 0 wheneverx 6= 0.

positive definite matrix a symmetric ma-trix A such thatxT Ax > 0 for any vectorxnot identically zero. The eigenvalues of apositive definite matrix are all strictly greaterthan zero.

positive photoresist a photoresist whosechemical structure allows for the areas thatare exposed to light to develop at a faster ratethan those areas not exposed to light.

positive real (PR) function a rationalfunctionH(s) of the complex variables =σ + ω is said to be positive real (PR) if itsatisfies the following properties:

1.H(s) is a real number whenevers is areal number, and

2. Re[H(s)] ≥ 0 wheneverRe[s] > 0,whereRe[·] represents the real part of[·].

positive semidefinite a scalar functionV (x, t) with continuous partial derivativeswith respect to all of its arguments is saidto be positive semi-definite if

1.V (0, t) = 0.2.V (x, t) ≥ 0 wheneverx 6= 0.

positive semi-definite matrix a symmet-ric matrixA such thatxT Ax >= 0 for anyvectorx. The eigenvalues of a positive semi-definite matrix are all greater than or equal tozero.

positive sequence the set of balanced nor-mal (abc) sequence components used in sym-metrical components. Balanced load cur-rents, for example, are strictly positive se-quence.

positive transition angle the angular por-tion of the time-based output signal (in de-grees) that has a positive slope. This quantitycould be loosely interpreted as the “leadingedge” angle.

positive-sequence reactance the induc-tive reactance offered by a circuit to the flow

of positive-sequence currents alone. Thepositive-sequence reactance is a function ofthe operating frequency of the circuit and theinductance of the circuit to positive-sequencecurrents.

positivity a systemH : X → X where

〈x,Hx〉 →≥ 0 ∀x ∈ X

See alsoinner product space, and passivity.

positron emission tomography (PET)(1) a form of tomographic medical imagingbased upon the density of positron-emittingradionuclides in an object.

(2) an imaging modality that uses in-jected positron-emitting isotopes as markersfor physiological activity. The isotopes emitpairs of gamma photons which are detectedusing a gamma camera and coincidence de-tector.

possibility theory a theory evolved fromthe concepts of fuzzy sets and approximatereasoning which provides a mathematicalframework for possibility analysis.

POST Seepower on self-test.

post insulator an electrical insulatorwhich is supported by its firmly-bolted base,either in an upright position or cantileveredout horizontally from a utility tower.

post-apply bake Seeprebake.

post-compensator a compensator posi-tioned on the process output signal. ForMIMO systems, a given compensator willhave a different effect depending on whetherit is positioned before or after the process.Hence the importance of the prefix “post.”See alsocompensation, compensator, andpre-compensator.

post-exposure bake (PEB) the process ofheating the wafer immediately after exposurein order to stimulate diffusion of the products


of exposure and reduce the effects of stand-ing waves. For a chemically amplified resist,this bake also causes a catalyzed chemicalreaction.

post-metal anneal (PMA) a process usedin semiconductor processing after metalliza-tion to improve device properties and circuitperformance. Typically it is performed at amodest temperature (300–450 C◦) in a form-ing gas (e.g., 5% hydrogen with 95% nitro-gen) ambient.

postbake Seehard bake.

posterior statistics the empirical statisticsof a random quantity (scalar, vector, process,etc.), based on the a priori statistics supple-mented with experimental or measured ob-servations.See alsoprior statistics, Bayes’rule.

postfix Seepostfix notation.

postfix notation a notational or program-ming scheme in which both operands of atwo-operand operation are written before theoperator is specified. Example:ab+ is thepostfix representation of a sum; this could beimplemented in a programming model basedupon an evaluation stack by the operation se-quence PUSH a, PUSH b, ADD. Postfix no-tation is used in programming zero-addresscomputers.

postincrementation an assembly lan-guage addressing mode in which the addressis incremented after accessing the memoryvalue. Used to access elements of arrays inmemory.

pot head a fork-shaped transition betweena three-phase buried cable and an overheadthree-phase electric power line.

potassium dihydrogen phosphate (KDP)a strong linear electro-optic material, be-

longing to the same crystal class as am-

monium dihydrogen phosphate (ADP). Itschemical form isKH2PO4.

potential an auxiliary scalar or vectorfield that mathematically simplifies the solu-tion process associated with vector boundaryvalue problems.See alsoHertzian potential.

potential coil a long, finely wound,straight coil, similar in operation to a Chat-tuck coil, that is used with a fluxmeter to mea-sure magnetic potential difference between apoint in a magnetic field and a flux-free pointin space.

potential source rectifier exciter a sourceof energy for the field winding of a syn-chronous machine obtained from a rectifiedstationary AC potential source. The AC po-tential can be obtained from the machinephase voltages, or from an auxiliary source.The components of the exciter are the po-tential source transformer and the rectifiers(including possible gate-circuitry).

potential transformer (PT) a devicewhich measures the instantaneous voltage ofan electric power line and transmits a signalproportional to this voltage to the system’sinstruments.Seevoltage transformer.

potentially unstable active circuit or de-vice, where particular impedance termina-tion placed on its input and/or output portscauses it to become unstable and oscillates.

potentiometric sensor a chemical sen-sor that measures the concentration of a sub-stance by determining the electrical poten-tial between a specially prepared surface anda solution containing substance being mea-sured.

Potier reactance the leakage reactanceobtained in a particular manner from a teston a synchronous machine at full load witha power factor of zero lagging. The test re-quires little power but supplies the excitationfor short circuit and for normal rated volt-


age both at full-load current at zero powerfactor. The Potier reactance is determinedby a graphical manner from the open circuitcharacteristic and the short circuit point forfull-load current.

power (1) a measurable quantity that is thetime rate of increase or decrease in energy.Units are in watts.

(2) ratio of energy transferred from oneform to another (i.e., heat, radio waves, light,etc.) to the time required for the transfer,expressed in watts.

power added efficiency dimensionlessratio of RF power delivered from a device tothe load minus the input incident RF powerversus the total DC power dissipated in thedevice.

power amplifier amplifier operating ateither audio or radio frequency range deliv-ering power to a termination such as speaker,antenna, or resistive load.

power angle the angular displacement ofthe rotor from the stator rotating magneticfield while the machine is on load. The powerangle is also the angle between the terminalvoltageVt of a synchronous machine and thegenerated voltageEg orEm, respectively, fora generator or motor. This angle denoted byδ is also referenced as power angle or torqueangle or the load angle in a synchronous ma-chine. It signifies the limits of the machineto remain in synchronism.See alsotorqueangle.

power angle curve a curve shown the re-lationship between the active power outputof a generator and its power angle.

power broadening increase in linewidthof a homogeneously broadened transition dueto stimulated emission.

power conditioner a device designed tosuppress some or all electrical disturbancesincluding overvoltages, undervoltages, volt-

age spikes, harmonics, and electromagneticinterference (EMI). Example power condi-tioners are active filters for the reductionof harmonics, metal-oxide varistors (MOVs)and isolation transformers for the protectionagainst voltage spikes, and EMI filters.

power control (1) the use of a mechanismto adjust a transmitter’s power, usually in or-der to improve the overall performance of acommunications network.

(2) in CDMA, a technique to increase theradio capacity. This is due to the fact thata CDMA system is interference-limited andthat all users in a cell operate at the samefrequency. The power control scheme usedat the forward link of each cell can reduce theinterference to the other adjacent cells. Theless the interference generated in a cell, thehigher the capacity.

(3) in a TDMA system, used to reduce co-channel interference or interference in adja-cent cells. By having all cells (which oper-ate at different frequencies with a cell reusefactor of 7 for a TDMA system) at approx-imately the same power, interference is re-duced.

power density generally refers to the av-erage power density, which is a measure ofthe power per unit area of a propagating EMwave. Mathematically, it is defined as thetime average of the Poynting vector.

power disturbance a variation of thenominal value of the voltage or current.

power divider passive electronic circuitconsisting of one input and two or more out-puts. A signal applied to the input is dividedinto equiphase output signals, generally ofequal amplitude.

power factor in an AC system, the ratioof the (active component) real powerP to theapparent powerS; it is given by the cosine ofthe angle subtended byS on the real,P axis.See alsoapparent power, real power, reactivepower.


power factor correction the addition ofreactive load to bring the combined powerfactor nearer unity. Since most industrialloads are inductive, capacitors are often em-ployed as passive devices for power factorcorrection.

power fault arc an arc through soilextending from a power lines’s lightningground to a buried, grounded structure.These may form when lightning strikes anenergized overhead electric power line.

power flow studies solutions of transmis-sion line active and reactive power flow andbus voltages giving system load.

power flow study the circuit solution ofan electric power system which yields thevoltage of each bus and thus the power flowsthroughout the system.

power follow a fault condition, especiallythrough a lightning arrester, in which powerline current flows along a path through airor other insulation broken down by a highvoltage impulse such as a lightning stroke toa conductor.Seepower fault arc.

power follow transformer a rugged,high-current power transformer used in testsof lightning arresters to test the arrester’spower follow arc suppression capability.

power flux density a vector that givesboth the magnitude and direction of an elec-tromagnetic field’s power flow. The units arewatts per square meter.

power fuse a protective device that con-sists of a fusible element and an arc quench-ing medium. An overload or fault currentin the fuse melts the fusible element, whichcreates an arc. The quenching medium theninterrupts the current at a current zero, andprevents the arc from restriking.

power gain dimensionless ratio of RFpower delivered to a load versus the input in-

cident RF power, normally expressed in deci-bels. Thus a power gain of 100 is expressedas 20 dB.

power line conditioner (PLC) equipmentinstalled on the customer’s side of the me-ter to eliminate momentary over- and under-voltages to critical loads.

power monitoring unit (PMU) a devicethat is installed in at a consumer’s site thatwill indicate to the utility whether or not anoutage condition exists. The units commer-cially available at the present time typicallyhave the facility to call power off and poweron status to a central facility via a telephonelink.

power noise factor for a photodetector,the ratio of signal-to-noise power ratio at theinput (SNPRin) to the signal-to-noise powerratio at its output (SNPRout). If the powernoise factor is denoted byk, then

k = SNPRin

SNPRout

power on self-test (POST) a series of di-agnostic tests performed by a machine (suchas the personal computer) when it powers on.

power optical power carried by an opticalfrequency beam.

power output useful output from a laseroscillator.

power quality (1) the concept of main-taining appropriate voltage and current wave-forms and frequency in transmission, dis-tribution, and generation systems, and usu-ally taken to mean undistorted and balancedwaveforms.

(2) a measure of an electric supply to meetthe needs of a given electrical equipment ap-plication. As delivered by the utility, powerquality is the faithfulness of the line voltageto maintain a sinusoidal waveform at ratedvoltage and frequency.


power signal suppose thatf (t) is a con-tinuous time signal and let

PT = 1

2π

∫ T

−Tf (t)2dt

represent the power dissipated byf (t) dur-ing the interval[−T , T ]. The signalf (t) isa power signal if limT→∞ PT exists and sat-isfies

0< P limT→∞PT < ∞ .

Furthermore, limT→∞ PT is the averagepower of the signal. Notice that an energysignal is not a power signal since necessarilylimT→∞ PT = 0.

Suppose thatf [k] is a discrete time signaland let

PT = 1

2T

T∑k=−T

f (tk)2 .

The signal f [k] is a power signal iflimT→∞ PT exists and satisfies

0< P limT→∞PT < ∞ .

Furthermore, limT→∞ PT is the averagepower of the signal. All discrete time pe-riodic signals are power signals.See alsoenergy signal, periodic signal.

power spectral analysis computation ofthe energy in the frequency distribution of anelectrical signal.

power spectral density (1) a function as-sociated with a random process that speci-fies the distribution of the signal power overthe different frequency components of thespectrum of the random process. The in-tegral of the power spectral density func-tion is the average power of the signal. TheWiener–Khinchine theorem states that thepower spectral densitySX(ω) of a randomprocessX(t) is equal to the Fourier transformof the autocorrelation functionRX(τ) of theprocess; i.e., for the processX(t) SX(ω) =∫∞−∞ RX(τ)e

−jωτ dτ , wherej2 = −1.

(2) the power spectral density function ofa signal is its power spectrum.

power spectrum (1) a representation ofpower contributed by each harmonic of thespectrum.

(2) for a wide sense stationary randomprocess is the Fourier transform of the auto-correlation function. The power spectrum isa measure of the average power of a randomprocess as a function of frequency. The ex-pected value of the periodogram tends to thepower spectrum as the periodogram windowlength tends to infinity.

power splitter passive device that acceptsan input signal and delivers multiple outputswith equal power.

power supply an electronic module thatconverts power from some power source toa form which is needed by the equipment towhich power is being supplied.

power supply unit (PSU) Seepowersupply.

power system damping a torque actionwhich works to reduce power system oscil-lations as time progresses.

power system oscillation variations in themachine rotor angle caused by small distur-bances in the system. The oscillations areusually at low frequency.

power system security an operational in-dex that determines the capability of a systemto withstand equipment failures and externaldisturbances. It could be measured by thereserve capacity available or the number ofoperating constraint violations under a givenprevailing condition or under a contingentprobability of disturbances.

power system stabilizer a control devicethat provides an additional input signal to theAVR to damp power system oscillations.


power transfer function any function ofinput to output power in ratio form, expressedas a dimensionless ratio.

power transformer a transformer that isused to transmit power from one voltage levelto another. Power transformers can be of ei-ther single phase or three phase design, andinclude either two or three windings.

Poynting vector a vector, proportionalto the cross product between the electricand magnetic field intensity vectors, indicat-ing the density and flow of electromagneticpower.

PPL Seephysical placement of logic.

PR function Seepositive real function.

practical intrafield predictor techniqueused to improve the image quality. Inpractical intrafield predictors, several is-sues are considered. For example, in two-dimensional images, prediction may not leadto significant improvements in entropy of theprediction error; however, there is a decreasein the peak prediction error. By choosingproper coefficients, it is possible to get im-proved prediction and a decay of the effectsof the transmission bit errors in the recon-structed picture.

practical stability Seeuniform ultimateboundedness.

practical stabilization deterministic ap-proach to robust systems design based on theuse of Lyapunov functions. Design feed-back control law ensures desirable behaviorof the systems for all admissible variations ofbounded uncertainty in the following sense.

There exists a neighborhood of the originB(d) such that

1. For any initial condition there exists asolution of the state equation for the closed-loop system in the given time interval.

2. All solutions for a given set of initialconditions can be extended over infinite timeinterval.

3. All solutions are uniformly boundedfor the given set of initial conditions (Seeuniform boundedness).

4. All solutions for the given set of initialconditions are uniformly ultimately boundedwith respect to a ballB(d0)with d0 > d ( Seeultimate boundedness).

5. All solutions from the ballB(d0) areuniformly stable.

If the given set of initial conditions couldbe extended to the whole state space, thenthe closed-loop system is globally practicallystable for the designed control law.

PRAM model a multiprocessor model inwhich all the processors can access a sharedmemory for reading or writing with uniformcost.

Prandtl number a nondimensional char-acteristic of fluids, relating the rate of mo-mentum diffusion to heat diffusion.

praseodymium doped fiber amplifieroptical amplifier based on the rare-earth ele-ment praseodymium used to amplify signalsin the 1310 nm telecommunications window.The region of useful gain is 1300–1350 nm.

preamplifier an amplifier connected toa low level signal source to present suitableinput and output impedances and to providegain so that the signal may be further pro-cessed without appreciable degradation in thesignal-to-noise ratio. A preamplifier mayinclude provisions for equalization and fre-quency discrimination.

prebake the process of heating the waferafter application of the resist in order to driveoff the solvents in the resist. Also called soft-bake and post-apply bake.

pre-compensator a compensator posi-tioned on the process input signal. ForMIMO systems, a given compensator will


have a different effect depending on whetherit is positioned before or after the process.Hence the importance of the prefix “pre-.”See alsocompensation, compensator, andpost-compensator.

pre-emphasis a technique of process-ing baseband signals through a network thatprovides a frequency-dependent amplitudetransfer characteristic.

precise interrupt an implementation ofthe interrupt mechanism such that the pro-cessor can restart after the interrupt at exactlywhere it was interrupted. All instructions thathave started prior to the interrupt should ap-pear to have completed before the interrupttakes place and all instructions after the in-terrupt should not appear to start until afterthe interrupt routine has finished.Comparewith imprecise interrupt.

predecrementation an addressing modeusing an index or address register in which thecontents of the address are reduced “decre-mented” by the size of the operand beforethe access is attempted.

predicate a logical expression that is tobe evaluated, often to determine whether animportant condition is satisfied. For exam-ple, “IF speed is high THEN reduce inputsupply.”

prediction (1) an estimation procedure inwhich a future value of the state (see the def-inition) is estimated based on the data avail-able up to the present time.

(2) in branching, the act of guessing thelikely outcome of a conditional branch deci-sion. Prediction is an important technique forspeeding execution in overlapped processordesigns. Increasing the depth of the predic-tion (the number of branch predictions thatcan be unresolved at any time) increases boththe complexity and speed.

predictive coding compression of a sig-nal by coding differences between samples

and predictions from previously coded val-ues. For example, in still image coding, a pre-dictive encoder may predict a pixel by takingthe average of the pixel’s left neighbor andits above neighbor. With raster-order codingthese values are already available in the de-coder, which can form the same prediction.The difference or prediction error values maybe quantized. Their probability density func-tion is approximately Laplacian, and furthercompression can therefore be achieved withentropy coding. Also calleddifferential pulsecode modulation.

predictive control control policy (scheme),realized at a given control layer, involvingrepetitive usage of a decision mechanismbased upon considering, at each interven-tion instant, the future operation of the con-trolled process (or the control system as awhole) over specified period of time (pre-diction interval). Usually, predictive controlinvolves the use of optimization-based deci-sion tools and of the free input forecasting;predictive control is the term describing a va-riety of possible control schemes, in particu-lar open-loop-feedback control and limited-look-ahead-control.

predictive pyramid a limited amount ofdata is used to form a prediction image, andthen the difference between the predicted im-age and the original image is used to form aresidual image. This can then further be it-erated to form a pyramid of residuals calledthe Laplacian pyramid.

predictive scalar quantization Seedifferential pulse code modulation.

predictive SQ Seedifferential pulse codemodulation.

predictive vector quantization the gen-eralization of scalar predictive coding tovector coding. Seedifferential pulse codemodulation.


predictive VQ Seedifferential pulse codemodulation.

predistorter predistortion type linearizerused to compensate the distortion componentgenerating at high power amplifier (HPA).Since the predistorter adds the distortioncomponent to the signal at 180 degrees outof phase, the distortion components, gener-ated in predistorter and HPA, are canceledout at the output of the HPA.

prefetch Seefetch policy.

prefetch queue in the CPU, a queue ofinstructions that has been prefetched prior tobeing needed by the CPU.

prefetching in the CPU, the act of fetch-ing instructions prior to being needed by theCPU.See alsofetch policy.

preformat information such as sector ad-dress, synchronization marks, servo marks,etc., embossed permanently on the opticaldisk substrate.

preincrementation an assembly lan-guage addressing mode in which the addressis incremented prior to accessing the mem-ory value. Used to access elements of arraysin memory.

preliminary breakdown an electricaldischarge in the cloud that initiates a cloud-to-ground flash.

preprocessing a series of image en-hancements and transformations performedto ease the subsequent image analysis pro-cess through, e.g., noise removal or featureextraction/enhancement.

pressure broadening spectral broadeningof a transition in a laser medium due to elasticor inelastic collisions.

pressure vessel a steel tank which en-closes the core of a nuclear reactor and isgenerally filled with coolant under pressure.

pressurized water reactor (PWR) a nu-clear reactor which uses liquid water underpressure for a primary coolant and modera-tor.

preventive congestion control protocolsthat control the system and user traffic param-eters so that the probability of a congestionis minimized.See alsocongestion.

price method coordination coordinationby the price method amounts to iterating thecoordination variables (dual coordination in-struments) defined as the vector of prices bywhich the local interaction inputs to the sub-processes as well as their outputs are multi-plied and added to the local performance cri-teria; these local criteria are then minimizedwith respect to the local decisions and theresults are passed to the coordinator. The co-ordinator iterates the values of prices untilthe interaction equations and, if needed, anyother coupling constraints are satisfied.

primal sketch a hierarchical image repre-sentation that makes explicit the amount anddisposition of the intensity changes in the im-age. At the lowest levels of the represen-tation, primitives just describe raw intensitychanges and their local geometrical structure;at the higher levels they represent groupingsand alignments of lower-level ones.

primary (1) the source-side winding.

(2) refers to the portion of a nuclear powerplant containing the reactor and its equip-ment.

primary coolant the medium used to re-move energy, in the form of heat, from a nu-clear reactor core, e.g., water, helium, or liq-uid metal.

primary feeder Seefeeder.


primary system of equations a system ofalgebraic and differential equations obtainedby applying the Kirchhoff’s current and volt-age laws and the element I-V relations.

primary voltage in power distribution thevoltage at the primary winding of the distri-bution transformer.

primary winding the transformer wind-ing connected to the energy source.

prime mover the system that provides themechanical power input for a mechanical-to-electrical energy conversion system (gener-ator), e.g., the diesel engine of an engine–generator set.

prime-number interleaving Seeinterleavedmemory.

primitive polynomial a polynomialp(x)of degreem that gives a complete table of2m distinct symbols containing 0 and 1. Thereciprocal of the primitive polynomial is alsoprimitive. See alsoirreducible polynomial.

primitive-based coding edges, lines, andother local features of images. Schemes todetect such primitives, then use them to codethe image. For example, edges may be usedto segment the image into regions which arethen independently coded as simple surfaces,while the boundaries are compressed with achain code.

Princeton architecture a computer ar-chitecture in which the same memory holdsboth data and instructions. This is contrastedwith the Harvard architecture, in which theprogram and data are held in separate mem-ories.

principal axis the optical axis of a lens orcamera, usually normal to the image plane.

principal component notionally, the di-rection of greatest variability of a randomvector or among a set of sample vectors.

More specifically, the principal componentis the direction of the eigenvector associatedwith the largest eigenvalue of the covariancematrix of the random vector (or the samplecovariance of a sample set). More generally,then principal components of a distributionare the eigenvectors corresponding to then

largest eigenvalues. Principal componentsare frequently used for data clustering, pat-tern analysis, and compression.

principal component analysis (PCA)a technique applied ton-dimensional datavectors with the aim of finding a set ofm-orthogonal vectors (typicallym � n) thataccount as much as possible for the data’svariance. PCA can be carried out in an un-supervised fashion by implementing a nor-malized version of Hebb’s rule onm-linearneural units.

principal point the point at which the op-tical axis of the lens in a camera meets theimage plane: also, the corresponding pointin the image.

principle of locality Seelocality. Seealsosequential locality.

principle of superposition in a linearelectrical network, the voltage or current inany element resulting from several sourcesacting together is the sum of the voltages orcurrents from each source acting alone.

printed circuit board (PCB) a substratemade from insulating material that has one ormore sandwiched metallic conductor layersapplied that are etched to form interconnect-ing traces useful for interconnecting compo-nents.

printer an output device for printing re-sults on paper.

prior statistics the statistics of a randomquantity (scalar, vector, process etc.) be-fore any experimental or measured knowl-


edge of the quantity is incorporated.See alsoposterior statistics.

prioritization coding a coding schemewhereby the position of the symbol in thedata steam indicates its weight.

priority encoder an encoder with the ad-ditional property that if several inputs are as-serted simultaneously, the output number in-dicates the numerically highest input that isasserted.

prism | air | metal (PAM) system thetwo-interface model of an ATR (attenuatedtotal reflection) system comprised of prism|air | metal. Commonly known as PAM sys-tem.

prismatic joint a joint characterized bya translation that is the relative displacementbetween two successive links. This transla-tion is sometimes called the joint offset.

private key cryptography also known assecret key cryptography. In such a crypto-graphic system, the secret encryption key isonly known to the transmitter and the receiverfor whom the message is intended. The se-cret key is used both for the encryption of theplaintext and for the decryption of the cipher-text. See alsopublic key cryptography.

privileged instruction an instruction thatcan be executed only when the CPU is in priv-ileged mode.

privileged mode a mode of execution ofmachine instructions in the CPU in whichcertain special instructions can be executedor data accessed that would otherwise be pro-hibited. See alsouser mode.

PRMA Seepacket reservation multipleaccess.

probabilistic metric space a generaliza-tion of the notion of metric spaces onto theuncertain systems by replacing a metric on a

given setS by a distance distribution func-tion F , and a triangle inequality by a gen-eralized inequality defined by triangle func-tion τ . A distance distribution functions be-tween two elementsp, q ∈ S is defined asa real functionFpq whose valueFpq(x) forany real numberx is interpreted as the proba-bility, the membership function, or the gradeof membership (depending on the type of theuncertainty model) that the distance betweenp andq is less thanx. The simplest distancedistribution function is given by the unit step(Heaviside) function1 as follows:

Fpq(x) = 1(x − d(p, q))

whered is a standard metric. Then a proba-bilistic metric space reduces to the standardmetric space. More precisely, a probabilis-tic metric space (PMS) is defined as a triple(S, F, τ) endowed with the following prop-erties:

Fpq(x) = 1(x) ⇐⇒ p = q

Fpq = Fqp

Fpr ≥ τ(Fpq, Fqr

)for all p, q, r ∈ S. The types of the distri-bution function and the triangle function arerelated to the model of uncertainty and theway of composition of the standard opera-tions in the model.

probabilistic neural network a term ap-plied loosely to networks that exhibit someform of probabilistic behavior but also ap-plied specifically to a type of network devel-oped for pattern classification based upon sta-tistical techniques for the estimation of prob-ability densities.

probability density function (PDF) (1) afunction describing the relative probability ofoutcomes of an experiment. For experimentswith discrete outcomes, the PDF is analogousto a relative frequency histogram. For exper-iments with continuous outcomes, the PDFanalogous to a relative frequency histogramwhere the category bin widths are reduced to


zero. The total area underneath a PDF mustalways be unity.

(2) the derivative of the cumulative distri-bution function (when the derivative exists).More formally, for a randomx and any proba-bilistic eventA, the probability density func-tion px(x) satisfies

Pr(x ∈ A) =∫A

dpx(x).

