Haykin_index.

fm Page 3 Tuesday, January 8, 2013 2:06 PM

Index I3

Colored noise channels, information D Dirac delta function, 28–33

capacity Decision-feedback equalization, Discrete Fourier transform (DFT).
capacity of NEXT-dominated 473–474 See DFT (discrete Fourier
channel, 252–253 Delta modulation (DM). See DM transform).
introduction, 248–252 (delta modulation). Discrete memoryless channels
Communication process DFT (discrete Fourier transform). See binary symmetric channel, 225
digital communication, 9–11 also IDFT (inverse discrete introduction, 223–225
introduction, 2–4 Fourier transform). Discrete memoryless channels, error-
multiple-access techniques, 4–5 binary sequence for energy control coding
networks, 6–9 calculation, 19–21 channel coding theorem, 580–581
Composite hypothesis testing, Dirac delta function, 28–33 introduction, 579–580
132–133 interpreting, 70–72 notation, 582
Compound probabilistic codes, introduction, 16–19 Discrete multicarrier transmission
introduction, 644–645 linear time-invariant systems, (DMT). See DMT (discrete
Compound probabilistic codes, 37–41 multicarrier transmission).
LDPC codes pairs, 24 Distortionless baseband data
(10, 3, 5) codes, 669–671 periodic signals, 34–36 transmission, 450–454
introduction, 646–669 theorems, 23 Distribution functions, Bernoulli
irregular codes, 674–675 time functions, 24 random variable, 101–105
minimum distance, 671–672 unit Gaussian pulse, 21–22 DM (delta modulation)
probabilistic decoding, 672–674 DFT (discrete Fourier transform), adaptive DM, 308
Constrained optimization problem, DMT systems introduction, 305
484–487 description, 489–491 quantization errors, 307–308
Convolutional codes. See Error- DFT-based DMT systems, receiver, 307
control coding for convolutional 492–493 transmitter, 305–307
codes. DMT-based DSL, practical DMT (discrete multicarrier
Convolutional interleaving, A32–A33 applications, 493–494 transmission) system, DFT
Cosine transformation of a random frequency-domain channel description, 489–491
variable, 109–112 descriptions, 491–492 DFT-based DMT systems,
Cross-correlation functions, introduction, 487–489 492–493
autocorrelation function, DFT (discrete Fourier transform), DMT-based DSL, practical
155–157 numerical computation applications, 493–494
Cross-spectral densities property, computing the IDFT, 77–78 frequency-domain channel
weakly stationary stochastic FFT algorithms, 72–77 descriptions, 491–492
processing, 172–174 interpretation of DFT and IDFT, introduction, 487–489
Cyclic codes, error-control coding 70–72 DMT (discrete multicarrier
calculating the syndrome, introduction, 69–70 transmission) system, loading,
598–599 DFT-based DMT systems, 492–493 482–484
cyclic property, 593 Differential entropy DPCM (differential pulse-code
encoding, 597–598 mutual information, 237–240 modulation)
generator matrices, 596–597 uniform distribution, 238–240 DPCM receiver, 303
generator polynomials, 594–595 Differential phase-shift keying DPCM transmitter, 303
hamming codes, 599–603 (DPSK). See DPSK (differential introduction, 301–303
introduction, 593–594 phase-shift keying). processing gain, 304
linearity property, 593 Differential pulse-code modulation DPSK (differential phase-shift
maximal-length codes, 603–604 (DPCM). See DPCM keying). See also PSK (phase-
parity-check matrices, 596–597 (differential pulse-code shift keying), introduction.
parity-check polynomials, modulation). error probability, 412–413
595–596 Digital communication introduction, generating DPSK signals, 413
properties, 593 9–11 illustration, 412
Reed-Solomon codes, 604–605 Digital subscriber lines (DSL). See introduction, 411–412
DSL (digital subscriber lines). optimum receiver, 413–415
Haykin_index.fm Page 4 Tuesday, January 8, 2013 2:06 PM

