Solutions Manual for Communication Systems 4th Edition Simon Haykin McMaster University, CanadaPreface This Manual is written to accompany the fourth edition of my book on Communication Systems. It consists of the following: * Detailed solutions to all the problems in Chapters 1 to 10 of the book + MATLAB codes and representative results for the computer experiments in Chapters 1, 2, 3, 4, 6, 7,9 and 10 I would like to express my thanks to my graduate student, Mathini Sellathurai, for her help in solving some of the problems and writing the above-mentioned MATLAB codes. I am also grateful to my Technical coordinator, Lola Brooks for typing the solutions to new problems and preparing the manuscript for the Manual. Simon Haykin Ancaster April 29, 2000CHAPTER I roblem As an illustration, three particular sample functions of the random process X(t), corresponding to F = W/4, W/2, and W, are plotted below: sin (2nWe) t arty sin(an # t inton & sin(an ft) t -2 ° 2 Ww Ww To show that X(t) is nonstationary, we need only observe that every waveform illustrated above is zero at t = 0, positive for 0 < t < 1/M, and negative for -1/2W < t < 0. ‘Thus, the probability density function of the random variable X(t,) obtained by sampling X(t) at t, = 1/tW is identically zero for negative argument, whereas the probability density function of the random variable X(t) obtained by sampling X(t) at t = -1/IW is nonzero only for negative arguments. Clearly, therefore, fe? # fee) "D> and the random process X(t) is nonstationary.Problem 1.2 X(t) = A cos(2nf,t) efore, X, = A cos(2nf,t,) Since the amplitude A is uniformly distributed, we may write 1 waar 0 < ¥, < costar t,) fy (x) = cra °, otherwise Fx er cos (2mEt, = 4 ° cos (2r£_t,) Similarly, we may write Xi, =A coslant (ty) and 1 {aaah O ) = costam(t -tp)I >) of the procs Since every weighted sum of the samples, Z(t) is Gaussian, it follows that Z(t) is a Gaussian process. Furthermore, we note that 2 2 eae) 7 EC = 1 This result is obtained by putting t,=t, in Eq. (1). Similarly, 2 2 E(Z2 e P28) * ElZ*(t,)] = 1 Therefore, the correlation coefficient of Z(t,) and Z(tp) is Covl2( t4)20 ty I @ 2 4)72(ty) Hence, the joint probability density function of Z(t,) and 2(tp) Fo¢ 4) 52g) BrP Be) = © exl-RC2 4422) where 1 awh ~cos"2e (t=t9)] 1 = Es By sintan(e,-ta) Q(24425) 2 sin“L2n(ty-to)](>) We note that the covariance of Z(t,) and Z(t) depends only on the time difference tyrtg. The process Z(t) is therefore widessense stationary. Since it is Gaussian it is also strictly stationary. Problem 1.5 @) Let X(t) = A + ¥(t) where A is a constant and Y(t) is a zero-mean random process. The autocorrelation function of X(t) is * Ry(e) = ELK(ter) X(t)] E((A + ¥(ter)] [A + ¥(t)]} Eta? + A Y(ter) +A Y(t) + Y(ter) Y(t) a RCD Which shows that Ry(t) contains a constant component equal to A°. (by Let X(t) = Ay cos(2uf .t + 6) + Z(t) where Ay cos(2nf .t+é) represents the sinusoidal component of X(t) and @ is a random phase variable. The autocorrelation function of X(t) is R(t) = E(K(ter) X(t)I EUIA, cos(2nft + 2nfjx + 6) + Z(t+r)] (A, cos(2nft + 8) + Z(t)]} ELAZ cos(2af,t + 2nfyr + 6) cos(2ift + 8)1 + ElZ(ter) Ay cos(2nf + 6)] + ELA, cos (anf t + ant. + 6) 2(t)] + EZ (ter) 2¢t)) 2 (Ag/2) cost2nt.t) + Ry(x) which shows that Ry(t) contains a sinusoidal component of the same frequency as X(t). Problem 1.6 (a) We note that the distribution function of X(t) isoO, x <0 Ad rAd Ace Faces® and the corresponding probability density function is 1 1 f(g) = FOX) + 3 Cx - A) which are illustrated below: a oe 1.0 (b) By ensemble-averaging, we have EIX(t)] = Sx fycgy (2) ax £ x Oh 6G) +b oex - a0) ax A - 2 The autocorrelation function of X(t) is Ry(t) = E(X(ter) X¢t)] Define the square function Sqq (t) as the square-wave shown below: 0Sa, (t) 1.0 Then, we may write Ryle) = EIA Sag Ct - 1 Say (t= ty +t) Say (t= ty) + goat, 5 ‘a Ty a * Ta 7 4 ea 24h vic 2. 