See alsocumulative distribution function.

procedure a self-contained code se-quence designed to be re-executed from dif-ferent places in a main program or anotherprocedure.See alsocall instruction, returninstruction.

procedure call in program execution, theexecution of a machine-language routine, af-ter which execution of the program continuesat the location following the location of theprocedure call.

process the context, consisting of allo-cated memory, open files, network connec-tions, etc., in which an operating systemplaces a running program.

process control block (PCB) an areaof memory containing information about thecontext of an executing program. Althoughthe PCB is primarily a software mechanismused by the operating system for the controlof system resources, some computers use afixed set of process control blocks as a mech-anism to hold the context of an interruptedprocess.

process environment part of the controlscene that is outside of the controlled process;within this environment are formed the un-controlled, free inputs to the process. Speci-fied quantities related to the environment maybe observed and used by the controller, forexample, when performing free input fore-casting.

process interaction a stream of energy,material, or information exchanged betweenthe sub-processes of a large-scale process.Relevant attributes of those streams are, re-spectively, interaction inputs or interactionoutputs. Interactions are described by the in-teraction equations, which relate interactioninputs to a given subprocess to interactionoutputs from other subprocesses.

process monitor in semiconductor man-ufacturing, wafers processed at a particularprocess step to permit the quality or someother attribute of the step to be monitored.For example, in a photolithographic etchingprocess, a process monitor would be a waferexposed and etched that permits the physicaldimensions of the resulting etched layer to bemonitored and measured.

process state the set of information re-quired to resume execution of a process with-out interfering with the results of the compu-tation.

process swap the act of changing the ex-ecution point from one process to another.

process window a window made by plot-ting contours corresponding to various spec-ification limits as a function of exposure andfocus. One simple process window, calledthe critical dimension (CD) process window,is a contour plot of the high and low CD spec-ifications as a function of focus and expo-sure. Other typical process windows includesidewall angle and resist loss. Often, sev-eral process windows are plotted together todetermine the overlap of the windows.

process-oriented analysis a method ofanalyzing application transformation pro-cessing as the defining characteristic of ap-plications.

processing amplifier high-performanceamplifier that regenerates as well as amplifiesthe signal being processed.


processing element (PE) a processingmodule, comprising at least a control section,registers, and arithmetic logic, in a multipro-cessor system. A processing element may becapable of operating as a stand-alone proces-sor.

processing ensemble a collection of pro-cessors under control of a single control unit.

processing gain Seespreading gain.

processor element Seeprocessing element.

processor farm a collection or ensem-ble of processing elements to which paral-lel processing tasks are assigned and dis-tributed for concurrent execution. In thismodel, tasks are distributed, “farmed out,” byone “farmer” processor to several “worker”processors, and results are sent back to thefarmer. This arrangement is suitable for ap-plications that can be partitioned into manyseparate, independent tasks. The tasks arelarge and the communications overhead issmall.

processor status word a register in theCPU that stores a collection of bits that, takenas a group, indicate the status of the machineat a given period of time.

product code two-dimensional burst andrandom error correcting code in the form ofa matrix with each row and column are codewords of two different linear codes.

product of sums (PS) the AND combi-nation of terms, which are OR combinationsof Boolean variables.

program counter (PC) a CPU registerthat contains the address in memory of thenext instruction to be fetched and executed.

program status word (PSW) a combina-tion program counter and status-flag registerprovided in IBM mainframe computers.

programmable array logic (PAL) aprogrammable logic array with no OR array,but with a fixed set of OR gates into whichare fed sets of product terms.

programmable gate array (PGA) Seefield-programmable gate array.

programmable logic array (PLA) a pro-grammable logic device that consists of anAND array forming logical products of theinput literals and an OR array that sums theseproducts to form a set of output functions.

programmable logic controller (PLC)a microprocessor based system comprised ofa set of modules for acquiring signals fromenvironment and other for producing effectson the environment. These effects are typ-ically used for controlling electromechani-cal machines by means of actuators (motors,heating element, etc.). The rules for spec-ifying the control law are Boolean expres-sions. These sequentially and cyclically areexecuted. More complex PLCs allow the de-scription of the control law by means of lad-der diagram.

programmable logic device (PLD) an in-tegrated circuit able to implement combina-tional and/or sequential digital functions de-fined by the designer and programmed intothis circuit.

programmable radio system radiosbased on digital waveform synthesis and dig-ital signal processing to allow simultaneousmultiband, multiwaveform performance.

programmable read-only memory(PROM) a semiconductor memory devicethat has a primary function of storing data in anonvolatile fashion that can be programmedto contain predetermined data by means otherthan photomasking. PROMs may be one-time programmable (OTP) or may be eitherUV or electrically eraseable, depending onthe particular semiconductor process tech-nology used for manufacturing.


programmed I/O transferring data to orfrom a peripheral device by running a pro-gram that executes individual computer in-struction or commands to control the trans-fer. An alternative is to transfer data usingDMA.

progressive coding ordering of coded val-ues such that the original signal can be recov-ered progressively. For example, in trans-form coding of a picture, transmission of thezero sequency coefficients for all blocks first,rather than transmission of all coefficients forthe first block first. This allows the receiver togenerate an approximate reconstruction earlyin the reception of data.

progressive scan method of scanning adisplay on a horizontal line-by-line basis instrict sequence during one vertical sweep ofthe scanning beam/element. Each scan pro-vides one complete video frame, and inter-laced fields do not occur in either the pro-duction or display of the picture element se-quence. Also known as sequential scan.

progressive transmission partial infor-mation of an image is transmitted. At eachstage, an approximation of the original imageis reconstructed at the receiver. The qualityof the reconstructed image improves as moreinformation is received.

projection in signal and image process-ing, the conversion of ann-dimensional sig-nal into ann−1 dimensional version throughsome integration in the continuous case, orsome summation in the discrete case. Forinstance, a 2-D image can be viewed as a(perspective) projection of 3-D scene via acamera. Another example exists in computedtomography. There a projection is a line in-tegral along a straight ray.

projection of a fuzzy set for fuzzy setQ ina Cartesian productXn,Xn = X1×· · ·×Xn,the, fuzzy set subspaceR in Xi of Xn, (i <n). The projection is usually denotedR =proj (Q;Xi), inXm with membership func-

tion obtained as the supremum of the mem-bership function for the dimensions to beeliminated.See alsocylindrical extension ofa fuzzy set, fuzzy set, membership function.

projection printing a lithographic methodwhereby the image of a mask is projectedonto a resist-coated wafer.

projection receiver multiple-access re-ceiver technique in which users are estimatedby first projecting the received signal ontothe null space of the unwanted or interferingusers.

projection television system a TV sys-tem in which a small but very bright imageis optically projected onto a large viewingscreen.

projective invariant a measure that is in-dependent of the distance and direction fromwhich a particular class of object is viewed,under perspective projection. The cross ra-tio is an important type of projective in-variant, which is constant for four collinearpoints viewed under perspective projectionfrom anywhere in space. Projective invari-ants are important in helping with egomotion(e.g., automatic guidance of a vehicle) and forinitiating the process of object recognition in3-D scenes.

PROM See programmable read-onlymemory, Pockels readout optical modulator.

prompt critical a nuclear reaction whichcan maintain criticality without the contribu-tion of delayed neutrons, i.e., the sort of re-action which takes place in nuclear weapons.

propagating mode pattern of electromag-netic field with most of the energy containedin the core of the fiber that is sustained forlong distances in an optical fiber (a few kilo-meters to 10’s of kilometers) with only theamplitude of the electromagnetic field grad-ually decreasing while the shape of the moderemains constant.


propagating wave a wave that exhibits aspatial shift as time advances. Mathemat-ically, a propagating wave is described byf (t − r/vp), wheref is some function,tis the temporal variable,r is the spatial vari-able, andvp is the phase velocity of the wave.

propagation the motion of electromag-netic waves through a medium or free space.

propagation constant a complex con-stant whose real part is the wave attenuationconstant (nepers attenuation per unit length)and whose imaginary part is the wave phaseconstant (radian phase shift per unit length).See alsophase constant.

propagation delay (1) the delay be-tween transmission and reception of a signal.Caused by the finite velocity of electromag-netic propagation.

(2) the delay time between the applica-tion of an input signal to a chain of circuitelements and the appearance of the resultingoutput signal.

(3) the time it takes for a transistor switchto respond to an input signal, symbolizedtpd .It is calculated between the 50% rise point tothe 50% fall point or vice-versa (see graph oftypical inverter gate).

Propagation delay for a typical inverter gate.

The time from when the input logic levelto a device is changed until the resultant out-put change is produced by that device.

propagation delay time Seepropagationdelay.

propagation factor the ratio of the elec-tric field intensity in a medium to its value ifthe propagation took place in free space.

propagation mode (1) in radio commu-nication, refers to the manner in which ra-dio signals propagate from the transmittingantenna to the receiving antenna. Radiowaves can be reflected from the ionized lay-ers within the ionosphere. Scatter modes ofpropagation also exist, and in these modes,propagation from the transmitting to receiv-ing antennas is largely the result of scatteringof the radio waves by dust and other particlesin the troposphere. The ionized trails of me-teors can also act as scatterers.

(2) in waveguides, used at microwave fre-quencies, propagation mode refers to the ar-rangement of electrical and magnetic fieldswithin the waveguide. Modes can be classi-fied as being transverse electric (TE) or trans-verse magnetic (TM), depending on whetherthe electric field or magnetic field is trans-verse to the direction of wave propagationwithin the waveguide. The mode that existsfor the lowest frequency that can be transmit-ted by the waveguide is known as the domi-nant mode of the waveguide.

propagation path the route along whicha radio wave propagates from the transmit-ting antenna to the receiving antenna. If thisis a straight line, the path is a line of sightpath. Otherwise, reflection, diffraction, andother phenomena may change the directionof propagation so that the path can be en-visioned as made up of several straight-linesegments.See alsoline of sight, multipathpropagation.

proper 2-D transfer matrix the 2-Dtransfer matrix

T (z1, z2) = N(z1, z2)

d(z1, z2)

N(z1, z2) ∈ Rp×m [z1, z2]


dz1,z2 =n1∑i=0

n2∑j=0

dij zi1zj

2

is called proper ifdn1,n2 6= 0 (d(z1, z2) isacceptable) and

degz1 nij (z1, z2) ≤ n1

degz2 nij (z1, z2) ≤ n2

for i = 1, . . . , p; j = 1, . . . , m ;wherenij (z1, z2) are the entries ofN(z1, z2)

and degz1(degz2) denotes the degree with re-spect toz1 (z2).

proper mode a mode as obtained from aboundary value problem defined over a finiteinterval and relative to a second-order dif-ferential equation with real coefficients.Seealso improper modes.

proper subgraph a subgraph that does notcontain all of the edges of the given graph.

proportional control control schemewhereby the actuator drive signal is pro-portional to the difference between the in-put/desired output and measured actual out-put.

proportional-integral-derivative (PID) con-trol a control scheme whereby the signalthat drives the actuator equals the weightedsum of

1. the difference,2. time integral of the difference, and3. time derivative of the difference be-

tween the input and the measured actual out-put.

It is a widely used control scheme in in-dustry that can be tuned to give satisfactoryperformance based on knowledge of domi-nant system time constant.

protection control access to informationin a computer’s memory, consistent with aparticular policy or mechanism. Ring num-bering was introduced in the Multics systemas one basis for limiting access and protect-ing information. The term “security” is used

when the constraints and policies are very re-strictive.

protection fault an error condition de-tected by the address mapper when the typeof request is not permitted by the object’s ac-cess code.

protective relay a device that monitorsthe condition of the electric power systemand determines the presence of faults or othersystem anomalies. The protective relay mon-itors current flow, voltage level, or other pa-rameter. When it operates due to a fault orother event, it initiates a trip signal intendedto open the appropriate circuit breaker(s) orother protective devices.

protective relaying a unit for discriminat-ing normal operating condition from faultedconditions in power systems.

protocol Seebus protocol.

protocol data unit (PDU) the unit of ex-change of protocol information between en-tities. Typically, a PDU is analogous to astructure in C or a record in Pascal; the pro-tocol in executed by processing a sequenceof PDUs.

protocol spoofing a technique used byVSAT networks to reduce the network de-lay. The satellite network emulates the hostcomputer front-end processor at the VSATlocations and emulates the multiple clustercontrollers at the hub location.

prototype filter a low-pass filter that ismodulated to form a modulated filter bank.

prototyping building an engineeringmodel of all or part of a system to prove thatthe concept works.

proximity effect a variation in the size ofa printed feature as a function of the sizes andpositions of nearby features.


proximity printing a lithographic methodwhereby a photomask is placed in close prox-imity (but not in contact) with a photoresistcoated wafer and the pattern is transferred byexposing light through the photomask intothe photoresist.

pruning self-generating neural networka methodology for reducing the size of self-generating neural network (SGNN).

During SGGN training, the network maygrow very quickly, and some parts of the net-work may become useful in neither train-ing nor classification, and the weights ofthese parts of the neurons of the networknever change after some training stage. Wecall these parts of the network dead sub-nets. It is obvious that the dead subnets ofthe network should be pruned away to re-duce the network size and improve the net-work performance. One way to achieve thisis to check the weights of each neuron inthe network to see whether they have beenchanged since last training epoch (or dur-ing the last few epochs). If they are un-changed, the neuron may be evidently deadand should be removed from the network. Ifa neuron is removed from an SGNN, all ofits offspring should also be removed.Seealsoself-generating neural network, trainingalgorithm of self-generating neural network.

PS Seeproduct of sums.

PSC Seepermanent split-capacitor motor.

PSC motor Seepermanent split-capacitormotor.

pseudo code a technique for specifyingthe logic of a program in an English-like lan-guage. Pseudo code does not have to followany syntax rules and can be read by anyonewho understands programming logic.

pseudo color the display in color of gray-level pixels to make certain gray-level pixels

or gray level patterns more visible. Typicalpseudocoloring schemes are

1. displaying all pixels in a range of graylevels in one color;

2. displaying each gray level at a fixed in-terval in a different color to produce coloredcontour lines;

3.displaying the entire gray-level range asa rainbow. Pseudocoloring is usually carriedout by look-up tables.See alsofalse color,look-up table model (LUT).

pseudo-exhaustive testing a testing tech-nique that relies on various forms of circuitsegmentation and application of exhaustivetest patterns to these segments.

pseudo-noise (PN) describes a class ofpseudo-randomly generated processes thatexhibit the statistical properties of noise.

pseudo-noise code phase the startingsymbol of a pseudo-noise (PN) code with re-spect to some time reference in the system.In the IS-95 CDMA system, the same PNcode with different PN code phases (typicallymultiples of 64 code symbols, or chips) is as-signed as the pilot sequence for different basestations.See alsospreading code.

pseudo-noise code tracking the act ofaccurately generating a local pseudo-noise(PN) code signal whose phase (timing rela-tive to some reference time) accurately tracksthe phase of an incoming PN code signal.

pseudo-noise sequence Seespreadingsequence.

pseudo-operation in assembly language,an operation code that is an instruction to theassembler rather than a machine-language in-struction.

pseudo-QMF a cosine-modulated filterbank with the prototype filter being even-length symmetric spectral factor of anM-th band filter. The aliasing of the adjacent


bands is cancelled while the aliasing of non-adjacent bands is very small.

pseudo-random describes a process thathas the statistical properties of being ran-dom, but is deterministic and can be recreatedgiven the requisite properties of its generator.

pseudo-random pattern generator de-vice that generates a binary sequence of pat-terns where the bits appear to be random inthe local sense (1 and 0 are equally likely),but they are repeatable (hence only pseudo-random).

pseudo-random skewing Seeinterleavedmemory.

pseudo-random testing (1) a testingtechnique (often used in BISTs) that baseson pseudo-randomly generated test patterns.Test length is adapted to the required level offault coverage.

(2) a technique that uses a linear feed-back shift register (LFSR) or similar structureto generate binary test patterns with statisti-cal distribution of values (0 and 1), acrossthe bits; these patterns are generated withoutconsidering the implementation structure ofthe circuit to which they will be applied.

PSK Seephase-shift keying.

PSNR Seepeak signal to noise ratio.

PSU power supply unit. See powersupply.

PSW Seeprogram status word.

psychovisual redundancy the tendencyof certain kinds of information to be rel-atively unimportant to the human visualsystem. This information can be elimi-nated without significantly degrading im-age quality, and doing so is the basis forsome types of image compression.Seealso image compression, Joint PhotographicExperts Group (JPEG), quantization matrix.

PT potential transformer. Seevoltagetransformer.

PU-239 a man-made, fissile element suit-able for use in nuclear weapons as well aspower reactors.

public key cryptography two differentkeys are used for the encryption of the plain-text and decryption of the ciphertext. When-ever a transmitter intends sending a receivera sensitive message that requires encryption,the transmitter will encrypt the message us-ing a key that the receiver makes availablepublically to anyone wanting to send themencrypted messages. On receiving the en-crypted message, the receiver applies a secretkey and recovers the original plaintext infor-mation sent by the transmitter. In contrastto secret key cryptography, public key cryp-tography does not suffer from the problem ofhaving to ensure the secrecy of a key.Seealsoprivate key cryptography.

public key cryptosystem system that usesa pair of keys, one public and one private, tosimplify the key distribution problem.

pull-in time the time required by an au-tomatic frequency control (AFC) system tolock or stabilize after a frequency change.

pull-in torque the amount of torqueneeded to change a synchronous motor’s op-eration from induction to synchronous whenself-started.

pull-out torque the maximum value oftorque that an AC motor can deliver. An in-duction motor operating at the pull-out torquewill operate at maximum slip, and load-ing it beyond the pull-out torque will causethe motor to stall. Synchronous motors re-main at synchronous speed up to the pull-outtorque. Exceeding the pull-out torque for asynchronous machine will lead to pole slip-ping and destruction of the machine.


pull-up torque the minimum torque gen-erated by an AC motor as the rotor accel-erates from rest to the speed of breakdowntorque. For an induction motor, this valueusually is less than the locked-rotor torque,and thus establishes the maximum load thatcan be started.

pulse a sudden change of an electricalvalue of short duration with a quick returnto the original value. A pulse injects a shortsharp burst of energy into a system and isusually quantified by its area rather than itsamplitude or its duration. In the limit as theamplitude tends to infinity and the durationto zero, it approaches the Dirac delta functionwhose Laplace transform is unity.

pulse amplitude modulation (PAM) aprocess in which an analog signal is sampled,the sampled values are then used to modulatethe amplitudes of pulses. PAM is one type ofanalog pulse modulation, which is differentfrom pulse code modulation.

pulse broadening Seepulse distortion.

pulse code modulation (PCM) a processin which an analog signal is sampled, thesampled value is then quantized, the quan-tized level is (usually binary) encoded, andfinally the code is represented by pulses ex-hibiting amplitude, phase or frequency mod-ulation.

pulse compression technique of correlat-ing identically coded signals, such as chirpsignals, to produce a sharply peaked corre-lation function, the peak width being muchshorter than that of the original coded pulsesby a ratio equal to the time-bandwidth of thecoded signals.

pulse distortion a spreading or lengthen-ing of the temporal shape of an optical signaltransmitted through an optical fiber causedby a combination of wavelength effects (dis-persion) and multimode and polarization ef-

fects. Alsocalledpulsebroadening.

pulse Doppler a coherent radar, usuallyhaving high pulse repetition rate and dutycycle and capable of measuring the Dopplerfrequency from a moving target. Has goodclutter suppression and thus can see a movingtarget in spite of background reflections.

pulse forming network (PFN) (1) a trans-mission line with different impedance alongits length.

(2) a lumped element circuit consisting ofinductors and capacitors designed to delivera square pulse.

pulse number of a converter number ofripples in DC voltage per cycle of AC voltage.A three-phase, two-way bridge is a six-pulseconverter.

pulse propagation time and space depen-dence of a travelling electromagnetic pulse.

pulse shape refers to the time domaincharacteristics of a pulse that is typically usedto transmit data symbols over a channel. Inmost cases the pulse is shaped using a trans-mitter filter so that it satisfies the Nyquist Icriterion for zero intersymbol interference.

pulse spreading increase with time or dis-tance of the length of a propagating electro-magnetic pulse.

pulse width width of the temporal regionover which a pulse has significant amplitude;sometimes represented more specifically as,for example, full width at half maximum.

pulse width modulated inverter an in-verter that uses pulse width modulation tocreate AC voltages from a DC supply.Seealsopulse-width modulation.

pulse-code modulation (PCM) (1) a dig-ital encoding technique whereby an analogsignal is converted into a sequence of digital


words in a periodic pulse train. The PCMprocess starts by bandlimiting the analog in-put waveform to one half the sampling rate( Seealiasing). The analog signal is peri-odically sampled, using flat-top PAM (Seepulse amplitude modulation). The height ofeach pulse is then quantized (estimated atan integer value) using an analog-to-digitalconverter, with the number of discrete quan-tizing levels determined by the resolution ofthe ADC used. The integer number associ-ated with each sample is converted into bi-nary format, then serially transmitted (MSBfirst). Note that for an 8-bit resolution ADC,each amplitude sample from the original ana-log waveform requires a sequence of 8 databits in the output pulse train to represent it;as such, the output data clock rate must beat least eight times faster than that of theclock used to sample the analog input wave-form. See alsoanalog-to-digital conversionandquantization.

(2) in image processing, a system thattransmits quantized amplitudes of each pixel.Each pixel is assigned a unique binary wordof finite length. Coding of image inten-sity samples in this manner is called pulsecode modulation or PCM. In the case of theDPCM, the difference resulting from the sub-traction of the prediction and the actual valueof the pixel is quantized into a set ofL dis-crete amplitude levels.See alsodifferentialpulse code modulation.

(3) in color signal encoding, PCM digiti-zation of R, G, and B using full bandwidth foreach component is common practice, as thePCM signals can be used for filtering, stor-age, compression, etc. But PCM encoding ofcolor signals may not always use all the R,G, and B signals. However, if all the compo-nents R, G, B have the same bandwidth thenfewer bits are used to quantize the red andblue compared with the green signal.

pulse-echo ultrasound using a probe con-taining a transducer to generate a short ultra-sound pulse and receive echoes of that pulse,associated with specular reflection from in-terfaces between tissues or scattering from

inhomogeneities within the tissue, to form adisplay of tissue backscatter properties.

pulse-width modulated inverter a self-commutated inverter where the amplitudeand frequency of the output are controlledby PWM. The switching frequency is gener-ally much higher than the output frequencyto ensure a smooth output waveform. In thefollowing example, PWM is performed bycomparing a sine reference with a trianglecarrier. The output before and after the low-pass filter is shown in the figure.

Waveforms of a PWM inverter.

pulse-width modulated switch an activeswitch driven by a pulse-width modulated(PWM) pulse train.

pulse-width modulation (PWM) a con-trol technique used in variable speed DC,AC, or other electrical variable speed drivesto control the harmonic content of the ap-plied voltage or current. Typically, the pulsewidth is modulated in three ways, trailing-edge modulation, leading-edge modulation,and double-edge modulation. Most popularis sinusoidal PWM for AC drives.


PWM is most frequently used in switchingconverter technology as the drive signal foractive switching elements.

pulsed laser laser designed to produceits output in the form of isolated or periodicpulses that may be either long or short com-pared to the length of the laser cavity.

pulsed network a network that conveysinformation between neurons by means ofpulse streams. The information may becoded using conventional pulse modulationtechniques or by using random bit streams,in which case the information is coded intothe statistics of the stream.

pulsed response the response of a system,or circuit to a pulsed input signal.

pumping process for achieving popula-tion inversion and gain in a laser medium;process for providing energy to an arbitrarynonlinear optical system; process in whichinput energy is used to obtain a populationinversion in a laser medium.

punched card a method, now obso-lete, used to represent data and programsas cardboard cards where the values wererepresented by punched holes at appropriateplaces.

punctured code a code constructed fromanother code by deleting one or more coor-dinates from each codeword. Thus an(n, k)original code becomes an(n−1, k) code afterthe puncturing of one coordinate.

punctured convolutional code a codewhere certain symbols of a rate 1/n convo-lutional code are periodically punctured ordeleted to obtain a code of higher rate. Be-cause of its simplicity, it is used in manycases, particularly in variable rate and high-speed applications.Compare withpuncturedcode.

puncturing periodic deletion of codesymbols from the sequence generated by aforward error control convolutional encoder,for purposes of constructing a higher ratecode. Also, deletion of parity bits in a for-ward error control block code.

pure ALOHA type of multiple accessprotocol. Any user is allowed to transmitat any time. The possible collisions resultin destroyed packets. Destroyed packetsare retransmitted at a later time.See alsoALOHA.

pure longitudinal wave ultrasonic planewave in which the particle motion is parallelto the wave vector and for which energy flowis parallel to the wave vector.

pure procedure a procedure that does notmodify itself during its own execution. Theinstructions of a pure procedure can be storedin a read-only portion of the memory and canbe accessed by many processes simultane-ously.

pure shear wave ultrasonic plane wavein which the particle motion is parallel to thewave vector and for which energy flow is per-pendicular to the wave vector.

purity in a color display, the productionof red image content (only) by the red pic-ture signal, blue image content (only) by theblue picture signal, and green image content(only) by the green picture signal. Pictureimpurity may be due to a color phosphor be-ing excited by an inappropriate color electronbeam.

Purkinje region the region of light inten-sity where, as illumination is reduced slowlyenough to allow the eye to adapt, sensitiv-ity to long wave light (red–orange) decreasesmore rapidly than sensitivity to short wavelight (blue–violet).

pursuit behavior the human operator’soutputs depend on system errors, as in com-


pensatory behavior, but may also be di-rect functions of system inputs and outputs.The human response pathways make theman–machine system a combined open-loop,closed-loop system.

pursuit display in the simplest case, adisplay that shows input command, systemoutput, and the system error as separable en-tities.

push brace a rigid brace, typically anotherutility pole, which is angled against a utilitypole to serve the purpose of a guy where aguy cannot be placed.

push instruction an instruction that storesthe contents of specified register/s on thestack.

push–pull amplifier an amplifier madeup of two identical class B amplifiers oper-ated as a balanced pair (180 degrees out ofphase with each other) into a common load,resulting in each amplifier operating over al-ternating half cycles of the input signal, andhaving the output signal combined such thatfull cycles are dissipated across the load. Oddorder harmonics are present at the load, whileeven order harmonics are suppressed, result-ing in an output signal that is an odd ordermultiple of the input frequency (either a fun-damental frequency amplifier or an odd orderfrequency multiplier).

push–push amplifier an amplifier madeup of two identical class B amplifiers oper-ated as a balanced pair (180 degrees out ofphase with each other) into a common load,resulting in each amplifier operating over al-ternating half cycles of the input signal, andhaving the output signal combined such thatrectified half cycles are dissipated across theload. Even order harmonics are present atthe load, while odd order harmonics are sup-pressed, resulting in an output signal that isan even order multiple of the input frequency(even order frequency multiplier).

pushdown stack a data structure contain-ing a list of elements that are restricted toinsertions and deletions at one end of the listonly. Insertion is called a push operation anddeletion is called a pull operation.