I4 Index

DSB-SC modulation, 60–61 UMTS turbo decoder, 653–657 generator polynomials, 594–595
DSL (digital subscriber lines) UMTS with binary PSK hamming codes, 599–603
band-limited channels, 475–477 modulation, 653–657 introduction, 593–594
DMT-based, practical Error-control coding for linearity property, 593
applications, 493–494 convolutional codes maximal-length codes, 603–604
E code tree, 607–608 parity-check matrices, 596–597
Entropy convolutional encoder, 606–607 parity-check polynomials,
Bernoulli random variable, introduction, 605–606 595–596
211–212 optimum decoding, 613–614 properties, 593
differential, 237–240 recursive systematic, 611–613 Reed-Solomon codes, 604–605
of extended source, 213–214 state diagrams, 609–611 Error-control coding for discrete
extension of a discrete trellis graph, 609. See also memoryless channels
memoryless source, 212–213 Trellis-coded modulation. channel coding theorem, 580–581
introduction, 207–209 Error-control coding for introduction, 579–580
properties of, 209–211 convolutional codes, maximum notation, 582
relative, 238–239 a posteriori probability decoding Error-control coding for linear coding
Envelopes algorithmic metrics, 627–628 blocks
band-pass signals, complex AWGN channel, branch metric hamming codes, 590–592
envelopes, 47–49 evaluation, 630–634 introduction, 582–585
low-pass signals, 47 BCJR algorithm, 623–624, 638 minimum distance considerations,
narrowband noise, 191–193 forward-backward recursions, 587–589
pre-envelopes, 45–47 626–630 syndrome decoding, 589–590
Equal gain combining, 538 introduction, 623–624 syndrome definition and
Ergodic processes, weakly stationary lattice-based framework for the properties, 585–587
stochastic processing, 157–158 derivation, 625–626 Exit charts
Error rates in band-limited channels log-MAP algorithm, 636–638 approximate Gaussian model,
due to channel noise in matched- MAP decoding algorithm, 661–663
filter receivers, 446–447 624–625, 635–638 developing, 658–661
Error-control coding. See also max-log-MAP algorithm, histogram computation method,
Compound probabilistic codes. 636–638, 639–644 663–666
forward error correction, 578–579 a posteriori L-value, finalizing, introduction, 657–658
introduction, 577–578 634 measuring, 664–666
LDPC codes. See LDPC (low- Error-control coding for Expectation
density parity-check) codes. convolutional codes, maximum introduction, 105–106
Error-control coding, exit charts a posteriori probability max- linearity, 107–108
approximate Gaussian model, decoding, 636–638 statistical independence, 108
661–663 Error-control coding for Exponential distribution, 110–111
developing, 658–661 convolutional codes, maximum Eye patterns
histogram computation method, likelihood decoding for binary systems, 467–469
663–666 asymptotic coding gain, 622–623 introduction, 463–464
introduction, 657–658 correct decoding of received all- for M-ary transmissions, 466
measuring, 664–666 zero sequences, 617–618 peak distortion for intersymbol
Error-control coding, turbo coding free distance, 620–621 interference, 465–466
extrinsic information, 649–650 incorrect decoding of received all- for quaternary systems, 467–469
introduction, 645–646 zero sequences, 619 timing features, 464
mathematical feedback analysis, introduction, 614–616 F
651–653 Viterbi algorithm, 616–617, 623 Fading channels
performance, 648–649 Error-control coding for cyclic codes comparison of modulation
serial concatenated codes, calculating the syndrome, schemes, 525–527
681–687 598–599 diversity techniques, 525
turbo decoder, 650–651 cyclic property, 593 effects of flat fading, 525–527
two-state encoder, 646–648 encoding, 597–598 introduction, 501–502
generator matrices, 596–597
Haykin_index.fm Page 5 Tuesday, January 8, 2013 2:06 PM