2 es 2 Since the wave is periodic with period Tg, Ry(t) must also be periodic with period Ty. (c) On a time-aver aging basis, we note by inspection of Fig. P/.4that the mean is ext = £ Next, the autocorrelation function Ty/2 a x(ter) x(t) dt Tye L 1 Kx(ter)x(t)> = has its maximum value of A°/2 at t = 0, and decreases linearly to zero at t = Tg/2. Therefore, 2 emcees éxtter) x(t)? #5 = 2 0Again, the autocorrelation must be periodic with period Ty. (4) We note that the ensemble-aver aging and time-averaging procedures yield the same set of results for the mean and autocorrelation functions. Therefore, X(t) is ergodic in both the mean and the autocorrelation function. Since ergodicity implies wide-sense stationarity, it follows that X(t) must be wide-sense stationary. Problem 1.7 (a) For tr] > T, the random variables X(t) and X(ter) occur in different pulse intervals and are therefore independent. Thus, E(X(t) X(ter)) = E(X(t)] E(K(ter)], lot Since both amplitudes are equally likely, we have E(X(t)] = Elx(ter)] for It] > T, 4/2, Therefore, 2 A Rt) =e For It] <1, the random variables occur in the same pulse interval if t, s ) Form the non-negative quantity EUUXCtar) + ¥(t)}7) = ELK? (ter) + ax(ter) Y(t) + ¥2(t)) E(x*(ter)] + 26(x(ter) ¥(t)) + ELY@(t] RCO) + ARyy(t) + RyCO) Hence , Ry(O) + Aye) + Ry(O) > 0 or Wy 1 ) Since Ryy(t) = Ryy(-t), we haveRyy(t) =F nCu) Ry(-r-u) du Since Ry(t) is an even function of t: Ryy(t) = nCu) Ry(t+u) du Replacing u by -u: Rye) =F now) Ry(e-u) du (c) If X(t) is a white noise process with zero mean and power spectral density No/2, we may write z Rye) = 260) Therefore, Rye) = ges aw) 6-u) du Using the sifting property of the delta function: N ‘0 Ryy(t) 2 xe hte) That is, 2 h(t) = Ge Ryy (1) Ny “yx This means that we may measure the impulse response of the filter by applying a white Power noise of, spectral density No/2 to the filter input, cross-correlating the filter output with the input, and then multiplying the result by 2M. Problem 1.12 (a) The power spectral density consists of two components: (1) A delta function 6(t) at the origin, whose inverse Fourier transform is one. (2) A triangular component of unit amplitude and width 2 oq, centered at the origin; the inverse Fourier transform of this component is fy sine@(f gt). Therefore, the autocorrelation function of X(t) is 13Ryle) = 1 + fq sine@(t ye) Wnieh is sketched below: R(t) T 1 1 i fy L fo (b) Since Ry(t) contains a constant component of amplitude 1, it follows that the de power contained in X(t) is 1. (c) The mean-square value of X(t) is given by EIX?(t)] = Ry (0) F14fy The ac power contained in X(f) is therefore equal to fo (4) If the sampling rate is f>/n, where n is an integer, the samples are uncorrelated. They are not, however, statistically independent. ‘They would be statistically independent if X(t) were a Gaussian process. ler The autocorrelation function of no(t) is + By Ce getg) = Elnatty) nat] E{fn (ty) cos(anfgt, +0) = ny(t,) sin(2nt.t40)] + [nj(tp) cos(arfgt, +0) - nyt.) sin(arf.t, + 0)]) Eln,(t,) nj(ty) cos(anf.t, +6) cos(anr.t, + 6) = nylt,) ny(ty) cos(arf,t, +0) sin(arf.t, + @) = nytty) ny(tp) sin(2nt.t, +6) cos(arf.t, +6) 14+ ny(ty) nylty) sin(ant ty +8) sin(art.t, +6)) = E{ny(ty) ny(to) coslarty(t yt.) = ny(ty) ny(tg) sinlant, (tty) + 26)) = Elny(ty) ny(to)] coslart,(t j-t3)1 ~ Elny(t,) ny(ty)] + Elsinfant,(t +t.) + 20]} Since © is a uniformly distributed random variable, the second term is zero, giving Ry (tyta) = Ry (tyty) coslant(t to] 2 1 2) Since n,(t) is stationary, we find that in terms of t tyrtet Ry @) Ne By cos (af ,t) Taking the Fourier transforms of both sides of this relation: Sy (f) = 15 (fet) + Sy, (F893 1 With Sy (f) as defined in Fig. Pfy% we find that Sy (f) is as show below: 1 2 20 20 MpaProblem 1.14 The power spectral