PV bus Seevoltage-controlled bus.

PWM Seepulse-width modulation.

PWM inverter Seepulse-width modulatedinverter.

PWM switch Seepulse-width modulatedswitch.

PWR Seepressurized water reactor.

pyramid coding any compression schemethat repeatedly divides an image into two sub-bands, one, a lowpass representation that issubsampled and used as input for the nextlevel, and the other an error (difference) im-age. A small lowpass image plus a pyra-mid of difference images of increasing size isgenerated, allowing the lowpass informationto be coded accurately with few codewords,and the highpass information to be codedwith coarse quantization.See alsoLaplacianpyramid.

pyramidal horn a horn antenna where theaperture is formed by flaring the walls in boththe E-plane and the H-plane. The flare anglesfor the E-plane and H-plane can be adjustedindependently.

pyroelectric a polar dielectric material inwhich the internal dipole moment is tempera-ture dependent. This leads to a change in thecharge balance at the surface of the materialwhich can be detected as either a potentialdifference or as a charge flowing in an exter-nal circuit.

pyrolytic grid a grid structure made ofpyrolytic (oriented) graphite that is laser-cut


to the proper geometry for use in a powervacuum tube.

pyrosensitive smart material materialthat self-adaptively (smartly) manages theelectromagnetic surface characteristics ofactive surfaces constituted by pyrosensi-tive inclusions, in response to an external

temperature-inducing stimulus applied perthe feedback information on electromagneticcharacteristics.

PZT generic name for piezoelectric mate-rials of the lead (Pb) zirconate titanate family.


Qq unit of electric charge.q = 1.602×10−19 coulombs.

Q Seequality factor.

q axis Seequadrature axis.

QL common symbol loaded quality factor.QL is dimensionless.

QU common signal for unloaded qualityfactor. QU is dimensionless.See alsoqualityfactor.

Q-factor a figure of merit that representsthe ratio of stored to dissipated energy percycle.

Q-switching rapid increase in the qual-ity of a laser cavity from below threshold toabove threshold for the purpose of obtaininga large output optical pulse; also called loss-switching.

QAM See quadrature amplitudemodulation.

QOS Seequality of service.

quad tree a tree data structure used to rep-resent an image, where each node is a squaresubset of the image domain. The root repre-sents the whole image domain, which is sub-sequently subdivided in square subsets untilall pixels in each node have the same value,i.e., until all nodes become “pure.”

quadratic detector a detector that makesuse of the second-order statistical structure(e.g., the spectral characteristics) of measure-ments. The optimum structure for detecting

a zero-mean Gaussian signal in the presenceof additive Gaussian noise is of this form.

quadratic gain profile gain that variesquadratically with distance away from theaxis of an optical medium.

quadratic index profile index of refrac-tion that varies quadratically with distanceaway from the axis of an optical medium.

quadratic measure of sensitivity whenthe multiparameter sensitivity row vector isknown and the componentsxi have toleranceconstantsεi (these are considered to be pos-itive numbers), i.e.,

xi0 (1 − εi) ≤ xi ≤ xi0 (1 + εi)

then the quadratic measure of sensitivity isdefined as

MQ =∫ ω2

ω1

[n∑i=1

(ReSF(jω,x)xi

)2ε2i ]dω

quadratic media propagation media inwhich the gain and/or index of refraction varyquadratically with distance away from theaxis.

quadratic phase coupling a measure ofthe degree to which specific frequencies in-teract to produce a third frequency.

quadratic stabilizability a property of adynamical (possibly uncertain) system thatcould be stabilized using state feedback de-signed via a quadratic Lyapunov function. Ifthe system is linear, then it is usually requiredthat it could be quadratically stabilizable vialinear feedback. It means that there existssuch a linear in statex feedback control lawand a positive definite matrixP that the re-sulting closed-loop system has the propertythat the derivative of the Lyapunov functionV (x) = x′Px is negative definite for all pos-sible uncertainties. The suitable matrixPmay be found, for example, by solving a re-spective game type Riccati equation or para-metrically scaled linear matrix inequalities.


quadrature a condition where there is a90 degree separation between items. That is,they are at right angles to one another.

quadrature amplitude modulation (QAM)a modulation technique in which the incom-ing symbols are split into two substreams,which are modulated in quadrature. In theconventional implementation, the “in-phase”symbols are modulated by cos 2πfct , andthe “quadrature” symbols are modulated bysin 2πfct , so that the transmitted phases as-sociated with the two substreams differ byπ/2.

quadrature axis (q axis) an axis placed90 degrees ahead of the direct axis of a syn-chronous machine.Seedirect axis.

quadrature axis magnetizing armature re-actance a reactance that represents all theinductive effects of the q-axis stator currentof a synchronous machine, except for thatdue to the stator winding leakage reactance.In Park’s q-axis equivalent circuit of the syn-chronous machine, this reactance is the onlyelement through which both the stator androtor currents flow. Its value may be deter-mined by subtracting the stator winding leak-age reactance from the steady-state value ofthe q-axis operational impedance or from thegeometric and material data of the machine,using expressions described in.

quadrature axis magnetizing reactanceSeequadrature axis magnetizing armaturereactance.

quadrature axis subtransient open circuittime constant a constant that character-izes the initial decay of transients in the q-axis variables of the synchronous machinewith the stator windings open-circuited. Theinterval characterized is that immediately fol-lowing a disturbance during which the effectsof all amortisseur windings are considered.A detailed (derived) closed-form expressionfor the subtransient open-circuit time con-stant of a machine with two q-axis amortis-

seur windings is obtained by taking the recip-rocal of the smallest root of the denominatorof the q-axis operational impedance. An ap-proximate (standard) value is often used, inwhich it is assumed one amortisseur windingresistance is very small relative to the otherand the detailed expression simplified. Ex-pressions for the q-axis subtransient open-circuit time constant of a machine with twoq-axis amortisseur windings are derived in.

quadrature axis subtransient reactancethe high-frequency asymptote of the q-axisoperational impedance of a synchronous ma-chine. The reactance characterizes the equiv-alent reactance of the q-axis windings of themachine during the initial time following asystem disturbance. In models in which therotor windings are represented as lumped pa-rameter circuits, the q-axis subtransient reac-tance is expressed in closed form as the sumof the stator winding leakage reactance, andthe parallel combination of the q-axis mag-netizing reactance and the q-axis rotor amor-tisseur leakage reactances.

quadrature axis synchronous reactancethe sum of the stator winding leakage reac-tance and the q-axis magnetizing (armature)reactance of a synchronous machine. This re-actance represents the balanced steady-statevalue of the q-axis operational impedance ofthe synchronous machine.

quadrature axis transient reactance avalue that characterizes the equivalent re-actance of the q-axis windings of the syn-chronous machine between the initial timefollowing a system disturbance (subtransientinterval) and the steady state. This reactancecannot be directly mathematically related tothe q-axis operational impedance. However,in models in which the q-axis contains twoamortisseur windings and the rotor windingsare represented as lumped parameter circuits,the q-axis transient reactance is expressed inclosed form as the sum of the stator windingleakage reactance, and the parallel combina-tion of the q-axis magnetizing reactance and


the primary q-axis amortisseur winding leak-age reactances.

quadrature detector mixer in a FM re-ceiver where the output voltage is a functionof the original modulation of two carrier in-puts that are 90 degrees apart in phase.

quadrature distortion interference insignal transmission caused by phase error inthe reference carrier.

quadrature FM demodulator FM de-modulation of two carrier inputs that are 90degrees apart in phase.

quadrature hybrid a directional couplerthat accepts an input signal and delivers twoequal power outputs that are 90 degrees outof phase.

quadrature modulation a modulationscheme that involves the modulation of twosinusoidal carriers with a 90-degree phasedifference by two independent message sig-nals. The two carriers are typically rep-resented asA cos(ωct) (the in-phase car-rier) and A sin(ωct) (the quadrature car-rier), where ωc is the carrier frequency.The modulated signal is written ass(t) =Amc(t) cos(ωct+φ)+Ams(t) sin(ωct+φ),wheremc(t) andms(t) are the in-phase andquadrature modulating signals andφ is a ran-dom phase.

quadrature multiplexing a type of mul-tiplexing technique, in which two messagesignalsm1(t) andm2(t) are transmitted onthe same carrier frequency with two quadra-ture carriersAc cos 2πfct andAc sin 2πfct .That is, the quadrature multiplexed signal,also known as quadrature amplitude modu-lated signal, is

v(t) = Acm1(t) cos 2πfct

+ Acm2(t) sin 2πfct.

It is noted that the term quadrature indicatesa 90◦ phase difference between the two carri-ers. For the demodulation refer to quadrature

demodulation. The term quadrature mul-tiplexing is sometimes also referred to asquadrature carrier multiplexing.

quadrature phased signals two indepen-dent signals that have identical frequenciesand have a fixed 90-degree phase difference.

quadrature signal in quadrature modu-lation, the signal component that multipliessin 2πfct , wherefc is the carrier frequency.

quadrature spreader a device that takesan information signal and two independentPN sequences (two spreading codes) to pro-duce two baseband direct sequence spreadspectrum signals. These signals constitutethe modulation signals for a quadrature mod-ulation scheme.

quadrature-axis subtransient short-circuittime constant a constant that characterizesthe initial decay of transients in the q-axisvariables of the synchronous machine withthe stator windings short-circuited. The in-terval characterized is that immediately fol-lowing a disturbance, during which the ef-fects of all amortisseur windings are con-sidered. A detailed (derived) closed-formexpression for the subtransient short-circuittime constant of a machine with two q-axisamortisseur windings is obtained by takingthe reciprocal of the largest root of the numer-ator of the q-axis operational impedance. Anapproximate (standard) value is often used,in which it is assumed one amortisseur wind-ing resistance is small relative to the otherand the detailed expression simplified.

quadrature-axis transient open-circuittime constant a constant that charaterizesthe decay of transients in the q-axis variablesof the synchronous machine with the statorwindings open-circuited. The interval char-acterized is that following the subtransient in-terval, but prior to steady-state, during whichthe fastest q-axis rotor circuit dynamics havesubsided. A detailed (derived) closed-formexpression for the transient short-circuit time


constant of a machine with two q-axis amor-tisseur windings is obtained by taking the re-ciprocal of the smallest root of the denomina-tor of the q-axis operational impedance. Anapproximate (standard) value is often used, inwhich it is assumed one amortisseur windingresistance is infinite and the detailed expres-sion simplified.

quadrature-axis transient short-circuittime constant a constant that character-izes the decay of transients in the q-axis vari-ables of the synchronous machine with thestator windings short-circuited. The inter-val characterized is that following the sub-transient interval, but prior to steady-state,during which the transients of the fastest dy-namic q-axis rotor circuits have subsided. Adetailed (derived) closed-form expression forthe transient short-circuit time constant of amachine with two q-axis amortisseur wind-ings is obtained by taking the reciprocal ofthe smallest root of the numerator of the q-axis operational impedance. An approximate(standard) value is often used, in which it isassumed one amortisseur winding resistanceis infinite and the detailed expression simpli-fied.

quadrupole a magnet system that pro-duces a pair of dipoles with like poles oppo-site each other.

quadrupole illumination a type of off-axis illumination where four circles of lightare used as the source. These four circles arespaced evenly around the optical axis.

quadtree a quadtree represents struc-ture in an image by repeatedly subdividingsquares into quarters until each square is ho-mogeneous. Quadtrees can be used for fac-simile coding, for representing contiguousregions in primitive-based coding or for rep-resenting motion vector distributions. In fac-simile, where the image consists of black andwhite values only, three codewords are re-quired:S,B,W , meaning split-block, black-block and white-block. Starting from large

squares, each is analyzed and an appropriatecodeword generated. If the codeword isS,then the same process is applied recursivelyto the four quarter-size blocks.

quadword a data unit formed from fourwords.

qualification all activities that ensure thatthe nominal design and manufacturing spec-ifications will meet or exceed the reliabilitygoals.

qualitative robustness an estimationscheme in which we optimize performancefor the least favorable statistical environmentamong a specified statistical class.

quality factor (Q) measure of the persis-tence of damped oscillations in a resonator;2π times the energy stored in an oscillatingsystem divided by the energy lost in one cy-cle of the oscillations. Q is a measure of theprecision of the frequency selectivity of anelectrical resonator or filter. Q is dimension-less.

quality of service (QOS) a means of spec-ifying the level of service provided by a net-work or required by a user of the network.Typical QOS parameters include bandwidth,delay, jitter, and error rate.

quantization (1) the process of convert-ing amplitude values that can take on manydifferent values (infinitely many for analogsignals) into a finite (or more coarse) repre-sentation. Quantization is necessary in dig-ital processing of inherently analog signals,such as voice or image.See alsolossy sourcecoding, analog-to-digital conversion.

(2) aspects of particle properties related totheir wavelike nature as described by quan-tum mechanics.

(3) dividing a signal value into equallysized portions, known as quanta.

(4) conversion of a continuous range ofvalues into a number of discrete levels, for ex-ample in analog-to-digital conversion; con-


version of finely spaced discrete values intomore coarsely spaced values, e.g., in datacompression.

quantization matrix a matrix of num-bers in the JPEG compression scheme thatspecifies the amount by which each discretecosine transform coefficient should be re-duced. The numbers are based on the hu-man contrast sensitivity to sinusoids and aresuch that coefficients corresponding to fre-quencies that people are not very sensitive toare reduced or eliminated.See alsocontrastsensitivity, discrete cosine transform, JointPhotographic Experts Group, human visualsystem, psychovisual redundancy.

quantization of transform coefficientstechnique that exploits the statistical redun-dancy in transform coefficients and alsothe subjective redundancy in pictures usingtransform coding. Unlike DPCM, quantiza-tion of transform coefficients affects everypixel in the transformed block. There is aspreading of the quantization error, thus mak-ing the design of transform coefficients quan-tizers quite different from that of DPCM.

quantizer a device that performs quan-tization (sense (1)). A quantizer takes ananalog input signal and produces a discretevalued output signal. The number of possi-ble levels in the output signal typically deter-mines the error incurred in the quantizationprocess. The output of the quantizer is typi-cally encoded into a binary number.

quantizer characteristics the perfor-mance of a scalar quantizerQ is typicallycharacterized by its mean-squared quantiza-tion error

E[(Q(x)− x)2

]=∫(Q(x)−x)2fX(x)dx,

wherefX(x) represents the probability dis-tribution of the inputX. In the special casewhereX is uniformly distributed, and whereQ is a uniform quantizer (that is, the intervalbetween successive quantization levels is1,

andQ(x) is the rounding ofx to the near-est level), then the mean-squared quantiza-

tion error is12

12 .

quantum coherence a condition that ex-ists when a quantum mechanical system isin a linear combination of two or more ba-sis states. The polarization produced whena 2-level atom interacts with a resonant elec-tromagnetic field is an example of a quantumcoherence.

quantum dot a particle of semiconductormaterial having all dimensions comparableto an average free carrier Bloch wavelength.

quantum effect macro effect on carriermovement created by the confinement of thecarriers within very small dimensions alongthe direction of current flow (length). Whenthe length dimension starts to approach anappreciable portion of the mean-free carrierpath length, then odd things start to happen.This effect is somewhat analogous to the tran-sition between low RF and microwave theory.

quantum efficiency the ratio of the av-erage number of free electrons produced inthe photodetector per photon. Since this is aquantized process, some photons produce nofree electrons, while other photons produceone or more free electrons when their energyis absorbed in the photodetector. Quantumefficiency is usually denotedQe.

quantum electronics the field of studyof the interactions of electromagnetic fieldswith matter, with emphasis placed on inter-actions that cannot be described classically.

quantum hall effect quantization of theHall resistivity observed in two-dimensionalhigh-mobility semiconductor systems, causedby novel many-body effects.

quantum interference the enhancementor suppression of a transition in a quantummechanical system that arises when there are


multiple paths from the initial to the finalstate.

quantum limit the minimum number ofphotons that on average must be received inan optical communication system for a spec-ified error probability (usually 10−9). Us-ing on–off keying and an ideal photon count-ing receiver, the quantum limit is approxi-mately 10 photons/bit.

quantum mottle Seespeckle.

quantum numbers a group of integrat-ing constants in a solution of Schrodinger’sequation.

quantum well infrared photodetectormid- to long-wavelength photodetector ex-ploiting optically induced transitions be-tween confined quantum well states. Amen-able to focal plane array manufacture dueto compatibility with GaAs epitaxial growthand processing techniques.

quantum well laser a laser that uses quan-tum wells as the gain medium.

quantum wire a piece of semiconductorhaving two dimensions comparable to an av-erage free carrier Bloch wavelength, the otherdimension being much larger.

quarter-wave symmetry a periodic func-tion that displays half-wave symmetry and iseven symmetric about the 1/4 and 3/4 periodpoints (the negative and positive lobes of thefunction).

quarter-wave transformer a piece oftransmission line with an electrical length of1/4 of the operating frequency wave length(or odd multiple) used for impedance match-ing.

quasi-ballistic motion in a metal or semi-conductor in which the scattering is infre-quent, and for which the phase coherence is

mostly maintained during the time in whichthe carrier is in the active volume.

quasi-Fermi levels energy levels thatspecify the carrier concentration inside asemiconductor under nonequilibrium condi-tions.

quasi-linear nearly linear, easily approx-imated as linear with insignificant errors.Many circuits, elements and devices operatedunder small signal conditions are consideredquasi-linear for many parameters or appli-cations. A low noise amplifier under smallsignal operation still exhibits signal distor-tion (nonlinearity), but the gain characteris-tics can be determined from S-parameters,which are linear, with little or no error.

quasi-optics a low loss method of trans-mission and manipulation of millimeter andsub-millimeter waves.

quasi-TEM a mode of a structure com-prising multiple dielectrics that behaves likethe TEM mode of an equivalent structurecomprising a single material with an appro-priate effective dielectric constant. In cer-tain types of transmission lines, for exam-ple, microstrip lines, the propagating wavesare not pure TEM waves. Therefore, thegeneral static approximations cannot be ap-plied; however, in most practical cases onecan make an approximation by assuming thatthe fields are quasi-TEM. In this case, goodapproximation for the phase velocity, propa-gation constant, and impedance is obtained.

quasiquantum device electron device inwhich the principle of the device function ismore easily explainable by quantum mechan-ics than by Newtonian classical dynamics.

queue a data structure maintaining a first-in-first-out discipline of insertion and re-moval. Queues are useful in many situations,


particularly in process and event scheduling.See alsoFIFO memory.

quick gun a tool for installing crimpedwire connectors.

quincunx five points, four at the cornersof a square and one in the center.

quincunx lattice a horizontally and verti-cally repeated pattern of quincunxes, whichis identical to a square lattice oriented at45 deg.

quincunx sampling the downsampling ofa 2-D or 3-D signal on a quincunx lattice byremoving every even sample on every oddline and every odd sample on every even line.In the frequency domain, the repeat spectrumcenters also form a quincunx lattice. Quin-cunx sampling structures have been used forTV image sampling on the basis that theylimit resolution on diagonal frequencies butnot on horizontal and vertical frequencies.


RRS commonly used symbol for sourceimpedance.

RT commonly used symbol for transfor-mation ratio.

R-ALOHA Seereservation ALOHA.

RL Typical symbol for load resistance.

Rabi frequency the characteristic cou-pling strength between a near-resonant elec-tromagnetic field and two states of a quan-tum mechanical system. For example, theRabi frequency of an electric dipole allowedtransition is equal toµE/hbar, whereµ is theelectric dipole moment andE is the maxi-mum electric field amplitude. In a stronglydriven 2-level system, the Rabi frequency isequal to the rate at which population oscil-lates between the ground and excited states.

race condition a situation where multipleprocesses access and manipulate shared datawith the outcome dependent on the relativetiming of these processes.

raceway a channel within a buildingwhich holds bare or insulated conductors.

radar an instrument that transmits elec-tromagnetic waves and receives properties ofthe reflected electromagnetic wave from thetarget, which can be used to determine thenature and distance to the target. Radar is anacronym that stands for radio detection andranging.

radar cross section (RCS) a measure ofthe reflective strength of a radar target; usu-ally represented by the symbols, measuredin square meters, and defined as 4π times the

ratio of the power per unit solid angle scat-tered in a specific direction of the power unitarea in a plane wave incident on the scattererfrom a specified direction.

RADHAZ radiation hazards to personnelas defined in ANSI/C95.1-1991 IEEE Stan-dard Safety Levels with Respect to HumanExposure to Radio Frequency Electromag-netic Fields, 3 kHz to 300 GHz.

radial basis function network a fullyconnected feedforward network with a sin-gle hidden layer of neurons each of whichcomputes a nonlinear decreasing function ofthe distance between its received input anda “center point.” This function is generallybell-shaped and has a different center pointfor each neuron. The center points and thewidths of the bell shapes are learned fromtraining data. The input weights usually havefixed values and may be prescribed on thebasis of prior knowledge. The outputs havelinear characteristics, and their weights arecomputed during training.

radial intensity histogram a histogramof average intensities for a round object incircular bands centered at the center of theobject, with radial distance as the running in-dex. Such histograms are easily constructed,and, suitably normalised, form the basis forscrutinizing round objects for defects, andfor measuring radius and radial distances ofcylindrically symmetrical features.

radial system a network of straight wiresor other conductors radiating from the baseof a vertical monopole antenna that simulatesthe presence of a highly conducting groundplane beneath the antenna. Typically, ra-dial wires are approximately a quarter wave-length long and are arranged to have equian-gular spacing between them. The radial wireends at the base of the monopole are electri-cally bonded together and to one conductorof the feedline.


radiated emission an electromagneticfield propagated through space.

radiating aperture a basic element of anantenna that in itself is capable of effectivelyradiating or receiving radio waves. Typicalexample of a radiating apertures are a slot,and horn antennas or a truncated metallicwaveguide.

radiation the phenomenon by whichsources generate energy, which propagatesaway from them in the form of waves.

radiation boundary condition (RBC) aboundary condition that is imposed to trun-cate the computational domain of a differen-tial equation method so that it satisfies theSommerfield radiation condition.

radiation condition the condition thatspecifies the field behavior at infinity; in fact,for an unbounded domain it is necessary tospecify the field behavior on the surface atinfinity. By assuming that all sources arecontained in a finite region, onlyoutgoingwavesmust be present at large distances fromthe sources; hence the field behavior at largedistances from the sources must meet thephysical requirement that energy travel awayfrom the source region. This requirement isthe Sommerfield “radiation condition” andconstitutes a boundary condition on the sur-face at infinity. Let us denote withA anyfield component transverse to the radial dis-tancer and withk the free-space wavenum-ber. The transverse field of a sphericallydiverging wave in a homogeneous isotropicmedium decays as 1/r at large distancesrfrom the source region; locally, the spher-ical wave behaves like a plane wave travel-ling in the outwardr direction. As such, eachfield component transverse tor must behavelike exp(−jkr)/r; this requirement may bephrased mathematically as

limr→∞ r

(∂A

∂r+ jkA

)= 0

radiation efficiency in antenna theory, theratio of radiated power to the amount of inputpower to the antenna. Has a value between 0and 1, inclusive.

radiation intensity in antenna theory, afar-field quantity that is a function of anglethat gives the level of radiation in a specificdirection. Radiation intensity is the radialcomponent of the time average Poynting vec-tor with all the terms associated with the dis-tance from the antenna normalized out. Theunits are watts per square radian.

radiation pattern a plot of the far-field ra-diation as a function of the angle phi or thetawhile the other angle is held constant. A ra-diation pattern can be either polar or rectan-gular, and can be either logarithmic or linear.

radiative broadening a spectral linebroadening mechanism that arises due tospontaneous decay of the excited state.

radiative heat transfer the process bywhich long-wave electromagnetic radiationtransports heat from a surface to its surround-ings.

radio astronomy the study of celestial ob-jects based on the investigation of their radiofrequency electromagnetic waves spectrum.

radio frequency (RF) a general term usedto refer to radio signals in the general fre-quency range from thousands of cycles persecond (kHz) to millions of cycles per second(MHz). It is also often used generically andinterchangeably with the term microwave todistinguish the high frequency AC portion ofa circuit or signal from the DC bias signalor the downconverted intermediate frequency(IF) signal.

radio frequency choke an inductance (acoil of wire) intended to present a very highimpedance at radio frequencies. Such fre-quencies span the range of kilohertz to hun-dreds of megahertz.


radio frequency integrated circuit (RFIC)integrated circuit designed to operate at ra-

dio frequencies as amplifiers, mixers, oscil-lators, detectors or combinations of above.Typically, RFICs are configured for specificapplication to operate as a complete RF sys-tem.

radio frequency interference (RFI)electromagnetic phenomenon that either di-rectly or indirectly contributes to degrada-tion in the performance of a receiver or otherRF system, synonymous with electromag-netic interference.See alsoelectromagneticinterference.

radio horizon the maximum range, fromtransmitter to receiver on Earth’s surface, ofdirect (line-of-sight) radio waves. This isgreater than the optical horizon, because theradio waves follow a curved path as a resultof the continuous refraction it undergoes inthe atmosphere.

radio local loop Seewireless local loop.

radio waves electromagnetic radiationsuitable for radio transmission in the rangeof frequencies from about 10 kHz to about300 Mhz.

radiography an imaging modality thatuses an X-ray source and collimator to createa projection image. The image intensity isproportional to the transmitted X-ray inten-sity.

radiology the monitoring and control ofradioactivity, especially in regard to humanexposure, in a nuclear power plant.

radiometer a passive receiver that detectsenergy from a transmitting source or rera-diated energy from a target. Typically, ra-diometers are used in remote sensing appli-cations.

radius of curvature one of the parame-ters characterizing the reflecting surface of aspherical mirror.

radix the base number in a number sys-tem. Decimal (radix 10) and binary (radix 2)are two example number systems.

radix complement in a system that usesbinary (base 2) data negative numbers, can berepresented as the two’s complement of thepositive number. This is also called a truecomplement.