Index I5

propagation effects, 502–505 Fading channels, space diversity-on- Rayleigh processes, 524
RAKE receiver and multipath transmit receive systems Rician-Jakes doppler spectrum
diversity, 564–566 Alamouti code, 540–541 model, 524–525
Fading channels, code division full-rate complex code, 541 Fourier series, 13–16
multiple access introduction, 538–539 Fourier transform. See DFT (discrete
Gold codes, correlation properties, linearity, 542–546 Fourier transform); IDFT
563–564 maximum likelihood decoding, (inverse discrete Fourier
Gold sequences, 562–563 545–546 transform).
introduction, 560–561 QPSK (quadriphase-shift keying), Frequency-domain
Walsh-Hadamard sequences, 539 description, 56–58, 268–271
561–562 receiver considerations, Alamouti relation to time-domain, 25–28
Fading channels, FIR modeling of code, 542–545 FSK (frequency-shift keying). See
doubly spread channels unitarity (complex orthogonality), also AWGN channel signaling.
generating tap coefficients, 541 introduction, 375–377
523–524 Fading channels, spread spectrum noncoherent detection of binary
introduction, 520–523 signals FSK, 410–411
practical matters, 523 classification of spread spectrum FSK (frequency-shift keying)
Rayleigh processes, 524 signals, 557–558 coherent detection. See also
Rician-Jakes doppler spectrum introduction, 557–558 PSK (phase-shift keying),
model, 524–525 processing gain of the DS/BPSK, introduction.
Fading channels, Jakes model 559 bandwidth efficiency, M-ary FSK
illustrative generation of fading Fading channels, statistical signals, 396–397
processes, 510–511 characterization of wideband M-ary FSK, introduction,
implemented as a FIR filter, wireless channels 395–397
509–511 classification of multipath M-ary FSK versus M-ary PSK,
introduction, 506–509 channels, 519–520 398–399
Fading channels, MIMO capacity Doppler power spectrum, minimum shift keying, 382–383
channel known at the transmitter, 517–519 phase trellis, 383–384
555–556 introduction, 511–512 power spectra, M-ary FSK signals,
ergodic capacity, 551–553 multipath correlation function of 396
log-det formula capacity, the channel, 512 FSK (frequency-shift keying)
553–554 power-delay profile, 516–517 coherent detection, binary FSK
outage capacity, 554–555 scattering function of the channel, error probability, 378–380
Fading channels, MIMO systems 514–516 generation and detection, 377–378
basic baseband channel model, spaced-frequency, spaced-time power spectra, 380–382
547–551 correlation function of the FSK (frequency-shift keying)
CAI (coantenna interference), channel, 514 coherent detection, MSK
546–547 uncorrelated scattering, 513 error probability, 390–391
introduction, 546 wide-sense stationarity, 512–513 Gaussian filtering, 392–395
Fading channels, OFDM FFT (fast Fourier transform) generation and detection, 389–390
introduction, 556 algorithms, 72–77 power spectra, 391–392
PAPR problem, 556–557 Filtering two jointly weakly signal-space diagram, 384–388
Fading channels, space diversity-on- stationary processes, 174 waveforms, 388–389
receive systems FIR (finite-duration impulse Full-rate complex code, 541
equal gain combining, 538 response) modeling, G
introduction, 528 introduction, 456–458 Gaussian distribution
maximum-ratio combining, FIR (finite-duration impulse introduction, 113
533–537 response) modeling of doubly jointly Gaussian random
outage probability for maximal- spread fading channels variables, 116
ratio combiner, 537 generating tap coefficients, linear function of a Gaussian
outage probability of selection 523–524 random variable, 114
combiner, 532 introduction, 520–523 mean, 114
selection combining, 528–532 practical matters, 523
Haykin_index.fm Page 6 Tuesday, January 8, 2013 2:06 PM

I6 Index

Gaussian distribution (cont.) introduction, 248–252 unipolar NRZ signaling, 311

random variables, 239–240 Information capacity law unipolar RZ signaling, 311
standard distribution, table of, 117 capacity of binary-input AWGN Linear coding blocks, error-control
sum of independent Gaussian channels, 244–248 coding
random variables, 114 implications of, 244–248 hamming codes, 590–592
variance, 114 introduction, 240–243 introduction, 582–585
Gaussian process PCM noise, 292–294 minimum distance considerations,
independence, 179 sphere packing, 243–244 587–589
introduction, 176–177 Information theory, history of, 1–2 syndrome decoding, 589–590
linear filtering, 177–178 Integrals, table of, A57 syndrome definition and
multivariate distribution, 178 Interleaving properties, 585–587
stationarity, 179 block, A30–A32 Linear function of a Gaussian random
Geometric representation of AWGN convolutional, A32–A33 variable, 114
channel signals introduction, A29–A30 Linear modulation theory
2B1Q code, 331–332 random, A33–A34 DSB-SC modulation, 60–61
Gram-Schmidt orthogonalization, Intersymbol interference, band- introduction, 58–60
329–331 limited channels, 447–449 SSD modulation, 64–66
introduction, 324–328 Inverse discrete Fourier transform summary of modulation methods,
Schwarz inequality, 328–329 (IDFT). See IDFT (inverse 66
Gold codes, correlation properties, discrete Fourier transform). VSB modulation, 61–64
563–564 J Linear time-invariant filter,
Gold sequences, 562–563 Jakes model, fading channels transmitting weakly stationary
Gram-Schmidt orthogonalization, illustrative generation of fading stochastic processing, 158–160
329–331 processes, 510–511 Linear time-invariant systems, 37–41
Group delays, 66–69 implemented as a FIR filter, Linearity, expectation, 107–108
H 509–511 LMS (least-mean-square) algorithm,
Hilbert transform introduction, 506–509 470–472
introduction, 42–44 Jointly Gaussian random variables, Log-normal distribution, A3–A6
low-pass signals, 44–45 116 Lossless data compression algorithms
Huffman coding, lossless data K Huffman coding, 219–220
compression, 219–220 Kraft inequality, lossless data Huffman tree, 220–221
Huffman tree, lossless data compression, 217–219 introduction, 215–216
compression, 220–221 Kraft inequality, 217–219
L Lempel-Ziv coding, 221–223
Hypothesis testing Lagrange multipliers, A19–A20
binary, 130–132 prefix coding, 216–217
LDPC (low-density parity-check) Low-density parity-check (LDPC).
composite, 132–133 codes
introduction, 126–129 See LDPC (low-density parity-
(10, 3, 5) codes, 669–671 check).
I history of, 645
Ideal band-pass filtered white noise, Low-pass signals
introduction, 646–669 envelopes, 47
189–190 irregular codes, 674–675
Ideal low-pass filtered white noise, Hilbert transform, 44–45
minimum distance, 671–672
181–182 probabilistic decoding, 672–674 M
Ideal Nyquist pulse, band-limited Least-mean-square (LMS) algorithm, Manchester code, 311
channels, 450–454 470–472 MAP (maximum a posteriori
IDFT (inverse discrete Fourier Lempel-Ziv coding, lossless data probability) decoding algorithm,
transform). See also DFT. compression, 221–223 624–625, 635–638
computing, 77–78 Line codes Mathematical tables
interpreting, 70–72 bipolar RZ signaling, 311 integrals, A57
Information capacity, colored noise introduction, 309–310 series expansions, A56
channels Manchester code, 311 trigonometric identities, A55
capacity of NEXT-dominated polar NRZ signaling, 311 unit prefixes, A58
channel, 252–253 split phase, 311 useful constants, A58
Haykin_index.fm Page 7 Tuesday, January 8, 2013 2:06 PM