density of the random telegraph wave is S(t) =F Ryle) expl-Jertt) a o =f exp(2vt) exp(-janft) de +f expl-2ur) exp(-Jjarfr) dr 0 0 1 = Spape) leepl@vr - Senfe)) 1 ~ Bwaget) Lexar - gered) 5 © 20=5nf) * 204jnt) van The transfer function of the filter is HOD) = 79a T+ jarf rc Therefore, the power spectral density of the filter output is 2 SyCf) = THCA I? scr) [1 + (nfRC)?1 an) To determine the autocorrelation function of the filter output, we first expand Sy(f) in partial fractions as follows v 1 — a - . Taner? Cyan) ete ve Sy(f) = Recognizing that 15v 2 yee exp(-2t tt) > v2 snr 172Rc 2, 2pe exp(-1t{ /RC) => (1 /2RC) “ag We obtain the desired result: L expe In) = 2Rc exp #1) Ryle) = Kr 5 1-4R 70% S 16Problem 1.15 We are given y= fsa For x(t) = 6(2), the impulse response of this running integrator is, by definition, a(n =f 8(t)de 1 =1 for 1-TSOSt or, equivalently, 0S1 Next, we note that Ry(x) is related to the power spectral density by Ry) = f Sy(f) cos(anft) df (2) power Therefore, comparing Eqs. (1) and (2), we deduce that the, spectral density of X(t) is When the frequency assumes a constant value, f, (say), we have e a1 1 fpf) = 5 b(t.) + 3 6( fet) Ci 19and, correspondingly, 2 2 A A Sy() = FF Olof.) + ster) Problem L18 Let 6x? denote the variance of the random variable X, obtained by observing the random process X(t) at time t,. The variance ,? is related to the mean-square value of X,, as follows ox = IX?) - ne wherefx = EX]. Since the process X(t) has zero mean, it follows that of = EX?) Next we note that EIX{] = [7 Sx(Oaf We may therefore define the variance ox” as the total area under the power spectral density Sx(f) as oe £ Sx(Oat @ ‘Thus with the mean px = 0 and the variance o, defined by Eq. (1), we may express the probability density function of X, as follows fx) = me oian ox 20x 20Problem 1.19 The input-output relation of a full-wave rectifier is defined by { X(tyy— X(ty) 20 [xii xt) <0 Y(t) = XC Y= Tne probability density function of the random variable X(t,), obtained by observing the input random process at time t,, is defined by To find the probability density function of the random variable Y(t,), obtained by observing the output random process, we need an expression for the inverse relation defining X(t.) in terms of Y(t,). We note that a given value of Y(t,) corresponds to 2 values of X(t,), of equal magnitude and opposite sign. We may therefore write X(ty) = KCI, KC) <0 XE, = Wb, Ky > 0. In both cases, we have - px | aes) eae ‘The probability density function of Y(t,) is therefore given by ! X(t, ) | Bore fy OY) = fyq s(x = oy) [ae]. (ezys X(t) X(t) art) XC) j He,which is illustrated below: fa) Ye) 0.791 Problem 1.20, (a) Tne probability density function of the random variable Y(t,)+ obtained by observing the rectifier output Y(t) at time t,, is (2 1 exp(- dy 20 (z oy cc aa f, (y) = X(t) oOo, y, is given by by) = J X(tga) hy(u) du The expected value of Z(t») is my, * Elatt,)1 HAO) my where Hg(0) =F pCu) au The covariance of Y(t.) and Ut) is Cove) Zby)] = EL((ty) ay MZ(ty) ~My II 1 2 ELL Slt yt) Ay) (tye) = Ay) by) hy Cu) dt du) Lf ELK Ht) ay) K(tgeu) = 44) 46D hy(u) dt du LF Cy(tyatgeteu) ny Cr) hy(u) at du mee X al 1 2! where Cy(t) is the autocovariance function of X(t). Next, we note that the variance of X(t.) is 2 2 Of = LCE) may 97 FEF Sylora) yD Cw) de du and the variance of Z(t) is 2 2. oy = EL(Z(ty) — a, °7 2, 2 25 es fete) h(t) ho(u) dt du 26The correlation coefficient of ¥(t4) and Z(t) is covi¥(t4)2(t5)1 3 YT, % pe Since X(t) 18 a Gaussian process, it follows that Y(t,) and 2(t,) are 4 i a : ointly Caussi with a probability density function given by ' a sonny fewseten FyC4 4) ,2¢ty) Ya22) = K expl-O0y92)1 where anoy a 41 1 | ein 2 — »? - 206 2a-e)| ¥yn1. 227 z, 171 22, a 22 2,2 Y. 2. 7. 