Radon transform the Radon transform ofa functionf (x, y), r(d, φ) is its line integralalong a line inclined at angleφ from theyaxis and at a distanced from the origin.

radwaste a contraction for "radioactivewaste," usually referring to mildly radioac-tive sludge removed from the coolant in anuclear reactor.

RAID Seeredundant array of inexpensivedisks.

RAKE receiver a receiver type in whichthe received signal from each (or a few domi-nating) resolvable propagation ray is individ-ually demodulated and subsequently com-bined into one decision variable accordingto some criterion. RAKE receivers are com-monly used in wideband transmission sys-tems such as spread-spectrum systems wherethe large bandwidth allows several rays to beresolved, thus creating a diversity gain (fre-quency diversity).

RAM Seerandom access memory.

RAM neuron a random access memorywith n inputs and a single output. The in-puts define 2n addresses and presentation ofa particular input vector allows the contentsof the 1-bit register at that address to be read,or to be written into. Training consists ofwriting 1s or 0s into the 1-bit registers, as


required for the various input vectors in thetraining set.

RAMA See resource auction multipleaccess.

ram’s head the top of a transmission linetower

Raman echo a type of photon echo inwhich the stimulated emission is assisted bya Raman transition.

Raman laser laser in which the amplifi-cation mechanism is considered to be Ramanscattering.

Raman scatter frequency-shifted lightscatter, utilized as powerful analytical chem-istry technique.

Raman scattering scattering of light bymeans of its interactions with the vibrationalresponse of a molecular system. The scat-tered light is typically shifted to lower fre-quencies ( See Stokes Law); the frequencyshift is equal to the vibrational frequencyof the molecule, typically 1013 to 1014Hz. In spontaneous Raman scattering (Seespontaneous light scattering), the scatteredlight is emitted in nearly all directions. Instimulated Raman scattering (Seestimulatedlight scattering), the scattered light is emittedin the form of an intense beam. The emittedbeam tends to be intense because it experi-ences large amplification by an amount givenby egIL, whereg is the gain factor of stimu-lated Raman scattering,I is the intensity ofthe incident laser beam in units of power perunit area, andL is the length of the interactionregion.

Raman, Sir Chandrasekhara Venkata(1888–1970) author of 500 articles and in-dependent investigator of light scatter andacoustics. Winner of a Nobel Prize for thediscovery of Raman scattering.

Raman–Nath diffraction in acousto-optics, the regime where many diffraction or-ders exist due to the thinness of the acousticgrating relative to the acoustic wavelength inthe direction of light.

Raman–Nath diffraction regime regimewhere the acoustic beam width is sufficientlynarrow to produce many diffracted beamswith significant power.

Raman–Nath mode acousto-optic spec-trum analyzer similar to the acousto-optic (AO) Bragg-mode spectrum analyzer,but uses illumination at normal incidence tothe face of the AO cell that results in multi-ple diffracted orders, with the two first orderbeams being used at the Fourier plane to ob-tain the input RF signal spectrum.

Raman–Nath scattering the scatteringof light from a periodically varying refrac-tive index variation in a thin medium, as con-trasted with Bragg scattering, for instance inthe operation of an acousto-optic modulator.

ramp a linear function of grey level,usually connecting two contrasting regions.Named after its appearance in one dimension,it is often used as one model of an edge.

Ramsey fringe the spectral feature gener-ated when a quantum mechanical transitionis excited by two identical-frequency, time-separated electromagnetic pulses. Ramseyfringes are used in cesium atomic beamclocks.

random access (1) term describing a typeof memory in which the access time to anycell is uniform.

(2) a method for allowing multiple usersto access a shared channel in which trans-missions are not coordinated (or perhaps arepartially coordinated) in time or frequency.

random access device See randomaccess.


random access memory (RAM) direct-access read/write storage in which each ad-dressable unit has a unique hardwired ad-dressing mechanism. The time to access arandomly selected location is constant andnot dependent on its position or on any previ-ous accesses. The RAM has a set ofkaddresslines (m = 2k), nbidirectional data lines, anda set of additional lines to control the direc-tion of the access (read or write), operationand timing of the device.

RAM is commonly used for the mainmemory of a computer and is said to be staticif power has to be constantly maintained inorder to store data and dynamic if periodicabsences of power do not cause a loss ofdata. RAM is usually volatile.See alsostaticrandom access memory, dynamic randomaccess memory, nonvolatile random-accessmemory.

random behavior response without aspectral or amplitude pattern or relationshipto the excitation. The excitation may be in-ternal, and may be thermal in nature.

random coding coding technique inwhich codewords are chosen at random ac-cording to some distribution on the codewordsymbols. Commonly a tool used in the devel-opment of information theoretic expressions.

random logic a digital system con-structed with logic gates and flip-flopsand other basic logic components intercon-nected in an non-specific manner.See alsomicroprogramming.

random process a mathematical pro-cedure for generating random numbers toa specific rule called a process,x whichis defined on continuousx(t), t ∈ Rn,or discretex(k), k ∈ Zn space / time.The value of the process at each point inspace or time is a random vector.See alsorandom variable, random vector. Seealsocorrelation, covariance, autocorrelation,autocovariance.

random replacement algorithm in acache or a paging system, an algorithm thatchooses the line or page in a random man-ner. A pseudo-random number generatorsmay be used to make the selection, or otherapproximate method. The algorithm is notvery commonly used, though it was used inthe translation buffers of the VAX11/780 andthe Intel i860 RISC processor.

random signal a signalX(t) that is eithernoiseN(t), an interfering signals(t), or asum of these:

X(t) = s1(t)+· · ·+sm(t)+N1(t)+· · ·+Nn(t)

random testing the process of testing us-ing a set of pseudo-randomly generated pat-terns.

random variable a continuous or discretevalued variable that maps the set of all out-comes of an experiment into the real line (orcomplex plane). Because the outcomes ofan experiment are inherently random, the fi-nal value of the variable cannot be predeter-mined.

random vector a vector (typically a col-umn vector) of random variables.See alsorandom variable, random process.

randomized decision rule a hypothesisdecision/classification rule that is not deter-ministic (that is, the measurement or obser-vation does not uniquely determine the de-cision). Although typically not useful givencontinuous observations, a randomized rulecan be necessary given discrete observations.See also receiver operating characteristicscurve.

range filter an edge detection filter thatfinds edges by taking the difference betweenthe maximum and minimum values in a lo-cal region of the image. The range filter alsoaccepts a weight mask the size of the localimage region that controls pixel values be-fore they enter the minimum and maximum


calculations. The weight mask allows edgesin certain directions to be searched for.

range image an image in which the in-tensity of pointx is a function of the distancebetweenx and the corresponding point in thescene (object) projected on the image plane.

range of Jacobian denotedJ , it is definedas a subspaceR(J ) inRr of the end-effector-velocities that can be generated by the jointvelocities, in the given manipulator posture.Seefull rank Jacobian.

rank filter an image transform used inmathematical morphology. Assume that toevery pixelp one associates a windowW(p)containing it. Letk be an integer> 1 whichis less than or equal to the size of each win-dowW(p). The rank filter with rankk andwindowsW(p) transforms an imageI intoa filtered imageI ′ whose gray-levelI ′(p)at pixelp is defined as thek-th least valueamong all initial gray-levelsI (q) for q in thewindowW(p). In a dual version, thek-thgreatest value is selected. When eachW(p)

is the translate byp of a structuring elementW of sizen, three particular cases are note-worthy:

1. k = 1: the rank filter is the erosion byW .

2. k = n: the rank filter is the dilation byW ,the symmetric ofW .

3. n is odd andk = (n + 1)/2: the rankfilter is a median filter.Seedilation, erosion,median filter, structuring element.

rare gas one of the rare gases specified inthe periodic table, He, Ne, Ar, Kr, Xe.

rare gas halides excimer moleculeformed by a reaction of a rare gas atom anda halogen atom: e.g., XeF, ArF, KrF.

rare gas molecule excimer moleculeformed by a reaction of an excited and a neu-tral rare gas atom, usually with a third atompresent. For example, Ar2, Kr2.

rare gas oxides an excimer moleculeformed by the reaction of a rare gas atomand an oxygen atom.

rare-earth magnet a magnet that has anyof the rare-earth elements in its composition.Typically stronger than other magnet materi-als, these include neodymium iron boron andsamarium cobalt.

rare-earth permanent magnet magnetmade of compounds of iron, nickel, andcobalt with one or more of the rare-earth el-ements such as samarium. These materialscombine the high residual flux density of thealnico-type materials with greater coercivitythan ferrites.

raster a predetermined pattern of scan-ning lines used to provide uniform coverageof the area used for displaying a televisionpicture.

raster coordinates coordinates in a dis-play system that specifically identify a phys-ical location on the display surface.

raster graphics a computer graphics sys-tem that scans and displays an image peri-odically in a raster, or left-to-right, top-to-bottom fashion.

raster image Seebitmapped image.

raster width (1) physical distance be-tween raster lines on a display surface andbetween distinguishable points in the sameraster line; the two distances are frequentlydifferent.

(2) the physical distance between rasterlines on a display surface and between dis-tinguishable points in the same raster line.The two distances are frequently different.

rate distortion function the minimumrate at which a source is represented by oneof a set of discrete points.


rate distortion theory a theory aimedat quantifying the optimum performance ofsource coding systems. Using informationtheory, for several source models and distor-tion measures, rate distortion theory providesthe optimum distortion function and the op-timum rate function. The distortion functionis optimum in that the distortion for a givenrate is the theoretical minimum value of dis-tortion for encoding the source at the given orlower rate. The rate function is optimum inthat the rate at a given distortion is the min-imum possible rate for coding the source atthe given or lower distortion.

rate equation approximation assump-tion, in a semiclassical model for the inter-action of light with atoms, that all fields andpopulations change negligibly within the co-herence time of the wave functions, loses allinformation about the phase of the fields andwave functions.

rate equation model model for the in-teraction of light with atoms in which theatoms are represented only by their popula-tions or population densities and the electro-magnetic field is represented only by its in-tensity, power, energy, or photon density.

rate equations coupled ordinary nonlin-ear differential equations governing the in-teraction of an electromagnetic field (repre-sented by an intensity, energy density, or pho-ton density) with an atomic or molecular lasermedium (represented by populations of theenergy states); phase information relating tothe fields and wavefunctions is absent fromthese equations.

rate split multiple access coding tech-nique for the multiple-access channel inwhich each user splits their informationstream into two or more streams, whichare independently encoded. These encodedstreams are multiplexed according to somerule prior to transmission. Used to show thattime-sharing is not required to achieve cer-

tain points in the capacity region of multipleaccess channels.

rate-adaptive digital subscriber line(RDSL) a digital subscriber line (DSL)in which the rates in each direction are ad-justed according to the quality of the channel.In general, longer loops are associated withlower rates.

rate-compatible punctured convolutional(RCPC) code one of a family of punc-tured convolutional codes derived from onelow-rate convolutional parent code by suc-cessively increasing the number of puncturedsymbols, given that the previously punc-tured symbols should still be punctured (rate-compatibility). These codes have applica-tions in, for example, variable error protec-tion systems and in hybrid automatic repeatrequest schemes using additional transmittedredundancy to be able to correctly decode apacket. Also called RCPC code.

rate-distortion theory Claude Shannon’stheory for source coding with respect to afidelity criterion, developed during the late1940s and the 1950s. Can be viewed asa generalization of Shannon’s earlier theory(late 1940s) for channel coding and infor-mation transmission. The theory applies tothe important methods for vector quantiza-tion and predicts the theoretically achievableoptimum performance.

rated voltage the voltage at which a powerline or electrical equipment is designed to op-erate.

ratio detector a circuit for recovering(demodulating) baseband information (usu-ally audio) from a frequency modulated (FM)wave.

rational function a function that is theratio of two polynomials. Rational func-tions often arise in the solution of differentialequations by Laplace transforms.See alsoLaplace transform.


ray one of a family of lines (rays) used torepresent the propagation of an electromag-netic wave; most useful in real media whenthe wave amplitude varies slowly comparedto the wavelength (see geometrical optics).

ray equation set of second-order differ-ential equations governing the trajectory of alight ray propagating along an arbitrary path.

ray optics approximate representation ofelectromagnetic wave propagation in termsof light rays, most useful in real media whenthe wave amplitude varies slowly comparedto the wavelength.

ray tracing (1) a high-frequency elec-tromagnetic analysis technique in which thepropagation path is modeled by flux lines or“rays.” The ray density is proportional tothe power density, and frequently, bundles ofthese rays are called ray tubes.

(2) a rendering technique in which thepaths of light rays reaching the viewpoint arecomputed to obtain realistic images. Givena 3-D description of a scene as a collectionof surfaces characterized by different opticalproperties, rays are traced backward from theviewpoint through the image plane until theyhit one of the surfaces or go off to infinity.

ray transfer matrix real two-by-two ma-trix governing the transformation of the raydisplacement and slope with respect to a fixedaxis.

Rayleigh criterion a method of distin-guishing between rough and smooth surfacesin order to determine whether specular re-flection will occur. A surface is consid-ered smooth if the phase difference betweenwaves reflected from the surface is less thanninety degrees.

Rayleigh distribution the probability dis-tribution of the magnitude of a complex quan-tity whose real and imaginary parts are inde-pendent Gaussian random variables with zeromean. Frequently used to approximate multi-

path fading statistics in non-line-of-sight mo-bile radio systems.

Rayleigh length distance over which thespot size of a Gaussian beam increases fromits value at the beam waist to a value 2.5larger; a measure of the waist size of a Gaus-sian beam,π times waist spot size squareddivided by wavelength; half of the confocalparameter.

Rayleigh noise the envelope of a zeromean, wide-sense stationary, narrowbandGaussian noise process. The probability den-sity function of a sinusoid in narrowbandnoise is a generalized Rayleigh distribution,

p(z) = zσe− z2

2σ2 , z ≥ 0. Also known as aRician distribution.

Rayleigh scattering (1) theory for the in-teraction between light and a medium com-posed of particles whose size is much smallerthan the wavelength. According to it, thescattering cross section is proportional to thefourth power of the wavelength of the scat-tered light. This explains both the red andblue colors of the sky.

(2) an intrinsic effect of glass that con-tributes to attenuation of the guided opticalwave. The effect is due to random localizedvariations in the molecular structure of theglass which acts as scattering centers.

Rayleigh–Ritz procedure a procedurefor solving functional equations.See alsomoment method.

Rayleigh-wing scattering of light thescattering of light with no change in centralfrequency, and with moderate (of the orderof 1011 Hz) broadening of spectrum of thelight. Rayleigh-wing scattering occurs whenlight scatters from anisotropic molecules.

RBC Seeradiation boundary condition.

RC time constant the time needed forsignal traveling from an end to the other endof a wire is constant when the wire and the


whole chip is scaled down. As the lengthof a wire shrinks by a factor ofk and thecross-sectional area of the wire is reduced bya factor ofk2, the capacitance of the wire de-creases by a factor ofk while the resistanceincreases by a factor ofk. The RC time re-mains constant, and thus the input chargingtime remains the same, independent of scal-ing. Consequently, the scaling down of thechip cannot increase the speed of the chip ifwire is used. Optical interconnects can speedup the chip.

RCPC code Seerate-compatible puncturedconvolutional code.

RCS Seeradar cross section.

RCT Seereverse conducting thyristor.

RDSL Seerate-adaptive digital subscriberline.

re-entrancy term describing the numberof times that a multiplex armature windingof a commutated machine closes upon itselfvia the commutator ring. For example, du-plex windings can be either singly or dou-bly re-entrant. In a doubly re-entrant du-plex winding, the ends of the two windingcircuits close only on themselves and noton each other, creating two distinct circuitsthrough the commutator and two distinct cir-cuit closures. Conversely, in a singly re-entrant duplex winding, the two windings areconnected in series through the commutatorring creating only a single circuit closure.

reachability a term that indicates that adynamical system can be steered from zeroinitial state to any final state in a given time in-terval. For many dynamical systems, reach-ability is equivalent to controllability. Thisis always true for linear finite-dimensionalcontinuous-time dynamical systems. How-ever, in discrete case, controllability may bethe stronger notion than reachability. In thiscase, the two concepts are equivalent if andonly if rankA = n. For dynamical systems

with delays, these two notions are essentiallydifferent. For infinite-dimensional systems,the relations between reachability and con-trollability depend on the properties of thesemigroupS(t) generated by the operatorA.

reactance grounded an electrical systemin which the neutral is intentionally groundedthrough a reactance. Frequently used inthe neutral of generators and transformers tolimit the magnitude of line to ground faultcurrents.

reactance modulator modulator nor-mally using phase or frequency modulationwhere the reactance of the circuit is depen-dent on changes in the input modulating volt-age.

reaction a functional in electromagneticsthat relates a set of fields and sources to oneanother. Reaction concepts are often used inthe discussion of field reciprocity.

reaction range sum of end-to-end round-trip delay and processing time.

reactive compensation process of coun-teracting the reactive component of a deviceby means of capacitors and inductors. Bothseries and shunt compensation are prevalent.

reactive congestion control in packet net-works, a congestion control system whoseactions are based on actual congestion oc-currence.

reactive ion etching the process of etch-ing materials by the use of chemically reac-tive ions or atoms. Typically, the reactiveions or atoms are generated in a RF plasmaenvironment or in a microwave discharge.

reactive load a load that is purely capac-itive or inductive.

reactive matching impedance transfor-mation achieved by employing a matching


network constructed of only reactive ele-ments.

reactive near field the region close to anantenna where the reactive components of theelectromagnetic fields from charges on theantenna structure are very large compared tothe radiating fields. Considered negligible atdistances greater than a wavelength from thesource (decay as the square or cube of dis-tance). Reactive field is important at antennaedges and for electrically small antennas.

reactive power (1) electrical energy perunit time that is alternately stored, then re-leased. For example, reactive power is as-sociated with a capacitor charging and dis-charging as it operates on an AC system.Symbolized byQ, with units of volt-amperesreactive (VAR), it is the imaginary part of thecomplex power.

(2) the power consumed by the reactivepart of the load impedance, calculated bymultiplying the line current by the voltageacross the reactive portion of the load. Theunits are vars (volt-ampere reactive) or kilo-vars.

reactor a container where the nuclear re-action takes place. The reactor converts thenuclear energy to heat.

reactor containment Seecontainmentbuilding.

reactor core an array of nuclear fuel rodsthat are arranged so as to encourage a chainreaction and thus heat water to supply a powerfor the steam turbine in a nuclear power plant.

reactor refueling the process of shuttingdown a nuclear reactor for maintenance andfuel replacement, typically every 12 to 24months.

read ahead on a magnetic disk, readingmore data than is nominally required, in thehope that the extra data will also be useful.

read instruction an assembly languageinstruction that reads data from memory orthe input/output system.

read phase the first portion of a trans-action during which the executing processobtains information that will determine theoutcome of the transaction. Any transactioncan be structured so that all of the input in-formation is obtained at the outset, all thecomputation is then performed, and finallyall results are stored (pending functionalitychecks based on the locking protocols in use).

read-after-writ ehazard See truedatadependency.

read-modify-write cycle a type of mem-ory device access that allows the contents at asingle address to be read, modified, and writ-ten back without other accesses taking placebetween the read and the write.

read-mostly memory memory primarilydesigned for read operations, but whose con-tents also can be changed through proceduresmore complex and typically slower than theread operations. EPROM, EEPROM, andflash memory are examples.

read-only (ROM) memory semiconduc-tor memory unit that performs only the readoperation; it does not have the write capabil-ity. The contents of each memory location isfixed during the hardware production of thedevice and cannot be altered. A ROM hasa set ofk input address lines (that determinethe number of addressable positions 2k) anda set ofn output data lines (that determinethe width in bits of the information stored ineach position). An integrated circuit ROMmay also have one or more enable lines forinterconnecting several circuits and make aROM with larger capacity. Plain ROM doesnot allow erasure, but programmable ROM(PROM) does. Static ROM does not require aclock for proper operation, whereas dynamicROM does.See alsorandom access memory,programmable read-only memory.


read/modify/write an uninterruptiblememory transaction in which information isobtained, modified, and replaced, under theassurance that no other process could haveaccessed that information during the trans-action. This type of transaction is importantfor efficient implementations of locking pro-tocols.

read/write head conducting coil thatforms an electromagnet, used to record onand later retrieve data from a magnetic cir-cular platter constructed of metal, plastic, orglass coated with a magnetizable material.During the read or write operation, the headis stationary while the platter rotates beneathit. The write mechanism is based on themagnetic field produced by electricity flow-ing through the coil. The read mechanism isbased on the electric current in the coil pro-duced by a magnetic field moving relative toit.

Less common are magnetoresistive heads,which employ noninductive methods forreading. A system that uses such a head re-quires an additional (conventional) head forthe writing. See alsodisk head, magneticrecording code.

real address the actual address that refersto a location of main memory, as opposedto a virtual address that must first be trans-lated. Also called a physical address.Seealsomemory mapping, virtual memory.

real power consider an AC source con-nected at a pair of terminals to an otherwiseisolated network. The real power, equal tothe average power, is the power dissipatedby the source in the network.

real-time refers to systems whose cor-rectness depends not only on outputs, butthe timeliness of those outputs. Failure tomeet one or more of the deadlines can re-sult in system failure.See alsosoft real-timesystem, firm real-time system, hard real-timesystem.

real-time clock a hardware counter thatrecords the passage of time.

real-time computing support for envi-ronments in which response time to an eventmust occur within a predetermined amount oftime. Real-time systems may be categorizedinto hard, firm, and soft real-time.

realization for a linear continuous or dis-crete stationary finite-dimensional dynami-cal system, a set of four constant matricesA,B,C,D of the state and output equations.The matrices may be calculated using certainalgorithms. The realization is said to be min-imal if the dimensionn of the square matrixA is minimal.

realization problem for 2-D Fornasini–Marchesini model a problem in con-trol. The transfer matrixT (z1, z2) of the 2-DFornasini–Marchesini model

xi+1,j+1 = A1xi+1,j + A2xi,j+1

+ B1ui+1,j + B2ui,j+1

yij = Cxij +Duij (1)

i, j ∈ Z+ (the set of nonnegative integers) isgiven by

T (z1, z2) = C [Iz1z2 − A1z1 − A2z2]−1

× (B1z1 + B2z2)+D (2)

where xij ∈ Rn is the local state vector,uij ∈ Rm is the input vector,yij ∈ Rp isthe output vector, andAk, Bk (k = 1,2) arereal matrices. Matrices

A!, A2, B1, B2, C,D (3)

are called a realization of a given transfermatrix T (z1, z2) if they satisfy (2). A re-alization (3) is called minimal if the ma-tricesA1 andA2 have minimal dimensionamong all realizations ofT (z1, z2). The(minimal) realization problem can be statedas follows. Given a proper transfer matrixT (z1, z2) ∈ Rp×m(z1, z2), find matrices (3)(with minimal dimension ofA1 andA2) thatsatisfy (2).


realization problem for 2-D Roesser modelthe transfer matrixT (z1, z2) of the 2-D

Roesser model[xhi+1,jxvi,j+1

]=[A1 A2A3 A4

][xhijxvij

]

+[B1B2

]uij (1)

yij = C

[xhijxvij

]+Duij

i, j ∈ Z+ (the set of nonnegative integers) isgiven by

T (z1, z2) = C

[In1z1 − A1 −A2

−A3 In2z2 − A4

]−1

×[B1B2

]+D (2)

wherexhij ∈ Rn1 andxvij ∈ Rn2 are the hori-zontal and vertical state vectors,uij ∈ Rm isthe input vector, andyij ∈ Rp is the outputvector. Matrices

A =[A1 A2A3 A4

], B =

[B1B2

], C, D (3)

are called a realization of a given transfer ma-trix T (z1, z2) if they satisfy (2). A realiza-tion (3) is called minimal if the matrixA hasminimal dimension amongst all realizationsofT (z1, z2). The (minimal) realization prob-lem can be stated as follows. Given a propertransfer matrixT (z1, z2) ∈ Rp×m(z1, z2),find matrices (3) (with minimal dimensionof A) that satisfy (2).

received signal strength indicator (RSSI)ratio of signal power level for a single fre-

quency or a band of frequencies to an es-tablished reference; the reference is typically1 mW, and the resultant value is expressed indecibels. RSSI is often used in mobile com-munications to make assessments such as towhich base station a call should be connectedor which radio channel should be used forcommunication.

receiver noise thermal (Boltzmann-type)noise in a receiver, a function of its physi-

cal temperature above absolute zero and thenoise bandwidth of the receiver’s electronicdevices. Receiver noise causes finite receiversensitivity.See alsothermal noise.

receiver operating characteristics curveplot of the probability of detection (likeli-hood of detecting the object when the ob-ject is present) versus the probability offalse alarm (likelihood of detecting the objectwhen the object is not present) for a particularprocessing system.

receiver sensitivity the minimum radiosignal power at the input to a receiver that re-sults in signal reception of some stated qual-ity.

reciprocity (1) a consequence of Maxwell’sequations, stipulating the phenomenon thatthe reaction of the sources of each of twodifferent source distributions with the fieldsgenerated by the other are equal, provided themedia involved have certain permeability andpermittivity properties (reciprocal media).Referring to reciprocal circuits, reciprocitystates that the positions of an ideal voltagesource (zero internal impedance) and an idealammeter (infinite internal impedance) canbe interchanged without affecting their read-ings.