Index I7

Maximal-length sequences ideal band-pass filtered white P

choosing, A50–A54 noise, 189–190 PAM (pulse-amplitude modulation),
code generation, A51–A54 introduction, 183–189 274–277
correlation property, A48–A50 phase components, 191–193 PAPR (peak-to-average power ratio)
introduction, A47 plus sine wave, 193–195 problem
properties of, A47–A50 Rayleigh distribution, 192–193 clipping-filtering, PAPR
Maximum likelihood decoding, Rician distribution, 194–195 reduction, A37–A38
545–546 Networks, introduction, 6–9 fading channels, 556–557
Maximum-ratio combining, 533–537 NEXT-dominated channel, capacity introduction, A35
Mean functions, weakly stationary of, 252–253 maximum PAPR using M-ary
stochastic processing, 149–157 Noise. See also Narrowband noise; PSK, A36–A37
Mean-square value property White noise. properties of OFDM signals,
autocorrelation function, 151 definition, 179 A35–A36
weakly stationary stochastic shot, 180 Parameter estimation
processing, 164 thermal, 180 in additive noise, 124–125
Method of Lagrange multipliers, Noise, PCM introduction, 122–124
A19–A20 error threshold, 291–292 Partitioning continuous-time
MIMO (multiple input, multiple information capacity law, channels
output) capacity, fading channels 292–294 geometric SNR, 481–482
channel known at the transmitter, introduction, 290–291 introduction, 478–481
555–556, A24–A28 Noncoherent detection, binary FSK, loading the DMT system,
ergodic capacity, 551–553 410–411 482–484
log-det formula capacity, Noncoherent orthogonal modulation, PCM (pulse-code modulation)
553–554, A21–A24 AWGN channel signaling, encoding the transmitter, 288
outage capacity, 554–555 404–410 introduction, 285–286
Mixing random processes with Nonlinear solid-state power inverse operations in the receiver,
sinusoidal, weakly stationary amplifiers, A39–A43 288–289
stochastic processing, 167–169 Nonnegativeness property, weakly quantization of the transmitter,
Monotonicity of the distribution, 99 stationary stochastic processing, 286–288
Monte Carlo integration, A45–A46 164 regeneration along the transmitter
m-sequences. See Maximal-length Nonnegativity function, 99 path, 288–290
sequences. Normalization function, 99–100 PCM (pulse-code modulation), noise
MSK (minimum shift keying), FSK Normalization property considerations
coherent detection autocorrelation function, 152 error threshold, 291–292
error probability, 390–391 weakly stationary stochastic information capacity law,
Gaussian filtering, 392–395 processing, 165 292–294
generation and detection, 389–390 O introduction, 290–291
power spectra, 391–392 OFDM (orthogonal frequency Periodic signals, Fourier transform,
signal-space diagram, 384–388 division multiplexing), PAPR 34–36
waveforms, 388–389 problem Phase components, narrowband
Mutual information clipping-filtering, PAPR noise, 191–193
continuous random ensembles, reduction, A37–A38 Phase delays, 66–69
237–240 fading channels, 556–557 Phase-shift keying (PSK). See PSK
differential entropy, 237–240 introduction, A35 (phase-shift keying).
expansion, 228–229 maximum PAPR using M-ary Poisson process, weakly stationary
introduction, 226–227 PSK, A36–A37 stochastic processing, 174–176
nonnegativity, 228 properties of OFDM signals, Polar NRZ signaling, 311
symmetry, 227–228 A35–A36 Prediction-error filtering, redundancy
N Outage probability reduction
Nakagami distribution, A6–A9 for maximal-ratio combiner, 537 discrete time structure for
Narrowband noise of selection combiner, 532 predictions, 296–299
envelope, 191–193 introduction, 294–295
Haykin_index.fm Page 8 Tuesday, January 8, 2013 2:06 PM