(2) in antenna theory, the principal that thereceive and transmit patterns of an antennaare the same.

reciprocity in scattering law according towhich the source and detector points can beexchanged, providing the source amplitudeand phase are preserved.

reciprocity theorem in a network consist-ing of linear, passive impedances, the ratio ofthe voltage introduced into any branch to thecurrent in any other branch is equal in mag-nitude and phase to the ratio that results ifthe positions of the voltage and current areinterchanged.


recloser a self-contained device placed ondistribution lines that senses line currents andopens on overcurrent. Reclosing is employedto reenergize the protected line segment inthe case of temporary faults. Reclosers havethe capability for fast tripping for fuse sav-ing, and slow tripping to allow sectionalizingfuse operation for faults on laterals. The re-closer will retrip on permanent faults and goon to lockout. Reclosers are suitable for polemounting on overhead lines.

reclosing relay an auxiliary relay that ini-tiates circuit breaker closing in a set sequencefollowing fault clearing. Reclosing relays aretypically employed on overhead lines wherea high proportion of the faults are temporary.

recoil permeability the average slope ofthe minor hysteresis loop, which is roughlythe slope of the major hysteresis loop at zeroapplied field (H ), and is most often used todetermine the effect of applying and remov-ing a demagnetizing field to and from a mag-netic material.

recombination the process in which anelectron neutralizes a hole. Sometimes thisprocess causes light emission (i.e., throughradiative recombination), and sometimes itdoesn’t (i.e., through nonradiative recombi-nation).

recombination X-ray laser an X-raylaser made by gain from an inverted popu-lation where the upper level is inverted dueto recombining ions and electrons.

reconstruction the process of forming a3-D image from a set of 2-D projection im-ages. Also applies to the formation of a 2-D image from 1-D projections.Seeimagereconstruction, tomographyand computedtomography.

reconstruction from marker in a binaryimage, this is the operation extracting all con-nected components having a non-empty in-tersection with a marker. This operation can

be generalized to gray-level images by a mor-phological operator applying such a recon-struction on the gray-level slices of the im-age.

record unit of data, corresponding to ablock, sector, etc., on a magnetic disk, mag-netic tape, or other similar I/O medium.

recording code a line code optimized forrecording systems.See alsoline code.

recording density number of bits storedper linear inch on a disk track. In general, thesame number of bits are stored on each track,so that the density increases as one movesfrom the outermost to the innermost track.

recovery action that restores the state of aprocess to an earlier configuration after it hasbeen determined that the system has entered astate which does not correspond to functionalbehavior. For overall functional behavior, thestates of all processes should be restored in amanner consistent with each other, and withthe conditions within communication links ormessage channels.

rectangle detection the detection of rect-angle shapes, often by searching for cor-ner signals, or from straight edges presentin an image. Rectangle detection is impor-tant when locating machined parts in images,e.g., prior to robot assembly tasks.See alsopolygon detectionand square detection.

rectangular cavity a section of rectangu-lar waveguide closed on both ends by con-ducting plates.

rectangular window (1) in finite impulseresponse (FIR) filter design, the rectangu-lar window constituting the most straightfor-ward window function used usually as a ref-erence in studying other window functions.It is defined as 1 within an even interval cen-tered at the origin and 0 elsewhere.

(2) in image processing, an rectangulararea centered at a pixel under consideration.


This area is known as a window, or a mask,or a template. A square window of size 3×3is used most often.See alsoneighborhoodoperation.

rectifier a circuit that changes an AC volt-age to DC. Switching elements or diodes areused to create the DC voltage. Diode recti-fiers and thyristor rectifiers are the two mostcommonly used rectifiers.

recurrent coding old name for convolu-tional coding.

recurrent network a neural network thatcontains at least one feedback loop.

recursion the process whereby a programcalls itself.See alsorecursive procedure.

recursive equation a difference equationthat is of the form

y(k) =i=m∑i=0

aix(k − i)−i=n∑i=1

biy(k − i)

whereai andbi are some proper real con-stants. When all thebi = 0, it is called anonrecursive equation.

recursive filter a digital filter that is re-cursively implemented. That is, the presentoutput sample is a linear combination of thepresent and past input samples as well as thepreviously determined outputs. Tradition-ally, the term recursive filter is closely relatedto infinite impulse filter. In a nonrecursivefilter the present output sample is only a lin-ear combination of the present and past inputsamples.

recursive function Seerecursive procedure.

recursive method method that estimateslocal displacements iteratively based on pre-vious estimates. Iterations are performed atall levels, as in every pixel, each block of pix-els, along scanning line, from line to line, orfrom frame to frame.

recursive procedure a procedure that canbe called by itself or by another program thatit has called; effectively, a single process canhave several executions of the same programalive at the same time. Recursion providesone means of defining functions.

The recursive definition of the factorialfunction is the classic example:if n = 0

factorial(n)= 1else

factorial(n) = n * factorial (n−1)

recursive self-generating neural network(RSGNN) a recursive version of self-generating neural network (SGNN) that candiscover recursive relations in the trainingdata. It can be used in applications suchas natural language learning/understanding,continuous spoken language understanding,and DNA clustering/classification, etc.Seealsoself-generating neural network.

Red book SeeIEEE Color Books.

red head another name for a hot tap.

reduced characteristic table a tabularrepresentation used to illustrate the operationof various bistable devices.

reduced instruction set computer (RISC)processor relatively simple control unitdesign with a reduced menu of instructions(selected to be simple), data and instructionsformats, addressing modes, and with a uni-form streamlined handling of pipelines.

One of the particular features of a RISCprocessor is the restriction that all memoryaccesses should be by load and store instruc-tions only (the so called load/store architec-ture). All operations in a RISC are register-to-register, meaning that both the sources anddestinations of all operations are CPU regis-ters. All this tends to significantly reduceCPU to memory data traffic, thus improv-ing performance. In addition, RISCs usuallyhave the following properties: most instruc-tions execute within a single cycle, all in-


structions have the same standard size (32bits), the control unit is hardwired (to in-crease speed of operations), and there is aCPU register file of considerable size (32 reg-isters in most systems, with the exception ofSPARC with 136 and AMD 29000 with 192registers).

Historically, the earliest computers ex-plicitly designed by these rules were designsby Seymour Cray at CDC in the 1960s. Theearliest development of the RISC philosophyof design was given by John Cocke in the late1970s at IBM. However, the term RISC wasfirst coined by Pattersonet al. at the Uni-versity of California at Berkeley to describea computer with an instruction set designedfor maximum execution speed on a particularclass of computer programs. Patterson andhis team of researchers developed the firstsingle-chip RISC processor.

Compare withcomplex instruction setcomputer.

reduced-order model a mathematicalrepresentation of a system that is obtained byneglecting portions of a more explicit (de-tailed) model. In large-scale power systemanalysis, this term is typically used to indi-cate a model derived by neglecting the elec-tric transients in the stator voltage equationsof all machines and in the voltage equationsof power system components connected tothe stators of the machines.

reduced-voltage motor starter a de-vice designed to safely connect an electricmotor to the power source while limitingthe magnitude of its starting current. Var-ious electromechanical configurations maybe used: primary resistor, delta-wye, part-winding (requires special motor or dual volt-age windings). Power electronic devices mayalso be utilized to gradually increase the ap-plied voltage to system levels. The completestarter must also include fault and overloadprotection.

redundancy (1) the use of parallel or se-ries components in a system to reduce the

possibility of failure. Similarly, referringto an increase in the number of componentswhich can interchangeably perform the samefunction in a system. Sometimes it is referredto as hardware redundancy in the literature todifferentiate from so called analytical redun-dancy in the field of FDI (fault detection andisolation/identification). Redundancy can in-crease the system reliability.

(2) in robotics, the numbern degreesof mobility of the mechanical structure, thenumberm of operational space variables,and the numberr of the operational spacevariables necessary to specify a given task.Consider the differential kinematics mappingv = J (q)q in whichv is(r×1) vector of end-effector velocity of concern for the specifictasks andJ isr×n Jacobian matrix. Ifr < n,the manipulator is kinematically redundantand has(n−r) redundant degrees of mobility.Manipulator can be redundant with respect toa task and nonredundant with respect to an-other.See alsoredundant manipulator.

redundancy encoding any digital encod-ing scheme which takes advantage of redun-dancy in the digital signal. For example, inrun-length encoding, a gray scale digital im-age is represented by the gray level of a pixeland the number of times adjacent pixels withthat gray level appear. So an image contain-ing large regions of a single gray level can berepresented with a great reduction in digitalinformation.

redundancy statistics model refers tostatistical similarities such as correlation andpredictability of data. Statistical redundancycan be removed without destroying any in-formation.

redundancy-free channel coding refersto methods for channel robust source codingwhere no “explicit” error protection is intro-duced. Instead, knowledge of the source andsource code structure is utilized to counteracttransmission errors (for example, by meansof an efficient index assignment).


redundant array of inexpensive disks(RAID) standardized scheme for multiple-disk data base systems viewed by the operat-ing system as a single logical drive. Data isdistributed across the physical drives allow-ing simultaneous access to data from multipledrives, thereby reducing the gap between pro-cessor speeds and relatively slow electrome-chanical disks. Redundant disk capacity canalso be used to store additional information toguarantee data recoverability in case of diskfailure (such as parity or data duplication).The RAID scheme consists of six levels (0through 5), RAID0 being the only one thatdoes not include redundancy.

redundant manipulator the manipulatoris called redundant if more degrees of mo-bility are available than degrees of freedomrequired for the execution of a given task.Seealsodifferential kinematics, redundancy.

redundant number system the systemin which the numerical value could be repre-sented by more than one bit string.

Reed switch a magnetomechanical devicecomposed of two thin slats of ferromagneticmaterial within a hermetically sealed capsulethat attract each other when an external mag-netic field (from an electromagnet or perma-nent magnet) induces opposite poles at theoverlapping ends of both slats.

Reed–Solomon code an extension ofBAH codes to nonbinary alphabets devel-oped by Driving Reed and Gustave Solomonindependently of the work by Bose, Chaud-huri, and Hocquenghem. Arguably, the mostwidely used of any forward error controlcode.

reentrancy the characteristic of a blockof software code that, if present, allows thecode in the block to be executed by more thanone process at a time.

reentrant a program that uses concur-rently exactly the same executable code in

memory for more than one invocation of theprogram (each with its own data), rather thanseparate copies of a program for each invo-cation. The read and write operations mustbe timed so that the correct results are alwaysavailable and the results produced by an in-vocation are not overwritten by another one.

reference black level picture signal levelcorresponding to a specified maximum limitfor black peaks.

reference frame Seebase frame.

reference impedance impedance towhich scattering parameters are referenced.

reference matrix a triangular array of bitsused to implement the least recently used al-gorithm in caches. When theith line is refer-enced, all the bits in theith row are set to a 1and then all the bits in theith column are setto a 0. Having 0s in thejth row and 1s in thejth column identifies thejth line as the leastrecently used line.

reference monitor a functional modulethat checks each attempt to access memoryto determine whether it violates the system’ssecurity policy, intercepting it if a violationis imminent. The memory management unitcan provide this service, provided that theaccess control information contained there isknown to be consistent with the security pol-icy.

reference node one node in a network thatis selected to be a common point, and all othernode voltages are measured with respect tothat point.

reference point Seeset point.

reference white in a color matching pro-cess, a white with known characteristics usedas a reference. According to the trichro-matic theory, it is possible to match an arbi-trary color by applying appropriate amountsof three primary colors.


reference white level picture signal levelcorresponding to a specified maximum limitfor white peaks.

reflectance the physical property of mat-ter to reflect light, defined (punctually) as theratio between the intensityI (x) of a pointxof an image and the irradianceL(x) in thesame point.

reflected power power in the reflectedpart of an electromagnetic wave.

reflected wave the result that ensues whena high-speed electromagnetic wave reachesthe end of a transmission line, when the lineis not terminated with an impedance match-ing the surge impedance of the line. When asurge reaches an open circuited line terminal,the reflected voltage wave equals the incidentvoltage wave, resulting in a doubling of thelevel of the voltage surge at that point.

reflection in electromagnetic wave propa-gation, the change in direction of propagationof a plane wave due to the wave being inci-dent on the surface of a material. Typically,the effect is greater in the case of a materialthat has a high electrical conductivity.

reflection coefficient (1) the ratio of thereflected field to the incident field at a mate-rial interface.

(2) another way of expressing the impe-dance. The reflection coefficient is definedas how much signal energy would be re-flected at a given frequency. Like impedance,the reflection coefficient will vary with fre-quency if inductors or capacitors are in thecircuit. The reflection coefficient is alwaysdefined with respect to a reference or charac-teristic impedance (= (Z − Z0)/(Z + Z0)).For example, the characteristic impedanceof one typical TV transmission line is 75ohms, whereas another type of TV trans-mission line has a characteristic impedanceof 300 ohms. Hooking up a 75-ohm trans-mission line to a 300-ohm transmission linewill result in a reflection coefficient of value

(300− 75)/(300+ 75) = 0.6, which meansthat 60% of the energy received from the an-tenna.

reflection grating a diffraction gratingthat operates in reflection, i.e., the diffractedlight is obtained by reflecting off the grating.

reflective notching an unwanted notchingor feature size change in a photoresist patterncaused by the reflection of light off nearbytopographic patterns on the wafer.

reflectivity a property that describes thereflected energy as a function of the incidentenergy of an EM wave and a material body.The property may be quantified in terms ofthe magnitude of the reflection coefficient orthe ratio of the incident to the reflected field.

reflectometer instrument that measuresreflected power.

reflector antenna an antenna comprisedof one or more large reflecting surfaces usedto focus intercepted power around a smallregion where a primary feed antenna is thenused to input the power to the receiver elec-tronics. Also used to transmit power in nar-row angular sectors.

reflex klystron a high-power microwavetube oscillator.

refraction the process undertaken byan electromagnetic wave wherein the wavechanges direction of propagation as it is in-cident on the edge of a material. The waveundergoes a “bending” action, sometimes re-ferred to as knife edge refraction, and the“bending” angle is less than 90 degrees. Re-fraction may also occur as a wave propagatesthrough a media such as the atmosphere.

refractive index a parameter of a mediumequal to the ratio of the velocity of propaga-tion in free space to the velocity of propaga-tion in the medium. It is numerically equal tothe square root of the product of the relative


permittivity and relative permeability of themedium.See alsoindex of refraction.

refractivity the refractive index minus 1.

refractory period a period of time af-ter the initiation of an action during whichfurther excitation is impossible (absolute re-fractory period) or requires a greater stimulus(relative refractory period).

refresh refers to the requirement that dy-namic RAM chips must have their contentsperiodically refreshed or restored. Without aperiodic refresh, the chip loses its contents.Typical refresh times are in the 5–10 millisec-ond range.See alsomemory refresh.

refresh cycle (1) a periodically repeatedprocedure that reads and then writes back thecontents of a dynamic memory device. With-out this procedure, the contents of dynamicmemories will eventually vanish.

(2) the period of time taken to “refresh” aportion of a dynamic RAM chip’s memory.See alsorefresh.

refresh period the time between the be-ginnings of two consecutive refresh cyclesfor dynamic random access memory devices.

regeneration the process of returning en-ergy back into a system during a portion ofthe machine’s operating cycle.

regeneration loop a water purificationsystem used to maintain proper conditions ofthe cooling liquid for a power vacuum tube.

regenerative braking a method for ex-tracting kinetic energy from the load, con-verting it back to electricity, and returning itto the supply. Used widely in electric traindrives and electric vehicles.

region growing the grouping of pixelsor small regions in an image into larger re-gions. Region growing is one approachto image segmentation.See alsodilation,

erosion, image segmentation, mathematicalmorphology.

region of absolute convergence the setof complex numberss for which the mag-nitude of the Laplace transform integral isfinite. The region can be expressed as

σ+ < Re(s) < σ−

whereσ+ andσ− denote real parameters thatare related to the causal and anticausal com-ponents, respectively, of the signal whoseLaplace transform is being sought.Re(s)represents the real part ofs.

region of asymptotic stability Seeregionof attraction.

region of attraction the region around anequilibrium state of a system of differentialor difference equations such that the trajec-tories originating at the points in the regionconverge to the equilibrium state. Trajecto-ries starting outside the region of attraction ofthe given equilibrium state may “run away”from that equilibrium state.

region of convergence (ROC) (1) an areaon a display device where the image dis-played meets an accepted criteria for rastercoordinate deviation.Seeregion of absoluteconvergence.

(2) the set of complex numberss = a +jb for which the magnitude of the Laplacetransform integral is finite. The region canbe expressed as:

σ− < Real(s) < σ+,

whereσ− andσ+ denote real parameters thatare related to the causal and anticausal com-ponents, respectively, for the signal whoseLaplace transform is being sought.

(3) the set of complex numbersz = est

for which the magnitude of thez-transformsum is finite. The region can be expressedas:

R− < |z| < R+,


whereR− andR+ denote real parameters thatare related to the causal and anticausal com-ponents, respectively, for the signal whosez-transform is being sought.

region of interest (ROI) a restricted setof image pixels upon which image process-ing operations are performed. Such a set ofpixels might be those representing an objectthat is to be analysed or inspected.

region of support the region of variableor variables where the function has non-zerovalue.

register a circuit formed from identicalflip-flops or latches and capable of storingseveral bits of data.

register alias table Seevirtual register.

register direct addressing an instructionaddressing method in which the memory ad-dress of the data to be accessed or stored isfound in a general purpose register.

register file a collection of CPU registersaddressable by number.

register indirect addressing an instruc-tion addressing method in which the registerfield contains a pointer to a memory locationthat contains the memory address of the datato be accessed or stored.

register renaming dynamically allocat-ing a location in a special register file for aninstance of a destination register appearing inan instruction prior to its execution. Used toremove antidependencies and output depen-dencies.See alsoreorder buffer.

register transfer notation a mathemat-ical notation to show the movement of datafrom one register to another register by usinga backward arrow. Notation used to describeelementary operations that take place duringthe execution of a machine instruction.

register window in the SPARC architec-ture, a set or window of registers selected outof a larger group.

registration the process of aligning mul-tiple images obtained from different modali-ties, at different timepoints, or with differentimage acquisition parameters.Seefusion.

regression the methods that use backwardprediction error as input to produce an esti-mation of a desired signal. Quantitatively,the regression ofy onX, denoted byr(y), isdefined as the first conditional moment, i.e.,

r(y) = E(X|y).

regular controllability a dynamical sys-tem is said to be regularly controllable in timeinterval[t0, t1] if every dynamical system ofthe form

x′(t) = Ax(t)x(t)+ bjuj (t)

j = 1,2, . . . , m

is controllable wherebj is the j th columnof the matrixB anduj (t) is thej th scalaradmissible control.

regular cue any regular recurringpoint/element of a signal that can be used tosignal the start of a new signal sequence; e.g.,the leading edge of a 60-Hz square wave is aregular cue.

regular form a particular form of thestate space description of a dynamical sys-tem. This form is obtained by a suitabletransformation of the system state. The reg-ular form is useful in control design.

regularization a procedure to add a con-straint term in the optimization process thathas a stabilizing effect on the solution.

regulation the change in voltage from no-load to full-load expressed as a percentage offull-load voltage.


regulator a controller designed to main-tain the state of the controlled variable ata constant value, despite fluctuations of theload.

reinforcement learning learning on thebasis of a signal that tells the learning systemwhether its actions in response to an input (orseries of inputs) are good or bad. The signalis usually a scalar, indicating how good orbad the actions are, but may be binary.

rejection criteria criteria such as poorsurface texture, existence of scratch marks,and out-of-tolerance distance measures, whichconstitute reasonable grounds for rejecting aproduct from a product line.

relational model a logical data struc-ture based on a set of tables having commonkeys that allows the relationship between dataitems to be defined without considering thephysical database organization.

relative addressing an addressing mech-anism for machine instructions in which theaddress of the target location is given by thecontents of a specific register and an offsetheld as a constant in the instruction addedtogether. See alsoPC-relative addressing,index register, base address.

relative-address coding in facsimile cod-ing, represents the transition between levelson a particular scan line relative to transi-tions on the preceding scan line. A relative-address coding system has a pass mode code-word for indicating where a pair of transi-tions on the previous line does not have cor-responding transitions on the current line anda runlength coding mode applied when thereis no nearby suitable transition on the previ-ous line. CCITT Group IV facsimile uses aform of relative-address coding.

relative controllability a dynamical sys-tem with delays in which for a given timeinterval [0, t1] if for any initial condition(x(0), x0) and any final vectorx1 ∈ Rn there

exists an admissible controlu(t), t ∈ [0, t1],such that the corresponding trajectory satis-fies the condition

x (t1, x(0), x0, u) = x1

relative entropy information theoreticquantity representing the “distance” betweentwo probability distribution functions. Alsoknown as the Kullback–Liebler distance. Fortwo probability mass functionsp(x) andq(x), the relative entropyD(p||q) is givenby

D(p||q) =∑

p(x) logp(x)

q(x).

This quantity is only a pseudo-distance, asD(p||q) 6= D(q||p).relative intensity noise noise resultingfrom undesirable fluctuations of the opticalpower detected in an optical communicationsystem.

relative permeability the complex per-meability of a material divided by the per-meability of free space:µr = µ/µ0.

relative permittivity the complex permit-tivity of a material divided by the permittivityof free space:εr = ε/ε0.

relative refractive index difference theratio (n2

1 − n22)/2n

21 ≈ (n1 − n2)/n1 where

n1 > n2 andn1 andn2 are refractive indices.

relative sensitivity denotedSyx(y, x) anddefined as follows:

Sy

x(y, x) = ∂y

∂x

x

y= ∂y/y

∂x/x= ∂ ln y

∂ ln x

It is usually used to establish the approximaterelationship between the relative changesδy = 1y/y andδx = 1x/x. Here1y and1x are absolute changes. If these relativechanges are small, one writes that

δy ≈ Sy

xδx


This relationship assumes thatSyx is different

from zero. IfSyx = 0, the relative changesδy

andδx may be independent. The propertiesof relative sensitivity established by differen-tiation only are tabulated and may be foundin many textbooks and handbooks.See alsoabsolute sensitivity, sensitivity, sensitivitymeasure, semi-relative sensitivity.

relaxation (1) a general computationaltechnique where computations are iterateduntil certain parameter measurements con-verge to a set of values.

(2) the response of a linear time invari-ant system can be represented as the sum ofthe zero-input response (system response toa zero input function) plus the zero-state re-sponse (system response to an input func-tion when the system is in the zero state).Relaxation is the process of putting a sys-tem into its zero-state, i.e., all initial condi-tions are zero and there are no internal energystores. A system is considered relaxed if it isin the zero state.See alsorelaxation labeling,optimization.

relaxation labeling an iterative mathe-matical procedure in which a system of val-ues is processed, e.g., by mutual adjustmentof adjacent or associated values, until a stablestate is attained. Especially useful for achiev-ing consistent optimal estimates of pixel in-tensities or deduced orientation values forpoints on the surface of an object.See alsorelaxation, optimization.

relaxation oscillations the damped out-put oscillations that occur in some laser os-cillators when they are perturbed from steadystate.

relaxation time the time in which the ini-tial distribution of charge will diminish to 1/eof its original value.

relay a device that opens or closes a con-tact when energized. Relays are most com-monly used in power systems, where theirfunction is to detect defective lines or appa-

ratus or other abnormal or dangerous occur-rences and to initiate appropriate control ac-tion. When the voltage or current in a relayexceeds the specified “pickup” value, the re-lay contact changes its position and causesan action in the circuit breaker. A decision ismade based on the information from the mea-suring instruments and relayed to the trip coilof the breaker, hence the name “relay.” Otherrelays are used as switches to turn on or offequipment.

relay channel a multiterminal channel inwhich the receiver observes the transmittedsignal through two channels: one direct tothe transmitter, the other via an intermediatetransmitter/receiver pair.

reliability the probability that a compo-nent or system will function without failureover a specified time period, under statedconditions.

reliability criteria a set of operating con-ditions that the system operator adheres

to in order to guarantee secure operation.

relocatability the capability for a programto be loaded into any part of memory that isconvenient and still execute correctly.

relocation register register used to facili-tate the placement in varying locations of dataand instructions. Actual addresses are calcu-lated by adding program-given addresses tothe contents of one or more relocation regis-ters.

reluctance the resistance to magnetic fluxin a magnetic circuit; analogous to resistancein an electrical circuit.

reluctance motor a motor constructed onthe principle of varying reluctance of the airgap as a function of the rotor position withrespect to the stator coil axis. The torquein these motors arises from the tendency ofthe rotor to align itself in the minimum re-


luctance position along the length of the airgap.

reluctance torque the type of torque a re-luctance machine’s operation is based upon.A reluctance torque is produced in a magneticmaterial in the presence of an external mag-netic field, which makes it to line up with theexternal magnetic field. An induced field dueto fringing flux develops a torque that even-tually twists the magnetic material around toalign itself with the external field.

remanent coercivity the magnetic fieldrequired to produce zero remanent magne-tization in a material after the material wassaturated in the opposite direction.

remanent magnetism Seeresidual

remanent polarization the residual or re-manent polarization of a material after an ap-plied field is reduced to zero. If the materialwas saturated, the remanent value is usuallyreferred to as the polarization, although evenat smaller fields a (smaller) polarization re-mains.

remanence (1) in a ferromagnetic mate-rial, the value of the magnetic flux densitywhen the magnetic field intensity is zero.

(2) the magnetic induction (B) of a magnetafter the magnetizing field is removed andan air gap (hence self-demagnetizing field)is introduced to the magnetic circuit. Alsocalled retentivity or residual induction.

remote sensing the use of radar, satelliteimagery, or radiometry to gather data abouta distant object. Usually, the term refers tothe use of microwaves or millimeter wavesto map features or characteristics of planetarysurfaces, especially the Earth’s. Applicationsinclude military, meteorological, botanical,and environmental investigations.

remote terminal unit (RTU) hardwarethat gathers system-wide real-time data from

various locations within substations and gen-erating plants for telemetry to the energymanagement system.

removable disk disk that can be removedfrom disk drive and replaced, in contrast witha non-removable disk, which is permanentlymounted.See alsoexchangeable disk.

rename register Seevirtual register.

rendering (1) the preparation of the repre-sentation of an image to include illumination,shading, depth cueing, coloring, texture, andreflection.

(2) common techniques include Phongand Gouraud shading; more complex render-ing models such as raytracing and radiosityemphasize realistic physics models for cal-culating light interactions and texture inter-actions with objects.

renewable fuse a fuse consisting of areusable cartridge and a fusible element thatcan be replaced.

reorder buffer a set of storage locationsprovided for register renaming for holdingresults of instructions. These results maybe generated not in program order. At somestage, the results will be returned to the truedestination registers.

repeatability the ability of a sensor to re-produce output readings for the same value ofmeasurand, when applied consecutively andunder the same conditions.

repeater electromagnetic device that re-ceives a signal and amplifies it and retrans-mits it. In digital systems, the signal is re-generated.

repetition coding the simplest form of er-ror control coding. The information symbolto be transmitted is merely repeated an un-even number of times. A decision regardingthe true value of the symbol transmitted is


Unknown

magnetism.

then simply made by deciding which symboloccurred the greatest number of times.

representation singularity of φ a set oforientations for which the determinant of thetransformation matrix in the analytical Jaco-bian vanishes.See alsoanalytical Jacobian.

representative level one of the discreteoutput values of a quantizer used to representall input values in a range about the represen-tative level.See alsodecision level.

reprocessing the recycling of reactor fuelby separation into fissile, non-fissile, radioac-tive and non-radioactive components suchthat wastes can be isolated and fissile ma-terial re-used to make more reactor fuel.

repulsion–induction motor a single-phase motor designed to start as a repulsionmotor, then run as an induction motor. Therotor has a DC-type winding with brushesshorted together, in addition to the normalsquirrel cage winding. Although it is an ex-pensive design, it provides excellent start-ing torque with low starting current (similarto a universal motor) and relatively constantspeed under load.

requirements analysis a phase of soft-ware development life cycle in which thebusiness requirements for a software productare defined and documented.

reservation ALOHA (R-ALOHA) aclass of multiple access control protocol inwhich the user transmits the first packet inrandom access fashion. If successful, theuser will have a fixed part of the channel ca-pacity allocated.

reservation station storage locationsplaced in front of functional units and pro-vided to hold instructions and associatedoperands when they become available. Usedin a superscalar processor.

reset scrambling a technique of random-izing a source bit sequence by adding thesequence to a pseudo-random bit stream us-ing element-wise modulo-2 arithmetic. Thesource sequence is recovered at the de-coder through addition of the demodulatedbit stream with the same pseudo-random se-quence. The pseudo-random sequences areusually generated with linear feedback shiftregisters which must be aligned with a re-set or framing signal in order to recover thesource sequence accurately.

reset time for a line recloser or cir-cuit breaker, a time begins when the devicesuccessfully recloses following a temporaryfault. After the reset time elapses, the faultcycle is considered over, and any subsequentfault will be treated as the first fault in a newcycle.

residential loop two radial sources com-ing in to an open point used for switching.

residual current circuit breaker Euro-pean term for ground fault interrupter.

residual error the degree of misfit be-tween an individual data point and somemodel of the data. Also called a “residual”.

residual inductance Seeremanence.

residual magnetism a form of permanentmagnetism, referring to the flux remaining ina ferromagnetic material after the MMF thatcreated the flux is removed. For example,if a bar of steel is surrounded by a coil andcurrent is applied to the coil, the steel bar willcreate a magnetic field due to the rotation ofthe domains in the steel. After the current isremoved, some of the domains will remainaligned, causing magnetic flux.

residual overcurrent relay an overcur-rent relay that is connected to sense resid-ual current. Residual current is the sum ofthe three phase currents flowing in a currenttransformer secondary circuit, and is propor-


tional to the zero sequence current flowing inthe primary circuit at that point.

residual pyramid Seepredictive pyramid.

residual vector quantization Seeinter-polative vector quantization.

resistivity (1) the product of the resistanceof a given material sample times the ratio ofits cross-sectional area to its length.

(2) an electrical material property de-scribed by a tensor constant indicating theimpedance of free electron flow in the ma-terial. Resistivity relates the electric fieldstrength to the conduction current, and canbe expressed as the inverse of the conductiv-ity.

resistance ratio of the potential of an elec-trical current applied to a given conductor tothe current intensity value.

resistance ground a grounding scheme inwhich the neutral of Y-connected machinesis connected to ground through a resistancesuch that ground-fault currents are limited.

resistance grounded Seelow resistancegrounded system, high-resistance groundedsystem.

resistive mixer a device used to convertmicrowave frequencies to intermediate fre-quencies. Depending on the frequency ofmicrowaves to be converted, the intermedi-ate frequencies can be UHF, VHF, or HF.

resolution (1) the act of deriving from asound, scene, or other form of intelligence,a series of discrete elements from which theoriginal may subsequently be reconstructed.The degree to which nearly equal values of aquantity can be discriminated.

(2) the fineness of detail in a measurement.For continuous systems, the minimum incre-ment that can be discerned.

(3) the ability to distinguish between twounits of measurement.

(4) the number of pixels per linear unit (orper dimension) in a digital image.

(5) the smallest feature of a given type thatcan be printed with acceptable quality andcontrol.

(6) refers to the ability to resolve two pointtargets which are closely spaced in time orfrequency. For a linear system, resolutioncan be measured in terms of the width of theoutput pulse produced by a point target. Inimaging applications, resolution refers to thenumber of independently detectable, control-lable or displayable points.

resolvable component a component of asignal that exists in a group of components (asum of signals) such that the amplitude of thecomponent can be approximately determinedby correlation of the overall signal with acomponent signal of unit amplitude. Typi-cally, the different signal components are de-layed versions of the same signal received ata receiver in a channel with multipath prop-agation. The different delay components arethen resolvable if the relative delay betweenany two components is greater than the re-ciprocal of the transmitted signal bandwidth.

resolver an electric machine that is usedto provide information about the position ofa motor-driven system. The shaft of the re-solver is connected to the main motor eitherdirectly or through gears. The rotor containsa single winding, while the stator typicallycontains two windings in quadrature. Onestator winding and the rotor winding are ex-cited, while the remaining stator winding isshorted. When the resolver is rotated by themain motor, voltages are produced in the sta-tor windings that can be used to determinethe position. For multi-turn systems, two re-solvers (coarse and fine) may be required.

resonance in an RLC circuit, the reso-nance is the state at which the reactance ofthe inductor,XL, and the reactance of the


capacitor,XC , are equal.

XL = 2πfLXC = 1/(2πfC)

resonance fluorescence the modified flu-orescence produced when a quantum me-chanical system is strongly driven by one ormore near-resonant electromagnetic fields.

resonance Raman system Raman sys-tems that have a near-resonant intermediatestate. Resonance Raman systems are some-times referred to as lambda systems and canexhibit coherent population trapping.

resonant in any circuit or system underexcitation, the frequency at which a pair ofreactive components cancels (pole or zero)resulting in a natural mode of vibration.

resonant antenna linear antennas that ex-hibit current and voltage standing wave pat-terns formed by reflections from the open endof the wire.

resonant cavity cavity with reflecting sur-faces or mirrors that can support low-loss os-cillations. By closing a metallic waveguideby two metallic surfaces perpendicular to itsaxis, a cylindrical cavity is formed. Resonantmodes in this cavity are designated by addinga third subscript so as to indicate the num-ber of half-waves along the axis of the cav-ity. When the cavity is a rectangular paral-lelepiped, the axis of the cylinder from whichthe cavity is assumed to be made shouldbe designated, since there are three possiblecylinders out of which the parallelepiped maybe made. More generally, each closed cav-ity may sustain a discrete infinity of resonantfield distributions.

resonant frequency (1) a frequency atwhich the input impedance of an device isnonreactive, since the capacitive and induc-tive stored energy cancel each other.

(2) an oscillation frequency of the modesof a resonator.

resonant link inverter an inverter thatuses a resonant circuit to convert a constantDC voltage to a pulsating DC voltage. Theswitching elements in the inverter are thenturned off during the times that the input volt-age is zero, a technique referred to as soft-switching. Resonant switching techniquesreduce the switching losses and allow highswitching frequency operation to reduce thesize of magnetic components in the inverterunit.

resonant tunneling refers to the processof resonant enhancement of electron tunnel-ing by intermediate energy states. In the sim-plest case, it occurs when incoming electronscoincide in energy with the states created inthe well.

resonator (1) circuit element or combina-tion of elements, which may be either lumpedor distributed, that exhibit a resonance(s) atone or more frequencies. Generally, a reso-nant condition coincides with the frequencywhere the impedance of the circuit element(s)is only resistive.

(2) cavity with reflecting surfaces de-signed to support low-loss oscillation modes.See also bi-directional resonator, concentricresonator, confocal resonator, high-lossresonator, plane-parallel resonator, ringresonator, standing-wave resonator,unidirectional resonator, unstable resonator.

resonator stability perturbation stabilityof an axial light ray in a resonator; bounded-ness of ray trajectories; corresponds to con-finement of the resonator modes; not thesame as mode stability (unstable resonatorshave stable modes).

resource auction multiple access (RAMA)a multiple access protocol that stipulates

part of the frame for contention. UnlikePRMA and D-TDMA, the contention is notperformed by an ALOHA-type of protocolbut by an auction — a tree-sorting type ofalgorithm.


resource conflict the situation when acomponent such as a register or functionalunit is required by more than one instruc-tion simultaneously. Particularly applicableto pipelines.

response formula for general 2-D modelthe solution to the general 2-D model

xi+1,j+1 = A0xij + A1xi+1,j + A2xi,j+1

+B0uij + B1ui+1,j + B2ui,j+1 (1a)


i, j ∈ Z+ (the set of nonnegative integers)with boundary conditionsxi0, i ∈ Z+ andx0j , j ∈ Z+ is given by

xij =i∑

p=1

j∑q=1

(Ti−p−1,j−q−1B0

+ Ti−p,j−q−1B1 + Ti−p−1,j−qB2)upq

+i∑

p=1

(Ti−p,j−1 [A1 B1]

+ Ti−p−1,j−1 [A0 B0]) [ xp0up0

](2)

+j∑q=1

(Ti−1,j−q [A2 B2]

+ Ti−1,j−q−1 [A0 B0]) [ x0qu0q

]+ Ti−1,j−1 [A0 B0]

[x00u00

]wherexij ∈ Rn is the local state vector,uij ∈Rm is the input vector,Ak, Bk (k = 0,1,2)are given real matrices, and the transition ma-trix Tpq is defined by

Tpq =

I for p = q = 0A0Tp−1,q−1 + A1Tp,q−1+A2Tp−1,q for p, q ≥ 0(p + q 6= 0)

0 for p < 0 or/and q < 0

Substitution of (2) into (1b) yields the re-sponse formula.

restart the act of restarting a hardware orsoftware process.

restoration the act of restoring electricservice to a consumer’s facility. A dis-tinction is made between restoration andrepair as repair implies that whatever wasthe cause of the outage has been corrected,whereas restoration implies that the powerwas restored but it may have been throughsome means other than a repair, for ex-ample through switching.See alsoimagerestoration.

restore instruction an assembly languageinstruction that restores the machine state ofa suspended process to the active state.

retentivity Seeremanence.

reticle Seemask.

retiming the technique of moving the de-lays around the system. Retiming does notalter the latency of the system.

retire unit in modern CPU implementa-tions, the module used to assure that instruc-tions are completed in program order, eventhough they may have been executed out oforder.

retrace blanking blanking of a displayduring vertical retrace to prevent the retraceline from showing on the display.

retrace switch electronic method used toblank a display during retrace blanking.

retrace time amount of time that ablanked vertical retrace takes for a displaydevice. Note that this time is less than thetime the display is blanked.

retrograde channel for a MOSFET semi-conductor device, the channel region whosedoping is less at the surface of the channelthan it is at some depth below the surface.

retry a repetition of an operation, dis-turbed by a transient fault, to obtain goodresult.


return address the address of an instruc-tion following a Call instruction, where theprogram returns after the execution of theCall subroutine.

return difference matrix at the inputthe transfer function matrix relating the dif-ference between the injected and the re-turned signal in a feedback loop that has beenopened at the input. This transfer functionmatrix characterizes the effect of closing theloop and is the mathematical entity that iscentral to numerous control engineering the-ories. For a nonunity feedback configurationwith both pre- and postcompensation, it isgiven by

1 + FGK(s)

See alsoreturn difference matrix at theoutput.

return difference matrix at the outputthe transfer function matrix relating the dif-ference between the injected and the re-turned signal in a feedback loop that has beenopened at the output. For a nonunity feed-back configuration with both pre- and post-compensation, it is given by

1 +GKF(s)

See alsoreturn difference matrix at the input.

return instruction an instruction, whenexecuted, gets the address from the top of thestack and returns the program execution tothat address.

return loss usually expressed in decibels,it is the magnitude of the reflection coeffi-cient; the return loss is a measure of the powerreflected due to impedance mismatch of anantenna or other device.

return stroke in lightning, the upwardpropagating high-current, bright, potentialdiscontinuity following the leader that dis-charges to the ground some or all of thecharge previously deposited along the chan-nel by the leader.

return-to-bias recording Seemagneticrecording code.

return-to-zero recording Seemagneticrecording code.

reusability the possibility to use or easilyadapt the hardware or software developed fora system to build other systems.

reuse programming modules are reusedwhen they are copied from one applicationprogram and used in another. Reusability is aproperty of module design that permits reuse.

reuse ratio the ratio of the physical dis-tance between the centers of radio commu-nication cells and the nominal cell radius.This term is usually used with reference tothe reuse of particular radio channels (e.g.,cochannel reuse ratio).

reverberation inhomogeneities, such asdust, sea organisms, schools of fish, and seamounds on the bottom of the sea, form massdensity discontinuities in the ocean medium.When an acoustic wave strikes these inho-mogeneities, some of the acoustic energy isreflected and reradiated. The sum total ofall such reradiations is called reverberation.Reverberation is present only in active sonar,and in the case where the object echoes andis completely masked by reverberation, thesonar system is said to be “reverberation-limited.”

reverse breakdown the diode operatingregion in which significant current flows fromcathode to anode, due to an applied voltageexceeding the breakdown voltage.

reverse breakdown region the region ofthe I-V curve(s) of a device in which the de-vice is operating in avalanche or zener break-down.

reverse breakdown voltage Seereversebreakdown region.


reverse conducting thyristor (RCT) avariety of asymmetric silicon controlled rec-tifier in which the diode is integrated into thethyristor structure.

reverse engineering the reverse analysisof an old application to conform to a newmethodology.

reverse generation-recombination currentpart of the reverse current in a diode

caused by the generation of hole-electronpairs in the depletion region. This currentis voltage-dependent because the depletionregion width is voltage-dependent.

reverse isolation device or circuit loss ina path that is the reverse of the path normallydesired, expressed as the power ratio in deci-bels of the RF power delivered to a load atthe input port versus the RF power incidentat the output port. The negative sign resultsfrom the term “loss.”

reverse leakage current a nondestructivecurrent flowing through a capacitor subjectedto a voltage of polarity opposite to that nor-mally specified.

reverse link Seeuplink.

reverse saturation current part of thereverse current in a diode caused by diffusionof minority carriers from the neutral regionsto the depletion region. This current is almostindependent of the reverse voltage.

reverse translation buffer Seeinversetranslation buffer.

reverse voltage the voltage across the de-vice when the anode is negative with respectto the cathode.

reversible a motor capable of running ineither direction, although it may be neces-sary to rewire the connections to the motorto change the direction of rotation.See alsoreversing.

reversible loss a decrease in magneticinduction (B) of a permanent magnet whensubjected to thermal or magnetic demagneti-zation that is fully recovered (without remag-netization) when the detrimental conditionsare removed.

reversible temperature coefficient forpermanent magnets, a quantity that indicatesthe reversible change in magnetic inductionwith temperature.

reversing a motor that can be run in eitherdirection through the use of suitable switchesor contactors.

reversing motor starter a motor con-troller capable of accelerating a motor fromrest to normal speed in either direction ofrotation. Some reversing motor starters cango directly from forward to reverse (or viceversa), while others must be stopped beforea reversal of direction can take place. Bothelectromechanical and electronic reversingstarters are available.

revolute joint a joint characterized by arotation angle that is the relative displace-ment between two successive links.

revolving field the magnetic field createdby flow of a set of balanced three-phase cur-rents through three symmetrically displacedwindings. The created field revolves in theair-gap of the machine at an angular veloc-ity corresponding to the synchronous speedof the machine. The revolving field theory isthe basis of functioning of synchronous andinduction machines.

Reynolds number a nondimensional pa-rameter used to determine the transition toturbulence in a fluid flowing in pipes or pastsurfaces.

RF Seeradio frequency.

RF amplifier amplifier capable of provid-ing gain at radio frequencies.


RF choke a large-valued inductor that ex-hibits a large reactance at the operating fre-quency that effectively blocks the RF signal.

RF input power the difference betweenincident power and reflected power at theinput into a device or circuit, expressed inwatts. The specific point in the design atwhich RF input power will be measured isimportant, since it will affect gain and effi-ciency calculations.

RF output power the difference in thepower available under perfectly matchedconditions and the reflected power taking theoutput return loss into account, expressed inwatts. This is the RF power delivered to theload.

RF quadrature part of a radio frequencyreceiver that contains circuits that can betuned/varied to pass a desired signal carrierwave while rejecting other signals or carriers.

RF quadrature demodulation radio fre-quency demodulation of a signal that wasproduced by two signals separated in phaseby 90 degrees.

RF quadrature modulation radio fre-quency modulation of a signal that is pro-duced by two signals separated in phase by90 degrees.

RF tuner the part of a radio frequencyreceiver that contains circuits that can betuned/varied to pass a desired signal carrierwave while rejecting other signals or carriers.

RFI Seeradio frequency interference.

RFIC See radio frequency integratedcircuit.

RGB the most widely used image rep-resentation, where color is represented bythe combination of the three primary col-ors of the additive light spectrum (See alsotristimulus value). The RGB space is rep-

resented in Cartesian coordinates as a unitcube, where the origin represents black andthe point(1,1,1) represents white.See alsocolor space.

rheobase the minimum current necessaryto cause nerve excitation — applicable to along duration current (e.g., several millisec-onds).

rib waveguide a type of dielectric waveg-uide formed by several planar layers of di-electric media; the upper layer, instead ofbeing planar, presents a ridge (rib) where thefield is mostly confined. Used in integratedoptics.

Riccati equation a class of equations thatarises frequently in statistics and linear sys-tems theory, describing the evolution of thestatistics of measured systems. As with theLyapunov equation, several types exist. Ineach case,P,Q,R are symmetric positivedefinite;A,B,C are arbitrary.

Continuous-time, time-varying RiccatiEquation:

P = A(t)P (t)+ P(t)AT (t)

+ B(t)Q(t)BT (t)

− P(t)C(t)R−1(t)C(t)P (t).

Continuous-time algebraic Riccati Equation:

0 = AP + PAT + BQBT − PCT R−1CP.

Discrete-time algebraic Riccati Equation:

P = APAT + BQBT

− APCT(CPCT + R

)−1CPAT .

See alsoLyapunov equation, Kalman filter.

Rice distribution the probability distri-bution of the magnitude of a complex quan-tity whose real and imaginary parts are in-dependent Gaussian random variables withnonzero mean. Frequently used to approx-imate multipath fading statistics in line-of-sight mobile radio systems.


Rice factor for a Rice-distributed sig-nal, the parameter giving the ratio of thepower of the static (direct) signal componentto the power of the remaining signal com-ponents. The remaining signal components(which follow a Rayleigh distribution) are of-ten referred to as the diffuse signal. The Ricefactor is a parameter of the Rician probabilitydensity function.See alsofading Rician.

Rician distribution SeeRayleigh noise.

ridge detection Seeedge detection.

right hand circular polarization thestate of an electromagnetic wave in which theelectric field vector rotates clockwise whenviewed in the direction of propagation of thewave.

rigid body motion motion of bodies thatare assumed not to change their shape at all,i.e., deformation is absent or is neglected. Incontrast, non-rigid body motion takes defor-mation into consideration.

rigid link Seeflexible link.

ring bus a power transmission scheme inwhich a region is supplied by a continuous,closed loop of power transmission lines.

ring coupler a type of planar 180 degreehybrid that can easily be constructed in planar(micro strip or stripline) form.

ring network a network topology whereall nodes are connected in a loop. The topol-ogy is resilient to breaks in the loop as traf-fic can be rerouted in either direction. Alsoknown as loop network.

ring numbering for access control, pro-tection scheme in which every memory ob-ject is assigned a set of ring numbers andevery executing process is assigned a num-ber. The legality of an access attempt is de-termined by numeric comparisons betweenthe execution ring number and the ring num-

bers of the object to be accessed. A typicaldesign assigns more privilege to lower num-bers, where the system programs reside. Ifthere are three access modes (execute, read,write), three numbers(e, r, w) are assignedto each object. Letp denote the ring num-ber of the executing process. Then execute ispermitted ifp < e, read ifp < r, and writeif p < w. This scheme is simple to explain,but does not support general access controlpolicies, since they cannot be mapped to alinear sequence of integer values.

ring resonator resonator in which formuch of the mode volume the electromag-netic waves are described in terms of trav-elling rather than standing waves. Since thering has no open ends, radiative losses arevery small. In such a resonator, the modevolume of the electromagnetic waves aredescribed in terms of travelling rather thanstanding waves.

ringing (1) the phenomena in discrete-time (sampled data) systems in which a sys-tem containing a single pole on the negativereal axis in the z-plane can oscillate with aperiod of twice the sampling interval. Math-ematically, this discrete effect can be relatedback to the discontinuity that exists along thenegative real axis in the z-plane.

(2) in image processing, the occurrence ofripples near edges in an image processed bya lowpass filter with a steep transition band.

ripple the AC (time-varying) portion ofthe output signal from a rectifier circuit.

ripple current the total amount of alter-nating and direct current that may be appliedto an electrolytic capacitor under stated con-ditions.

ripple-carry adder a basicn-bit adderthat is characterized by the need for carries topropagate from lower- to higher-order stages.

RISC Seereduced instruction set computerprocessor.


rise time the time required for a digitalsignal to make the transition from a “low”value to a “high” value.

rise time degradation a measure of theslowing down of the pulse as it passes throughan I&P element. It includes both the increasein risetime of the pulse, as well as loss inamplitude.

rising edge in a clock or data signal, thatportion of the signal that denotes the changefrom the “low” state to the “high” state.

RLL Seerun-length limited code.

RMS Seeroot-mean-squared error.

RMS delay spread Seeroot-mean-squareddelay spread.

RMS Doppler spread Seeroot-mean-squared Doppler spread.

RMS gain ripple Seeroot-mean-squaredgain ripple.

RMS phase ripple Seeroot-mean-squaredphase ripple.

RMS power See root-mean-squaredpower.

RMW memory cycle Seeread-modify-write cycle.

robot a term originated from the Chechword robota,meaning work.

A definition used by the Robot Instituteof America gives a more precise descrip-tion of industrial robots: “A robot is a re-programmable multifunctional manipulatordesigned to move materials, parts, tools, orspecialized devices, through variable pro-grammed motions for the performance of avariety of tasks.”

The British Robot Association definesa robot as “A reprogrammable device de-signed to both manipulate and transport parts,

tools or specialized manufacturing imple-ments through variable programmed motionsfor the performance of specific manufactur-ing tasks.”

robot programming language a com-puter programming language that has spe-cial features which apply to the problemsof programming manipulators. Robot pro-gramming is substantially different from tra-ditional programming. One can identify sev-eral considerations that are typical to anyrobot programming method: The objectsto be manipulated by a robot are three-dimensional objects; therefore, a special typeof data is needed to operate an object. Robotsoperate in a spatially complex environment.The description and representation of three-dimensional objects in a computer are im-precise. Also, sensory information has tobe monitored, manipulated, and properly uti-lized. Robot programming languages can bedivided into three categories:

1. Specialized manipulator languagesbuilt by developing a completely new lan-guage. An example is the VAL language de-veloped by Unimation, Inc.

2. Robot library for an existing computerlanguage. It is a popular computer languageaugmented by a library of robot-specific sub-routines. An example is PASRO (Pascal forRobots) language.

3.Robot library for a new general-purposelanguage. These robot programming lan-guages have been developed by first creat-ing a new general purpose language, andthen supplying a library of predefined robot-specific subroutines. An example is AMLlanguage developed by IBM.

robot vision a process of extracting, char-acterizing, and interpreting information fromimages of a three-dimensional world. Thisprocess is also called as machine or computervision.

robot-oriented programming using astructured programming language that in-corporates high-level statements and has the


characteristics of an interpreted language, inorder to obtain an interactive environment al-lowing the programmer to check the execu-tion of each source program statement beforepreceding to the next one. Robot-orientedprogramming incorporates the teaching-by-doing method, but allows an interaction ofthe environment with physical reality.Seealsoteaching-by-showing programming.

robust control control of a dynamical sys-tem so that the desired performance is main-tained despite the presence of uncertaintiesand modeling inaccuracies.

robust controller design a class of designprocedures leading to control systems that arerobust in the sense of required performance.Robust design is a feedback process involv-ing robustness analysis. A specific techniqueused in robust controller design depends onthe type of model describing a system andits uncertainty, control objective, and a set ofadmissible controllers. The first requirementis to ensure robust stability; this could be fol-lowed by guaranteed cost, disturbance rejec-tion, robust poles localization, target sets ortubes reachability, or other demands.

Ackermann’s three basic rules of robustcontroller design are as follows:

1. Require robustness of control systemonly for physically motivated parameter val-ues and not with respect to arbitrarily as-sumed uncertainties of the model.

2. When you close a loop with actuatorconstraints, leave a slow system slow andleave a fast system fast.

3. Be pessimistic in analysis; then, youcan afford to be optimistic in design.

See alsoworst-case design, min-maxcontrol, practical stabilization, guaranteedcost control, H infinity design.

robust estimation an estimation schemein which we optimize performance for theleast favorable statistical environment amonga specified statistical class.

robust fuzzy controller a fuzzy controllerwith robustness enhancement or robust con-troller with fuzzy logic concepts.

robust fuzzy filter a fuzzy filter with ro-bustness enhancement or robust filter withfuzzy logic concepts.

robust stability a property of the family ofmodels for the system with uncertainty whichensures that it remains stable for all possibleoperating conditions for uncertain variablesranging over their sets. For time-invariantlinear systems with uncertainty, robust sta-bility means that the family of characteristicpolynomials generated by uncertain param-eters defined over the operating sets has allroots endowed with negative real parts.SeeKharitonov theorem. For nonlinear sys-tems, robust stability could be checked bysome techniques based on direct Lyapunovmethod or Popov criterion. In some cases,the requirement of robust asymptotic stabilitymay be weakened by more realistic practicalstability or ultimate boundedness demand.See alsopractical stabilization.

robust statistics the study of methodsby which robust measures may be extractedfrom statistical or numerical data, therebyexcluding measurements which are unlikelyto be reliable and weighting other measure-ments appropriately, thereby increasing theaccuracy of finally assessed values. Of spe-cific interest is the systematic eliminationof outliers from the input data.See alsorobustness, median filter.

robustness (1) a control system qualityof keeping its properties in the admissiblerange in spite of disturbances and other en-vironmental perturbations as well as uncer-tainties in the system model. The most fre-quent requirements deal with robust stabilityand robust performance expressed, for ex-ample, in terms of guaranteed cost. Robustsystems have the property of being insensi-tive to changes in the model parameters aswell as external disturbances. Usually the


system robustness is reached via robust con-troller design, although systems could be ro-bust in some sense (for example, robustly sta-ble) without use of any special design tech-niques. It is, for example, well known thatfor the majority of real-world plants, stan-dard PID controller suitably tuned ensuressufficient robustness. Nevertheless, in manysituations, robustness can be guaranteed onlyby sophisticated design techniques such asH infinity design, min-max control, practi-cal stabilization, guaranteed cost control, andothers.