I8 Index

Prediction-error filtering, redundancy Probability theory, second-order M-ary PSK versus M-ary FSK,
reduction (cont.) statistical averages 398–399
linear adaptive prediction, cosine transformation of a random PSK (phase-shift keying) coherent
300–301 variable, 109–112 detection, power spectra
theoretical considerations, exponential distribution, 110–111 binary PSK, 356–357
295–296 introduction, 108–109 M-ary PSK, 367–370
Pre-envelopes, 45–47 Processing gain, DPCM, 304 PSK (phase-shift keying) coherent
Prefix coding, lossless data Properties, weakly stationary detection, QPSK
compression, 216–217 stochastic processing error probability, 362–364
Probabilistic compound codes. See cross-spectral densities, 172–174 introduction, 357–359
Compound probabilistic codes. filtering two jointly weakly offset QPSK, 365–367
Probabilistic model, 90–97 stationary processes, 174 power spectra, 364–365
Probability theory introduction, 160–161, 170–172 signal-space diagrams, 358–359
characteristic function, 112–113 mean-square value of stationary waveforms, 359–365
introduction, 87–90 process, 164 PSK (phase-shift keying) coherent
probabilistic model, 90–97 mixing random processes with detection, signal-space diagrams
random variables, 97–98 sinusoidal, 167–169 binary PSK, 353–354
set theory, 88–90 nonnegativeness, 164 QPSK signals, 358–359
Probability theory, central limit normalization, 165 Pulse-amplitude modulation (PAM),
theorem random binary wave, 166–167 274–277
introduction, 118 sinusoidal wave with random Pulse-code modulation (PCM). See
sum of uniformly distributed phase, 165–166 PCM (pulse-code modulation).
random variables, 118–119 sum of two processes, 173 Q
Probability theory, distribution symmetry, 164 QAM (quadrature amplitude
functions Wiener-Khintchine theorem, modulation)
Bernoulli random variable, 169–170 average probability of error,
101–105 zero correlation among frequency 373–375
boundedness of the distribution, components, 162–163 introduction, 370–371
98 zero-frequency value, 164 M-ary QAM for M = 4, 371–373
introduction, 98 PSK (phase-shift keying), square constellations, 371
monotonicity of the distribution, introduction, 352. See also Q-function, bounds on, A11–A12
99 AWGN channel signaling; QPSK (quadriphase-shift keying),
nonnegativity, 99 DPSK (differential phase-shift PSK coherent detection
normalization, 99–100 keying); FSK (frequency-shift error probability, 362–364
uniform distribution, 100–101 keying). introduction, 357–359
Probability theory, expectation PSK (phase-shift keying), M-ary offset QPSK, 365–367
introduction, 105–106 QAM power spectra, 364–365
linearity, 107–108 average probability of error, signal-space diagrams, 358–359
statistical independence, 108 373–375 waveforms, 359–365
Probability theory, Gaussian introduction, 370–371 QPSK (quadriphase-shift keying),
distribution for M = 4, 371–373 space diversity-on-transmit
introduction, 113 QAM square constellations, 371 receive systems, 539
jointly Gaussian random square constellations, 371 Quadrature-modulated processes,
variables, 116 PSK (phase-shift keying) coherent autocorrelation function,
linear function of a Gaussian detection 156–157
random variable, 114 binary phase-shift keying, Quantization
mean, 114 352–357 errors, delta modulation, 307–308
standard distribution, table of, 117 error probability, binary PSK, introduction, 278–279
sum of independent Gaussian 354–356 noise, 279–281
random variables, 114 introduction, 352 scalar quantizers, optimality,
variance, 114 M-ary PSK, introduction, 282–285
367–370 sinusoidal modulating signal,
281–282