(2) the property of a process that results inits being able to suppress the effects of noisyor unreliable data, thereby arriving at reliablemeasures or interpretations, and degradinggracefully as more and more unreliable datais included. With image data, robust proce-dures are those which are able to detect ob-jects without becoming confused by partialocclusions, noise, clutter, object breakages,and other distortions.

ROC Seereceiver operating characteristicscurve, region of convergence.

ROI Seeregion of interest.

rollback Seebackward error recovery.

Rollett stability factor the inverse of theLinville stability factor (C), K is a measureof potential stability in a 2-port circuit operat-ing under small signal conditions, but stand-alone is insufficient to guarantee stability. A2-port circuit that is matched to a positivereal source and load impedance is uncondi-tionally stable ifK > 1, B1 > 0 (port 1stability measure), andB2 > 0 (port 2 stabil-ity measure). The design must provide suffi-cient isolation from the RF input and outputports to the bias ports to allow a reasonableinterpretation of the “2-port” device criteria.

K = 1

C= 1− | s11 |2 − | s22 |2 + | 1 |2

2· | s12 · s21 |where1 = s11 · s22 − s12 · s21.

rolling ball a method of determining thelightning protection to nearby structures af-forded by a tall, well-grounded structure likea steel tower. A shpere with a radius of 45 me-ters is imagined rolled against the tower. Anystructure which can fit within the space de-fined by the sphere’s point of contact with theearth, the base of the tower, and the sphere’spoint of contact with the tower is consideredto be protected by the tower against lightningstrikes.Seecone of protection.

rolloff transition from pass-band to stop-band.

ROM Seeread-only memory.

Romex cable a heavily insulated, non-armored cable used in residential wiring.

root locus the trajectory of the roots of analgebraic equation with constant coefficientwhen a parameter varies.

root sensitivity the dependence of thepoles and zeros of a lumped parameter cir-cuit function on the circuit elements.

Let a lumped parameter circuit functionbe represented as

F(s) = am∏mi=1(s − zi)

dn∏ni=1(s − pi)

wherezi are zeros andpi are poles. IfF(s) isalso a function of the circuit elementx, the lo-cation of these poles and zeros will depend onthis element. This dependence is describedby the semirelative root sensitivities

Sx(zi) = x∂zi

∂xSx(pi) = x

∂pi

∂x

The calculation of these sensitivities is sim-plified when zeros and poles are simple (notmultiple). For example, for a simple polepi ,the denominator of the circuit function satis-fies the relationship

D (pi) = D1 (pi)+ xD2 (pi) = 0


and one can find that

Sx(pi) = x∂pi

∂x= −xD2(pi)

D′(pi)whereD′(pi) is the derivative of the poly-nomialD(p) calculated at the pointpi . Formultiple poles, the calculation of the root sen-sitivity is more involved. The relationshipbetween the function sensitivity and polesand zeros sensitivities can also be estab-lished.

root-mean-squared delay spread a mea-sure of the width of a delay power spectrum.Computed for the delay power spectrum ina similar fashion as the standard deviation iscomputed for a probability density function.Usually known as RMS delay spread.

root-mean-squared Doppler spread ameasure of the width of a Doppler powerspectrum. Computed for the Doppler powerspectrum like the standard deviation is com-puted for a probability density function. Alsoknown as RMS Doppler spread.

root-mean-squared (RMS) error thesquare root of the mean squared error.

root-mean-squared gain ripple the dif-ference of the root-mean-squared (RMS) val-ues of the power gain peaks and gain dipsrelative to the RMS power across a specifiedband.

root-mean-squared phase ripple the dif-ference of the root-mean-squared (RMS) val-ues of the phase peaks and phase dips rela-tive to a best fit linear phase response acrossa specified band.

root-mean-squared power the averagepower, expressed in watts, delivered to theload over a complete period of time (T ), asif the current and voltage were constant overthat time period.

PRMS = 1

T

∫ T

0p · dt .

Also known as RMS power or average power.

rotate a logical operation on a data ele-ment that shifts each bit one position to theleft or right. The bit at the end of the loca-tion is transferred to the opposite end of theelement.

rotating excitation system an excitationsystem derived from rotating AC or DC ma-chines. The output of the system is still DCand connected to the rotor.

rotating wave approximation assump-tion in a semiclassical model for the interac-tion of light with atoms that all populations,field amplitudes, and polarization amplitudeschange negligibly within one optical cycle.

rotating-rectifier exciter an AC gener-ator, with rotating armature and stationaryfield, whose output is rectified by a solid-state device located on the same shaft to sup-ply excitation to a larger electrical machine,also connected to the same shaft.

rotational latency the time it takes forthe desired sector to rotate under the headposition before it can be read or written.

rotational loss one of several losses in arotating electric machine that are primarilydue to the rotation of the armature and in-clude the friction and windage losses. Alsocalled mechanical loss. They can be deter-mined by running the machine as a motor atits rated speed at no load, assuming the ar-mature resistance is negligible.

rotational position sensing a mechanismused in magnetic disks, whereby the disk in-terrupts the I/O controller when the desiredsector is under the read/write head. Used torecognize the different sectors in a track andsynchronize the different bits in a sector.

rotational transition transition betweenrotational states of a molecule.


Rotman lens a constrained cylindrical-lens antenna over a wide instantaneous band-width as derived by W. Rotman and P. Franchi(1980). The antenna configuration consistsof a stripline lens and line feed. The lens lay-ers can use a microstrip printed-circuit con-struction. Microstrip lines interconnect radi-ating elements.

rotor the rotating part of an electrical ma-chine including the shaft, such as the rotatingarmature of a DC machine or the field of asynchronous machine.

rotor power developed the amount ofpower developed by the rotor. In DC ma-chines, the developed power, frequently de-noted byPd , is calculated as the product ofthe induced EMFEa and the armature currentIa . In induction machines, the rotor powerdeveloped is obtained by subtracting the ro-tor copper losses from the air gap power.

rotor power input represents the totalpower delivered from the armature coil of aninduction motor across the air gap to the ro-tor via the air gap magnetic flux. It is repre-sented, on a per-phase basis, by

RPI = I2s

(R2eq/s

)= I2

s R2eq + I2s

(R2eq(1 − s)/s

)where Is is the armature coil current, R2eqis the equivalent per-phase rotor resistancereflected to the armature, ands is the rotorslip speed. The first term on the right-handside represents the ohmic heating power dis-sipated in the rotor windings, and the secondterm represents the conversion of electricalto mechanical power by the rotor.See alsorotor power loss.

rotor power loss represents the portion ofthe power transferred across the air gap to therotor of an induction motor that is lost eitherthrough ohmic heating of the rotor windingsor due to friction and windage losses in therotor. The mechanical power available at themotor shaft is the difference between rotor

power input and rotor power losses.See alsorotor power input.

rotor reference frame a two-dimensionalspace that rotates at the electrical angularvelocity of a specified machine rotor. Inelectric machines/power system analysis, anorthogonal coordinate axis is established inthis space upon which fictitious windings areplaced. A linear transformation is derivedin which the physical variables of the sys-tem (voltage, current, flux) are referred tovariables of the fictitious windings.See alsoPark’s transformation,arbitrary referenceframe, synchronousreference frame, stationary reference frame.

rotor speed quantification of the rota-tional operation of the moving part of a ro-tating electrical machine. The rotor speed ismeasured either in SI units in radians per sec-ond (rad/s) or in practical units in revolutionsper minute (rev/min).

rough surface surface whose corrugationis random and appreciable compared with thelight wavelength so as to produce light scat-tering. It is commonly characterized by theroot mean square and the correlation lengthof the random height profile.

round-robin arbitration a technique forchoosing which of several devices connectedto a bus will get control of the bus. After adevice has had control of the bus, it is notgiven control again until all other devices onthe bus have been given the opportunity to getcontrol in a predetermined order. The oppor-tunity to get control of the bus circulates in apredetermined order among all the devices.

rounding an operation that modifies afloating-point representation considered in-finitely precise in order to fit the requiredfinal format. Common rounding modes in-clude round to nearest, round toward zero,and round toward positive or negative infin-ity.


router a node, connected to multiple net-works, that forwards packets from one net-work to another. It is much more complexthan bridges that work between networkshaving compatible protocols. Also called agateway.

routing given a collection of cells placedon a chip, the routing routine connects theterminals of these cells for a specific designrequirement.

row decoder logic used in a direct-accessmemory (ROM or RAM) to select one ofa number of rows from a given row ad-dress. See alsotwo-dimensional memoryorganization.

row-access strobe Seetwo-dimensionalmemory organization.

RS flip-flop a single-bit storage element,usually formed by connecting two NOR orNAND gates in series. RS stands for reset–set. For state variableQ and next state vari-ableQ′, the simplified truth table is given as

R S Q Q′0 0 0 00 0 1 10 1 0 00 1 1 01 0 0 11 0 1 11 1 0 X1 1 1 X

the symbol “X” is used to denote an unknownstate for the flip-flop.See alsoJK flip-flop.

RS-170A technical standard developed bythe Electronics Industry Association that de-scribes in detail the relationship between ver-tical, horizontal, and subcarrier componentswithin a video signal. The standard permitssynchronization of two or more video signals.

RS-422 technical standard developed bythe Electronics Industry Association that de-fines the exact physical, electrical, and func-

tional characteristics for a 40-pin connec-tor that links a computer to communicationequipment.

RSGNN See recursive self-generatingneural network.

RSSI Seereceived signal strength indicator.

RTB reverse translation buffer. Seeinverse translation buffer.

RTU Seeremote terminal unit.

rubbers personal protective wear forline workers, including insulating gloves,sleeves, and rubber boots.

ruby amplifying medium employed in thefirst man-made optical frequency laser.

ruby laser first man-made optical fre-quency laser.

run winding the main winding of a single-phase induction motor.

run-length coding the assignment of acodeword to each possible run of 0s (whitepel sequence) or run of 1s (black pel se-quence) in a scan of the subject copy.

run-length encoding See run-lengthlimited code.

run-length limited (RLL) code a linecode that restricts the minimum and/or max-imum number of consecutive like-valuedsymbols that can appear in the encoded sym-bol sequence.

running digital sum the difference be-tween cumulative totals of the number oflogic 1s and number of logic 0s in a binarysequence. It is a common measure in the per-formance of a line code.

running integral for the functionx(t), therunning integral,y(t) isy(t) = ∫ t

−∞ x(τ)dτ .


An example of a running integral is the unitstepu(t) = ∫ t

−∞ δ(τ )dτ , whereδ(t) is theunit impulse function.


SS-100 a 100-pin bus formerly used bycomputer hobbyests and experimenters.

S-matrix Seescattering matrix.

S-parameters Seescattering parameters.

S-plane the domain of the Laplace trans-form F(s) of a complex-valued functionf (t). Sinces is a complex number, the do-main ofF(s) is the complex plane. The lines = jω, whereω is a real number, corre-sponds to whereF(s) is the Fourier transformof f (t).

S/H Seesample-and-hold amplifier.

saddle a U-shaped piece of wire which iscrimped to a main conductor so that a hot tapcf. can be readily attached at a later time.

saddle-point equilibrium an equilibriumin zero-sum game constituted by a pair of se-curity strategies of the players. The neces-sary and sufficient condition for existence ofthe saddle point is equality of the security lev-els for both players. IfJi denotes a cost func-tion of the zero-sum game anddi; i = 1,2 astrategy of theith player, then a paird∗

1, d∗2

is in saddle-point equilibrium iff

J(d∗

1, d2) ≤ J

(d∗

1, d∗2

) ≤ J(d1, d

∗2

)for all admissible(d1, d2). Depending of thetype of game, the meaning of the cost func-tional and strategies and the specific condi-tions for the existence of the saddle point mayvary. If the equilibrium in pure strategiesdoes not exist and the game is played manytimes in the same conditions, then the saddlepoint may be defined for average values of thecost function and for mixed strategies defined

as probability distributions on the spaces ofpure strategies.

Safe Medical Devices Act (SMDA) apublic law that imposes reporting require-ments on “device-user facilities” includ-ing hospitals, ambulatory surgical facilities,nursing homes, and outpatient clinics. Theyare required to report information that “rea-sonably suggests” the probability that a med-ical device has caused or contributed to thedeath, serious injury, or serious illness of apatient at that facility.

Safeguard a program administered by theInternational Atomic Energy Agency com-prising procedures and inspections which as-sure that fissile materials from power reactorsare not diverted to nuclear weapons use.

safety the probability that a system willeither perform its functions correctly or willdiscontinue its functions in a well-defined,safe manner.

safety-critical system a system that is in-tended to handle rare unexpected, dangerousevents.

sag a decline ranging from 0.1 to 0.9 puin RMS voltage or current at the supply fre-quency for a time period of 0.5 cycles to1 minute.

Sagittal projection a projection of a3-dimensional object onto a 2-dimensionalplane which intersects the object in a front toback direction dividing the objects into rightand left halves. Typically with reference toan animal or human body.

Sagnac interferometer a common pathinterferometer devised by Georges Sagnac tomeasure the ether wind, then adapted as anoptical gyroscope, and most recently utilizedfor hyperspectral imaging. A Sagnac inter-ferometer is composed of two coils of opti-cal fiber arranged so that light from a single


source travels clockwise in one and counter-clockwise in the other.

Sagnac logic gate an all-optical gatebased on a Sagnac interferometer.

Sagnac, Georges (1869–1928) Born:

French scientist and professor at the Uni-versity of Lille and the University of Paris,active in the investigation of radiation pro-duced by X-rays, and the optics of inter-ference. The inventor of the common pathSagnac interferometer.

sail switch a device used in control sys-tems that detects the flow of air, or other gas,and causes a relay to open or close as a resultof the motion of the sail.

salient feature a characteristic often localfeature on an object which can be detectedand used as part of the process of inferringthe presence of an object from its features.Typical salient features include point featuressuch as corners and small holes, or fiducialmarks (e.g., on printed circuit boards), butmay in addition include large-scale straight-forwardly detected features such as large cir-cular holes which can also aid the inferenceprocess.

salient-pole drive See synchronousdrive.

salient-pole rotor machine AC mo-tor/generator design in which the rotor is con-structed of outward-projecting pole piecesmounted on a shaft-mounted central spiderassembly. Spider assemblies are typicallyspoked. Pole pieces are built up from lam-inated sheets, which are bolted together be-tween a pole shoe on the outer end and dove-tail fixture on the inner end. The dovetailsare keyed into slots on the spider to mount thepole pieces to the rotor. Rotor windings aregenerally constructed from preformed, insu-lated coils that are fit over the pole piecesduring assembly. Salient rotors are typically

low-speed designs with short axial length andlarge diameter.

salt and pepper noise Seeimpulse noise.

SAM Seestandard additive model.

samarium cobalt a brittle, high-energymagnetic material that is best known forits performance at high temperatures, whichcomes in two compositions, SmCo5 and ahigher energy Sm2Co17.

sample a single measurement that is takento be representative of the measured propertyover a wider area, frequency range, or timeperiod. When recording digital sound, a sam-ple is a voltage measurement that reflects theintensity of the acoustic signal at a particu-lar moment, and has a time period associatedwith it, that is, the sample represents the sig-nal until the next measurement is made. In adigital image, a sample is a single measure-ment of light intensity at a particular pointin the scene, and that measurement is used torepresent the actual but unmeasured intensityat nearby points.

sample complexity the number of train-ing examples required for a learning systemto attain a specified learning goal.

sample space the set of all possible sam-ples of a signal, given the particular parame-ters of the sampling scheme.

sample-and-hold Seesample-and-holdamplifier.

sample-and-hold amplifier (SHA) aunity gain amplifier with a mode controlswitch where the input of the amplifier is con-nected to a time-varying signal. A triggerpulse at the mode control switch causes it toread the input at the instance of the triggerand maintain that value until the next triggerpulse. Usually the signal is electrical, butother forms, such as optical and mechanicalforms, are possible. Also denoted S/H.


sample-and-hold circuit a device with amode control switch that causes reading theinput at the instance of the trigger and main-taining that value until the next trigger pulse.

sampling the act of turning a time-continuous signal into a signal that is time-discrete or time-discontinuous. See the fol-lowing diagrams, whereTS is the samplingperiod.

The sampling process.

In order to maintain frequency compo-nents of interest in the time-discontinuoussignal, the Nyquist sampling criterion needsto be satisfied. This criterion states that therate of sampling of the time-continuous sig-nal has to be at least twice as great as thefrequency of the signal component with thehighest frequency which is of interest in thetime-continuous signal. Sampling is verycommon in digital recording and digital com-munication systems.

sampling frequency in analog-to-digitalconversion, the rate or frequency at which ananalog signal is sampled and converted intoa digital signal.

sampling function a mathematical func-tion used when sampling a signal. In partic-ular, a sampling functionS(t) can be multi-plied by the continuous function to be sam-pled,F(t), to obtain the sampled version ofF . S is most often a collection of equallyspaced impulses.

sampling period the period for which thesampled variable is being held constant.

sampling process Seesampling.

sampling rate Seesampling frequency.

sampling theorem See Shannon’ssampling theorem.

SAR See synthetic aperture radarorspecific absorption rate.

satellite cell cell with the radius largerthan 500 km where the cell is controlled bya satellite.See alsocell.

satellite imagery the acquisition of pic-tures of the earth from space. Satellite im-agery can be used to enhance maps, col-lect resource inventories (e.g., forestry, wa-ter, land use), assess environmental impact,appraise damage following a disaster, andcollect information on the activities of hu-mans. Satellite imagery tends to be multi-spectral, including a wide range of opticalfrequencies and, more recently, infrared andradar.See alsoremote sensing.

saturable absorber the nonlinear opticalphenomenon in which the absorption coeffi-cient of a material decreases as the intensityof the light used to measure the absorptionincreases.

saturable absorption the effect of therebeing less absorption in a material for largervalues of the incident illumination.

saturated gain value of the gain in a sat-urable amplifier for a particular value of in-tensity.

saturated logic logic gates whose outputis fully on or fully off, determined principallyby the external circuit.

saturating control a controller producinga bounded control signal. Finite limits onthe magnitude of the control signals that areprovided by the actuators are due to the factthat the actuators are physical devices andas such are subject to physical constraints.


Thus, the actuator saturates, that is, it has“limited authority.”

saturation (1) the failure of the output toincrease as fast as the input.

For example, often the current regulatorused in variable-speed drives is unable totrack the commanded current because of in-sufficient voltage difference between the mo-tor back EMF and the supply.

In an amplifier, saturation results in a re-duction of gain in an amplifier or loss in anabsorber due the intensity of the signal beingamplified or absorbed.

In ferromagnetic circuits, the magneticflux initially increases linearly with the ap-plied magnetomotive force (MMF), but even-tually most of the domains in the ferromag-netic material become aligned, and the rate ofincrease in flux decreases as the MMF con-tinues to increase. See figure below.See alsosaturation flux density.

Saturation effect in a ferromagnetic circuit.

(2) with respect to color, the amount orpurity of the color seen. A pure color is saidto be fully saturated, and the saturation de-creases as white is added to the mix. Thecolor “pink,” for example is a less saturatedversion of “red.”

saturation angle the angular portion ofthe time-based output signal (in degrees) overwhich the device is saturated. It is alwaysless than or equal to the conduction angle,

since the device must conduct before it cansaturate.

saturation flux density the maximumvalue of intrinsic induction (Bi) beyondwhich an increase in magnetizing field yieldsno further improvement, indicating that allmagnetic moments in the material have beenaligned.

saturation intensity the intensity of abeam of light above which saturation ef-fects become appreciable.See alsosaturableabsorption.

saturation magnetization the magneticmoment per unit volume of a material whenthe magnetization in the sample is aligned(saturated) by a large magnetic field.

saturation parameter reciprocal of thevalue of intensity for which the gain of anamplifying medium or the loss of an absorb-ing medium is reduced to one half of its un-saturated value.

saturation polarization the value towhich the externally measured electricaldipole moment of a ferroelectric body tendswhen subjected to an external electrical fieldgreater than the coercive field.

save instruction an assembly languageinstruction that saves information about thecurrently executing process.

SAW Seesurface acoustic wave.

SAW device Seesurface acoustic wavedevice.

saw-tooth coupler transmission line cou-pler consisting of two parallel transmissionlines placed in close proximity to one another.The adjacent edges of the two transmissionlines are shaped in a notch or saw-tooth pat-tern to equalize the phase velocity of the evenand odd mode voltage components.


SCA See subsidiary communicationauthorizationandstation control error.

SCADA acronym for supervisory controland data acquisition. A system which mea-sures critical power system parameters (e.g.,voltage, power flow, circuit breaker status,and generator outputs) at remote points in anelectric power system and transmits the datato a central control site where these condi-tions may be monitored.

scalable video coding compression ofvideo such that transmission at different datarates, or reception by decoders with differ-ing performance, is possible merely by dis-carding or ignoring some of the compressedbitstream, i.e., without recoding the data.The compressed data are prioritized suchthat low-fidelity reconstruction is possiblefrom the high-priority data alone; additionof lower-priority data improves the fidelity.

scalar network analyzer a test instru-ment designed to measure and process onlythe magnitude of transmitted and reflectedwaves. Used to measure such microwavecharacteristics as insertion loss, gain, returnloss, SWR, and power.

scalar processor a CPU that dispatchesat most one instruction at a time.

scalar quantization (SQ) (1) quantiza-tion of a scalar entity (a number; as opposedto vector quantization), obtained, e.g., fromsampling a speech signal at a particular time-instant. Each input value to the quantizer isassigned a reproduction value, chosen froma finite set of possible reproductions. A de-vice performing scalar quantization is calleda (scalar) quantizer.

(2) a type of quantization in which a scalarquantity is quantized into another scalarquantity.

scalar wave wave that can be describedby a single scalar function of space and time.

scalar wave equation in optics, a simpli-fication of the Maxwell–Heaviside equationsthat governs a single scalar function repre-senting an electromagnetic wave; sometimesa complex equation if the waves are harmonicin time.

scale (1) a property of an image relatingthe size of a pixel in the image to the size ofthe corresponding sampled area in the scene.A large scale image shows object features inmore detail than a small scale image. (Seealsoresolution. )

(2) to change the size (i.e., enlarge orshrink) of an image or object while main-taining the overall proportions.

(3) one of two parameters of a wavelet, theother being translation. The scale specifiesthe duration of the wavelet.

scaled processor architecture (SPARC)name for a proprietary class of CPUs.

scaling function the solution to the multi-scale equation; it can be obtained by iteratinga low pass filter in the two-channel filter bankan infinite number of times.

scan design a technique whereby storageelements (i.e., flip-flops) in an IC are con-nected in series to form a shift-register struc-ture that can be entered into a test mode toload/unload data values to/from the individ-ual flip-flops.

scan line in a digital image, a contiguousset of intensity samples reflecting one row orcolumn of the image. A class of image pro-cessing algorithms, called scan line or scanconversion methods, looks at the image oneor two scan lines at a time in order to achievethe goal.

scan tape Seehelical scan tape.

scan-based testing a mechanism for ac-cessing all the data in a hardware module bytreating it as one long shift register and thenshifting the data out of the module one bit at


a time. A device with this capability can alsobe set to any desired state by shifting in thedesired state. The method can also be appliedto software objects.

scan-test path a technique that enhancescircuit observability and controllability byusing a register with shift (in test mode) andparallel load (during normal operation) capa-bilities.

scanner (1) a device used for scanningwritten documents or printed pictures bytracing light along a series of many closelyspaced parallel lines.

(2) any device that deflects a light beamthrough a range of angles, using mecha-nisms such as diffraction from electro-opticor acousto-optic gratings or mechanical de-flectors.

(3) a type of projection printing toolwhereby the mask and the wafer are scannedpast the small field of the optical system thatis projecting the image of the mask onto thewafer.

scanning process for converting attributesof a display at raster coordinate locations,such as color and intensity, into a fixed set ofnumerical attributes for manipulation, trans-mission, or storage of the display.

scanning tunneling microscope extreme-ly sensitive method for measuring atomic po-sition at a surface by monitoring the electroncurrent due to tunneling between a moveablemetal tip and the surface semiconductor.

scara manipulator a robot with three par-allel revolute joints allowing it to move andorient in a plane(q1, q2, q3) with a fourthprismatic joint (q4) for moving the end-effector normal to the plane. Usually scaramanipulators can move very fast and they areused to assemble the parts.

scattering (1) the process by which aradio wave experiences multiple reflections

A scara manipulator.

from a surface or a volume of the propaga-tion medium.

(2) the area of electromagnetics that dealswith finding the total fields in a region thatcontains an incident field and one or more ob-jects (scatters) in the region. The total field isthe sum of the incident field when no objectsare present, plus the additional field producedwhen the objects are present.

scattering cross section total energy scat-tered in all directions, normalized to thewavenumber squared. It has dimensions ofarea.

scattering function the scattering func-tion is a function of two variables, delay andDoppler shift, characterizing the spread ofaverage received signal energy over time andfrequency. In a nondispersive channel (nomultipath and no Doppler shift), the scatter-ing function is simply an impulse at zero de-lay and zero Doppler shift.

scattering holography recovery of thephase of the scattered light by means of itsinterference with a reference beam, both em-anating from the same coherent source, suchas the same laser beam.

scattering matrix an× n square matrixS of complex numbers used to relate an ar-ray of incident waves,a, and one of reflectedwaves,b, for an n-port network represent-ing a waveguide component or discontinu-ity. Generally, each matrix element (called


a scattering parameter) is the amplitude ofa reflected or transmitted wave to that of anincident wave. The subscripts of a typicalcoefficientSij refer to the output and inputports related by the coefficient. These co-efficients are generally frequency-dependentand are relative at a specified set of input andoutput reference planes. The scattering ma-trix is defined with respect to a specific set ofport terminations (normalizations).

Physical interpretations can be given tothe scattering coefficients; for example,S2

ij

is the fraction of available power at portidue to a source at portj . The incident andreflected wave amplitudes at porti, ai , andbi , respectively, are obtained from the voltageand current,Vi, Ii , at the same port as

ai = 1

2|ReZi |−1/2 (Vi + ZiIi)

bi = 1

2|ReZi |−1/2 (Vi − Z∗

i Ii)

whereZi is the reference impedance and theasterisk denotes the complex conjugate. It isalso known as S-matrix.

scattering parameters parameters thatcharacterize a microwave network with anarbitrary number of ports by relating the volt-age waves incident on the ports to those re-flected from the ports. Also known as S-parameters. The scattering parameters areoften represented in terms of a scattering ma-trix. See alsoscattering matrix.

scattering resonance sharp increase (ordecay) of the scattered energy as a functionof either frequency or observation direction,as for the glory or the rainbow.

scatterometer a device to measure the an-gular distribution of scattered intensity. Itsmain component contains a photodetectormounted over a goniometer.

scene analysis the process of analyzing a3-D scene from 2-D images. Typically, thisprocess will involve object and feature detec-tion, inference of the presence of objects from

their features, projective invariance proper-ties, analysis of the play of light on surfaces,including approaches such as texture analy-sis, and many other types of procedure.

scheduler a part of the operating systemfor a computer that decides the order in whichprograms will run.

scheduling in an operating system,scheduling of CPU time among competingprocesses.

schematic a diagram that shows how anelectronic device is constructed.

schematic capture a design entry methodwherein the designer draws the schematic ofthe desired circuit using a library of standardcells. The program outputs a netlist of theschematic.

schematic diagram a circuit diagram, di-vorced of biasing subcircuits, that depictsonly the dynamic signal flow paths of an elec-tronic circuit.

Schmoo plot an X-Y plot giving thepass/fail region for a specific test while vary-ing the parameters in the X and Y coordi-nates.

Schottky barrier diode a two-terminaljunction barrier device formed by a junctionof a semiconductor and a metal. These diodesare widely used in integrated circuit applica-tions and in very high frequency mixer andmultipliers. Also called hot-carrier diode.

Schottky contact a metal-to-semiconductorcontact where, in order to align the Fermi lev-els on both sides of the junction, the energyband forms a barrier in the majority carrierpath.

Schottky noise Seeshot noise.

Schrodinger wave equation (SWE) animportant relationship in quantum mechanics


that describes the energy states of electrons.In its time-independent form it is expressedas

(h/2π)2

2m= ∇2φn + (En + qV ) φn = 0

whereφn is the wave function correspondingto the subbandn whose minimum energy isEn, V is the potential of the region,m is theparticle mass, andh andq are Planck’s con-stant and the electronic charge, respectively.

Schur matrix a square matrix with realelements and whose all eigenvalues have ab-solute values less than one.

scientific visualization the use of com-puter graphics techniques to represent com-plex physical phenomena and multidimen-sional data in order to aid in its understandingand interpretation.

scintillation the variation of electromag-netic signal strength with time due to randomchanges in time of refractive index of the at-mosphere. Apparent at optical frequenciesas the twinkling of stars.

scoreboard term originally used for a cen-tralized control unit in the CDC 6600 proces-sor which enabled out-of-order issue of in-structions. The scoreboard unit held variousinformation to detect dependencies. Nowsometimes used for the simpler mechanismof having a single valid bit associated witheach operand register.

scotopic formally, a description of lumi-nances under which human rod cells are ac-tive. Informally, describing dim or night-time luminances.

scotopic vision vision in the eye deter-mined by the number of and condition of therods in the eye. Also called night vision.

SCP Seeservice control point.

SCR Seesilicon controlled rectifier.

scram the emergency shutdown of a nu-clear reactor by the rapid insertion of all con-trol rods. The term is attributed to EnricoFermi.

scrambling randomization of a symbolsequence using reversible processes that donot introduce redundancy into the bit stream.See alsoreset scrambling, self-synchronizingscrambling.

scrubber a means of removing sulfurdioxide from coal-burning power plant ex-haust gas by forcing it through a chemicalsolution.

SCSI Seesmall computer systems interface.

SDH Seesynchronous digital hierarchy.

SDRAM See synchronous dynamicRAM.

SDS Seestructured distribution systems.

seal-in relay an auxiliary relay that re-mains energized through one of its own con-tacts, which bypasses the initiating circuit un-til deenergized by some other device.

sealing current the current necessary tocomplete the movement of the armature ofa magnetic circuit closing device from theposition at which the contacts first touch eachother.

sealing voltage the voltage necessary tocomplete the movement of the armature ofa magnetic circuit closing device from theposition at which the contacts first touch eachother.

search coil a solenoid that is wound withan air core or around a magnet or permeablecomponent of a magnetic circuit to measure


the change of flux within the coil; used witha fluxmeter.

SECAM Seesequential color and memory.

second harmonic component the signalcomponent of a periodic signal whose fre-quency is twice the fundamental frequency.

second order discrete time system a dis-crete system for which the difference be-tween the input and output signals is of sec-ond order.

second order system a continuous timesystem described by a second order differen-tial equation.

second-harmonic generation the processin which a laser beam of frequencyω inter-acts with a material system to produce a beamat frequency 2ωby means of the second-ordersusceptibility. Under carefully controlledcircumstances, more than 50% of the inci-dent intensity can be converted to the secondharmonic.See alsoharmonic generation.

second-order susceptibility a quantity,often designatedχ2, describing the second-order nonlinear optical response of a mate-rial system. It is defined through the re-lation P 2 = Cχ2E2, whereE is the ap-plied electric field strength andP 2 is the2nd-order contribution to the material polar-ization. The coefficientC is of order unityand differs depending on the conventionsused in defining the electric field strength.The second order susceptibility is a tensor ofrank 3, and describes nonlinear optical pro-cesses such as second-harmonic generation,sum- and difference-frequency generation,and optical rectification.See alsononlinearsusceptibility.

secondary (1) the load-side winding.(2) refers to the portion of a nuclear

power plant containing non-radioactive com-ponents such as turbines and generators.

secondary cache a buffer element be-tween slow-speed peripheral devices, such asdisks, and a high-speed computer.

secondary distribution system a distri-bution system in which a significant the sub-division of power to customers is done on thesecondary, or low-voltage side of the distri-bution transformer, as opposed to the prac-tice of assigning a small distribution trans-former to each customer or small group ofcustomers.

secondary memory generic term usedto refer to any memory device that providesbackup storage besides the main memory.Secondary memory is lower-level, larger ca-pacity, and usually a set of disks.

Only data and programs currently used bythe processor reside in main memory. Allother information (not needed at a specifictime) is stored in secondary memory and istransferred to main memory on a demand ba-sis. It is the highest (big but slow) level inthe memory hierarchy of modern computersystems.

secondary resistor a resistor connected tothe rotor of a wound-rotor induction machineto permit variation of the effective rotor re-sistance. By varying the resistance, machinecharacteristics may be optimized for startingor varying load conditions.

secondary selective service a redundantelectric service in a critical load is supplied bytwo distribution transformers, each of whichis served by a separate, independent distribu-tion primary circuit.

secondary service refers to areas servicedby skywaves and not subject to objectionableinterference.

secondary side that side of the packagingand interconnecting structure farthest fromlayer number one. (Also called the soldierside in through-hole component mountingtechnology.)


secondary storage Seesecondary memory.

secondary system of equations a systemof algebraic and differential equations ob-tained from the primary system of equationsby transformation of network variables.

secondary voltage in power distributionwork the voltage at the secondary of the dis-tribution transformer.

secondary voltage control an automaticvoltage control scheme that is similar in func-tion to the automatic voltage regulator, but itspurpose is to control a bus voltage which neednot have a synchronous generator connectedat the bus.

secondary winding the transformer wind-ing to which the loads are connected.Seealsoprimary winding.

sectionalizer a switch placed in distribu-tion lines and programmed to open during aline dead time. The sectionalizer will sensethe presence of current surges due to faults,and is programmed to open after a set numberof faults occur during a short period of time.When the fault is cleared by the protectingrecloser or circuit breaker, the sectionalizerwill open, allowing the recloser or breaker tosuccessfully reenergize the portion of the lineupstream from the sectionalizer.

sectionalizing fuse a sectionalizing fuse isa fuse employed on the primary distributionsystem to isolate laterals from the main feederin the event of a fault on that lateral.

sectionalizing switch a switch on primarydistribution systems used to isolate lateralsand segments of main feeder lines. On radialdistribution systems, sectionalizing switchesare placed to allow rerouting of power to min-imize extended outages following a line seg-ment failure.

sector Seedisk sector.

sector mapping a cache organization inwhich the cache is divided into sectors whereeach sector is composed of a number of con-secutive lines. A complete sector is not trans-ferred into the cache from the memory; onlythe line requested. A valid bit is associatedwith each line to differentiate between linesof the sector that have been transferred andlines from a previous sector. Originally ap-peared in the IBM System/360 Model 85.

sectorization the action of modifying anomnidirectional antenna in a cellular systemso that it is replaced by a number of direc-tional antennas each having a radiation pat-tern approximately covering a sector of a cir-cle. Common examples in cellular systemsare those of sectorization with three sectorsper cell (120 degree sectors) and sectoriza-tion with six sectors per cell (60 degree sec-tors).

security the ability of the power sys-tems to sustain and survive planned and un-planned events without violating operationalconstraints.

security analysis Seecontingency analysis.

SEED Seeself-electro-optic device.

seek time the time that it takes to positionthe read/write device over a desired track ofinformation.

segment a region in computer memory de-fined by a segment base, stored in a segmentbase register and, usually, a segment limit,stored in a segment limit register.See alsovirtual memory.

segment mapping table a memory tablewithin a computer that is used to translate log-ical segment addresses into physical memoryaddresses.


segment register a register that stores thebase, or starting memory address, of a mem-ory segment.

segment table a table that is used to storeinformation (e.g., location, size, access per-missions, status, etc.) on a segment of virtualmemory.

segmentation (1) an approach to vir-tual memory when the mapped objects arevariable-size memory regions rather thanfixed-size pages.

(2) the partitioning of an image in mutu-ally exclusive elements in which visual fea-tures are homogeneous. Region-based seg-mentation relies on the analysis of uniformityof grey level or color; contour-based segmen-tation relies on the analysis of intensity dis-continuities.

segmentation-based coding a codingscheme that is based on segmentation.Seesegmentation.

segmented architecture in computer ar-chitecture, a scheme whereby the computer’smemory is divided up into discontinuous seg-ments.

selective controllability a dynamical sys-tem is said to be selectively controllable withrespect to theith state variable byj th con-trol variable in [t0, t1] if all the dynami-cal systems with scalar admissible controlsuj (t), j = 1,2, . . . , m have theith statevariablexi(t), i = 1,2, . . . , n controllablein [t0, t1].

selective fuse coordination See fusecoordination.

selectivity the ability of a receiver to re-ceive only its desired band of frequencies andreject those on either side. This not only im-proves the signal properties but also the noisecharacteristics of the receiver.

selector channel I/O channel that handlesonly one I/O transaction at a time. Normallyused for high-speed devices such as disks andtapes.See alsomultiplexer channel.

self organizing map (SOM) a neural net-work that transforms a multi-dimensional in-put to a 1- or 2-D topologically ordered dis-crete map. Each neuron has a weight vectorEwj and for a given inputEx, the output is theindex of the neuron with minimum Euclideandistance. During training, the weights are up-dated as

Ewj (n+ 1) ={ Ewj (n)+ α(n)[Ex − Ewj (n)], Cj ∈ N(Ci, n)Ewj (n), Ci 6∈ N(Cj , n)

whereα is a learning parameter in the range0< α < 1, andN(Ci, t) is the set of classeswhich are in the neighborhood of the winningclassCi at timet . Initially, the neighborhoodmay be quite large during training, e.g., halfthe number of classes or more. As the train-ing progresses, the size of the neighborhoodshrinks until, eventually, it only includes theone class.

self-arbitrating bus a communicationbus that is capable of resolving conflictingrequests for access to the bus.

self-bias a technique employed wherebya transistor only needs a single bias sup-ply voltage between the drain terminal andground. This is commonly accomplished byplacing a parallel combination of a resistorand capacitor between the source terminaland ground.

self-biased amplifier an amplifier that uti-lizes a voltage-controlled current source asthe active device (such as a MESFET), andin which a series resistive feedback elementin the DC current path creates the DC voltagerequired to control the quiescent bias point,thereby resulting in the need for a single biassupply.


self-checking pertaining to a circuit, withrespect to a set of faults, if and only if it isfault-secure and self-testing.

self-commutated Seenatural commutation.

self-demagnetizing field a field inside ofa permanent magnet that is opposed to its ownmagnetization, which is due to internal cou-pling of its poles following the introductionof an air gap in the magnetic circuit.

self-electro-optic device (SEED) a bi-stable device that is a PIN photodetector andalso an optical modulator; the intrinsic regionis generally constructed as a quantum-wellstack. Detection of light alters the electri-cal bias on the PIN, which in turn alters thetransmission through the device; the opticaltransmission change exhibits hysteresis anda two-state transmission character.

self-focusing focusing of an electromag-netic beam in a nonlinear medium by the gainor index profile resulting from the action ofthe beam on the medium.

self-generating neural network (SGNN)networks of self-organizing networks, eachnode network of which is an incomplete self-organizing network. For this kind of networkof neural networks, not only the weights ofthe neurons but also the structure of the net-work of neural networks are learned from thetraining examples.

SGNN can be as complex as acyclic di-rected graph, but the most frequently usedSGNN takes a tree structure and is called aself-generating neural tree (SGNT), whichis very similar to self-organizing tree butwith much higher ratio of neuron utilization.Since many fewer neurons participate in thecompetition during the training and classifi-cation, the speed of SGNT is much faster.SGNN has found applications in diagno-sis of communication networks, image/videocoding, large-scale Internet information ser-vices, and speech recognition.See also

self-generating neural tree, self-organizingneural tree.

self-generating neural tree (SGNT) asimplified version of self-generating neuralnetwork with a tree structure. SGNT is nor-mally much faster in training and classifica-tion, but with less descriptive power com-pared with the corresponding SGNN becauseof its simple topological structure. How-ever, if the number of network nodes is thesame, SGNT has the same descriptive power,higher ratio of neuron utilization, higherspeed, and may end up with higher accuracy,since large-scale networks can be generatedand trained quickly.See alsoself-generatingneural network, self-organizing neural tree.

self-modifying code a program using amachine instruction that changes the storedbinary pattern of (usually) another machineinstruction in order to create a differentinstruction which will be executed subse-quently. Definitely not a recommended prac-tice and not supported on all processors.

self-organizing algorithm a training al-gorithm for a self-organizing system consist-ing of the following main steps:

1.Calculate the similarities of the trainingvector to all the neurons in the system andcompare them to find the neuron closest tothe training vector, i.e., the winner.

2. Update the weights of the winner andits neighborhood according to

wi(t + 1) = wi(t)+ α (xi(t)− wi(t)) ,

wherewi(t) is theith weight of the neuronat time t , xi(t) is the ith component of thetraining vector att , andα is a training rate.The neighborhood of the winner starts from abigger area and reduces gradually during thetraining period.

self-organizing neural tree (SONT) atree-like network of self-organizing neuralnetworks, each node of which is a Koho-nen network. Each of the neurons in thehigher level networks has its child network


in the lower level of the tree hierarchy. Thetraining method is similar to that of Koho-nen’s method, but is conducted hierarchi-cally. From the top (root) of the tree down,the winner of the current self-organizing net-work is found as the closest neuron to thetraining example. The weights of the winnerand its neighbors are updated, and then thechild network of the winner will be selectedas the current network for further examina-tion until a leaf node network is encoun-tered. This kind of network of neural net-works can be useful for complex hierarchicalclustering/classification. However, the uti-lization of the neurons may become poor asthe network size growing if the uniform treestructure is adopted. The utilization may beimproved if carefully designed structure isused but how to obtain an optimum struc-ture remains an issue. Self-generating neuralnetwork (SGNN) may be a solution to thisproblem. See alsoself-organizing system,self-generating neural network.

self-organizing system a class of unsu-pervised learning systems that can discoverfor itself patterns, features, regularities, cor-relations, or categories contained in the train-ing data, and organize itself so that the outputreflects these discovery.

self-phase modulation the nonlinear op-tical process in which a pulse of light trav-eling through a material characterized by anintensity-dependent refractive index under-goes spectral broadens as a result of the time-varying phase shift imparted on the beam.

self-pumped phase conjugator (SPPC)a phase conjugator that does not require theexternal pumping beams. Optical four-wavemixing is a convenient method for the gen-eration of phase conjugate waves. However,the process requires a pair of counterprop-agating pump beams. With a self-pumpedphase conjugator, an incident beam willpump the photorefractive medium and gen-erate its own phase conjugate wave. Thereare many physical mechanisms that can yield

self-pumped phase conjugation, for exam-ple, SPPC with resonators, SPPC with semi-resonators, SPPC with stimulated backwardscattering, etc.

self-synchronizing scrambling a tech-nique that attempts to randomize a source bitstream by dividing it by a scrambling polyno-mial using arithmetic from the ring of poly-nomials overGF(2). Descrambling is per-formed with only bit-level synchronizationthrough continuous multiplication of the de-modulated sequence by the same scramblingpolynomial. The division and multiplicationprocedures can be implemented with simpleshift registers, enabling this technique to beused in very high bit rate systems.

self-test a test that a module, either hard-ware or software, runs upon itself.

self-test and repair a fault-tolerant tech-nique based on functional unit active redun-dancy, spare switching, and reconfiguration.

self-testing pertaining to a circuit, for aset of faults, if and only if for any fault inthis set there exists a valid input code suchthat the circuit output is noncorrect (does notbelong to the valid output codes, i.e., can bedetected with a code checker).

Selfoc lens a type of gradient-index lens,using a refractive-index profile across thecross section of the element, typically a cylin-der; the profile is generally produced by im-plantation.See alsogradient index optics.

selsyn Seesynchro.

semaphore a synchronization primitiveconsisting of an identifier and a counter,on which two operations can be performed:lock, to decrease a counter, and unlock, toincrease a counter.

semi-classical model model for the in-teraction of light with atoms or molecules inwhich the atomic or molecular wave func-


tions are described by Schrodinger’s equa-tion, while the electromagnetic fields aredescribed by the Maxwell–Heaviside equa-tions.

semi-guarded machine a machine inwhich some of the ventilating openings, usu-ally in the top half, are guarded as in the caseof a guarded machine to prevent accidentalcontact with hazardous parts, but the othersare left open.

semi-insulator Seesemiconductor.

semi-magnetic semiconductor semicon-ductor alloy or superlattice, usually from theII-VI columns of the periodic table, in whichthere is a concentration of magnetically ac-tive atoms such as manganese that introducenew magneto-optical and magneto-transportproperties.

semi-relative sensitivity one of two sen-sitivity measures used in circuit and systemtheory. These are

Sx(y, x) = x∂y

∂x= ∂y

∂x/x= ∂y

∂lnx

which is frequently denoted bySx , and

Sy(y, x) = 1

y

∂y

∂x= ∂y/y

∂x= ∂lny

∂x

which is also denoted bySy . Both thesequantities are used in a way similar to us-ing relative sensitivity. See alsoabsolutesensitivity, relative sensitivity, sensitivity,sensitivity measure.

semi-rigid cable a coaxial cable with asolid metal outer-conductor. Typically usedwhere the cable is bent to fit the applicationonly once.

semiconductor a material in which elec-trons in the outermost shell are able to mi-grate from atom to atom when a modestamount of energy is applied. Such a materialis partially conducting (can support electri-cal current flow), but also has properties of

an insulator. The amount of current conduc-tion that can be supported can be varied by“doping” the material with appropriate mate-rials, which results in the increased presenceof free electrons for current flow. Commonexamples are silicon and GaAs. Also calledsemi-insulator.

semiconductor device a transistor, resis-tor, capacitor, or integrated circuit made froma semiconductor material.

semiconductor diode laser laser in whichthe amplification takes place in an electricallypumped semiconducting medium.

semiconductor laser Seediode laser.

semiconductor laser amplifier semicon-ductor laser in which feedback from the endfacets or other reflecting surfaces is unimpor-tant.

semiconductor laser oscillator a laser os-cillator in which the amplification takes placein a semiconducting medium.

semiorthogonal wavelets wavelets whosebasis functions in the subspaces are not or-thogonal but the wavelet and scaling sub-spaces spanned by these basis functions areorthogonal to each other.

sense amplifier in a memory system, cir-cuitry to detect and amplify the signals fromselected storage cells.

sensitivity a property of a system indicat-ing the combined effect of component toler-ances on overall system behavior, the effectof parameter variations on signal perturba-tions and the effect of model uncertaintieson system performance and stability. Forexample, in radio technology, sensitivity isthe minimum input signal required by the re-ceiver to produce a discernible output.

The sensitivity of a control system couldbe measured by a variety of sensitivity func-tions in time, frequency or performance do-


mains. A sensitivity analysis of the systemmay be used in the synthesis stage to mini-mize the sensitivity and thus aim for insen-sitive or robust design.See alsorobustness,robust controller design, objective function.

sensitivity bound a lower or upper limiton the sensitivity index for a system.

Such a bound exists, for example, for fil-ters whose passive elements are limited toresistors, capacitors, and ideal transformers,and whose active elements are limited togyrators (characterized by two gyration re-sistances considered different in sensitivitycalculations), and active voltage and currentcontrolled voltage and current sources.

sensitivity function a measure of sensitiv-ity of signals or performance functions due toparameter variations or external signals (dis-turbances, controls). For small changes inparameters, sensitivity functions are partialderivatives of signals or performance func-tions with respect to the parameters and couldbe found from linearized models of the sys-tem model under study.

For linear systems, the sensitivity func-tions in semilogarithmic form could be ob-tained from the original system model insome special nodes of its block scheme calledsensitivity points. Moreover, for linear time-invariant feedback systems, sensitivity maybe defined in frequency domain both forsmall and large parameter deviations. Oneway to do this is to compare the output er-rors due to plant parameter variations in theopen-loop and closed-loop nominally equiv-alent systems. The resulting sensitivity func-tion or for multi-input–multi-output systems,sensitivity matrixS(jω) is given by a returndifference function or its matrix generaliza-tion, i.e.,

S(jω) = (I +K(jω))−1

where K(jω) is an open-loop frequencytransfer function (respectively matrix) andIis a unit matrix of appropriate dimension. Inthe single-input–single-output systems and

for the small variations,S is equal to the clas-sical differential logarithmic sensitivity func-tion defined by Bode.See alsosensitivityindex, sensitivity invariant.

sensitivity index a sensitivity measure.For the multiparameter sensitivity row vectorone can determine the following two charac-teristics:

1. the worst-case sensitivity index

ν(F ) =n∑i=1

∣∣∣ReSF(jω,x)xi

∣∣∣2. the quadratic sensitivity index

ρ(F ) =[

n∑i=1

(ReSF(jω,x)

xi

)2]1/2

These two are used most frequently in sen-sitivity comparisons, yet other indices, forexample, for the case of random circuit pa-rameters, can be derived as well.See alsosensitivity function, sensitivity invariant.

sensitivity invariant feature of a circuitor system that is insensitive.

Let a circuit function beF(s, x) wherex = [x1, x2, . . . , xn]t is a vector of circuitparameters. A sensitivity invariant exists forthis circuit if there is a relationship

n∑i=1

SF(s,x)xi

= k

HereSF(s,x)xi

is a component of a multiparam-eter sensitivity row vector andk is a constant.In most of cases, this constant can have oneof three possible values, namely, 1, 0, and−1. The sensitivity invariants are useful tocheck the sensitivity calculations.See alsosensitivity, sensitivity function, sensitivitybound.

sensitivity measure a number used tocharacterize and compare circuits the func-tions of which depends of more than onecomponent. They are used when the compo-nents of multiparameter sensitivity row vec-tor and the circuits component tolerances are


known. The most frequently used are worst-case measure and quadratic measure of sensi-tivity. See alsosensitivity, sensitivity index.

sensitivity reduction reducing sensitivityis very desirable in active filter realization.Some general suggestions, at the stage of ap-proximation, ensuring that filter realizationswill have low sensitivities to component vari-ations include the following:

1. Increasing the order of approximationand introducing a redundancy.

2. Using predistorted (tapered) specifica-tions in the vicinity of the passband edge.

3. Using transfer functions with a limitedvalue of the maximumQ of the transfer func-tion poles. The realization itself gives the cir-cuits lower sensitivities if the filter is realizedas the doubly terminated lossless structure orits active equivalent.

sensor a transducer or other device whoseinput is a physical phenomenon and whoseoutput is a quantitative measurement of thatphysical phenomenon. Physical phenomenathat are typically measured by a sensor in-clude temperature or pressure to an internal,measurable value such as voltage or current.

sensor alignment alignment of sensors soas to correct the time delay differences arisingfrom spatial differences.

sensorless control a control method inwhich mechanical sensors are replaced by anindirect estimation of the required variable.

separability the separable property for thesignal or system such that the signal or systemrepresentation can be expressed by the prod-uct of component terms, each depending onfewer independent variables.

separable data a 2-D signal that can bewritten as a product of two 1-D signals.

separable filter a filter that can be ap-plied in two or more operations without anychange in overall function, thereby gaining

some computational advantage. In particu-lar, a 2-D mean filter can be re-implementedidentically as two orthogonal 1-D mean fil-ters, and is therefore separable. However,a 2-D median filter is non-separable, as itsaction is not in general identical to that oftwo orthogonal 1-D median filters.See alsoseparable transform.

separable image transform a 2-D sepa-rable transform used to transform images.

separable kernel for a 2-D transform akernel that can be written as the product oftwo 1-D kernels. For higher dimension trans-forms a separable kernel can be written as theproduct of several 1-D kernels.Seeseparabletransform.

separable transform a 2-D transform thatcan be performed as a series of two 1-D trans-forms. In this case the transform has a separa-ble kernel. The 2-D continuous and discreteFourier transforms are separable transforms.In higher dimensions a separable transformis one that can be performed as a series of1-D transforms.See alsoseparable filter.

separately excited DC machine a DCmachine where the field winding is suppliedby a separate DC source. Separately excitedgenerators are often used in feedback con-trol systems when control of armature volt-age over a wide range is required.

sequency in a transform, the number ofzero-crossings of a particular basis function.By extension used to refer to the transformcoefficient that corresponds to a particularbasis function. For example, in the discretecosine transform the zero sequency coeffi-cient is the one for which the basis functionis flat (and therefore has no zero crossings),often called the DC coefficient. Sequency isroughly analogous to frequency; higher se-quency basis functions correspond to higherfrequency components of signal energy.


Electrical Engineering Dictionary - Shariflcsl/useful/Electrical Engineering...Electrical...

Documents

Transcript of Electrical Engineering Dictionary - Shariflcsl/useful/Electrical Engineering...Electrical...