2 2 Ay 9z5) = 1 1 (b) The random variables ¥(t,) and Z(t,) are uncorrelated if and only if their covariance is zero, Since ¥(t) and Z(t) are jointly Gaussian processes, it follows that Y(t,) and 2p) are statistically independent if Cov[¥(t,)2(tp)] is zero. ‘Therefore, the necessary and sufficient condition for ¥(t,) and Z(t,) to be statistically independent is that is £ Sgt grtgeteu) h(a) hg(u) dt du = 0 for choices of t, and tp. .Problem 1.22 (a) The filter output is Y(t) = f h(x) X(tet) de a Tf Xtar) dt 0 + Then, the sample value of Y(t) at t=T equals ie 5 Yaqs xt) du ° ‘The mean of ¥ is therefore T EW] = EES x(u) au) oO T J E(XK(u)] du 0 aio The variance of Y is of = ELV?) - terry}? = Ry(0) = ce Sy(f) df £ S(t) ce? atH(f) = s W(t) exp(—Jj2nft) dt T J expl-jextt) dt 0 sI- pexptoseert 7 =jext 4 0 = jar (1 = exp(-j2fT)] = sine(fT) exp(-jafT) Therefore, og = S SyCt) sine@(eT) at (b) Since the filter input is Gaussian, it follows that Y is also Gaussian. probability density function of ¥ is mere of is detined shove. Problem 1.23 G@) Te power spectral density of the noise at the filter output is given by N = Moy jee 2 Se rorya Hence, they 2 sy(t) = 30 2xflmy T+(2qfL/RY zu Ls 14(2qf L/R)' Zz J ‘The autocorrelation function of the filter output is therefore x Bo too - Be expe Bet) ‘The mean of the filter output is equal to H(0) times the mean of the filter input. The process at the filter input has zero mean. The value H(0) of the filter’s transfer function H(f) is zero. It follows therefore that the filter output also has a zero mean. ‘The mean-square value of the filter output is equal to Ry(0). With zero mean, it follows therefore that the variance of the filter output is of = RCO) Since Ry(t) contains a delta function &(t) centered on t = 0, we find that, in theory, oy? is infinitely large. 30Problem 1.24 (a) The noise equivalent bandwidth is 4s wey ae (0)\? +» ipa ne + (4/8) at zs af ors (ety) fo * Bn sinGa/en) 2 * sino(i 72a) (b) When the filter order n approaches infinity, we have 0 Me SrRStT ERT Problem 1.25 ‘The process X(t) defined by xX®= Yo hit - wy, kee where h(t - %,) is a current pulse at time %, is stationary for the following simple reason. There is no distinguishing origin of time. 31Problem 1.26 (a) Let S,(f) denote the power spectral density of the noise at the first filter output. The dependence of S,(f) on frequency is illustrated below: s\(2) ¥/2 Let S,(f) denote the power spectral density of the noise at the mixer output. may write So(f) 1 TUS (fe) + 84 Ce which {s illustrated below: 8, (£) 32 Then, we‘The power spectral density of the noise n(t) at the second filter output is therefore defined by No ee 860 = 4 uc 0, otherwise ‘The autocorrelation function of the noise n(t) is RO = nce sine(2Bt) (b) The mean value of the noise at the system output is zero. Hence, the variance and mean-square value of this noise are the same. Now, the total area under S,(f is equal to (No/4)(2B) = NoB/2. The variance of the noise at the system output is therefore NyB/2. (c) The maximum rate at which n(t) can be sampled for the resulting samples to be uncorrelated is 2B samples per second, 33coblem 1.2’ (a) The autocorrelation function of the filter output is Ryle) = FF ley) Wlrg) Ryltoty+ty) dey Aty i} Since Ry(t) = (No/2) 6(t), we find that the impulse response h(t) of the filter must Satisfy the condition: Ry(r) = zh i h(t,) h(t.) S(tHt 445) dt, dty Ny * Bf Mart) h(ty) dtp (b) For the filter output to have a power spectral density equal to Sy(f), we have to choose the transfer function H(f) of the filter such that Ng 2 Sy(f) = so HCE) or (aCe)Problem 1.28 (@) Consider the part of the analyzer in Fig. 1.19 defining the in-phase component 742), reproduced here as Fig. 1 Narrowband noise no) filter > ni) 2cos(2nf.t) Figure | For the multiplier output, we have v(t) = 2n(t)cos(2mf,t) Applying Eq. (1.55) in the textbook, we therefore get SVP) = SF fo) +50 F + fod] Passing v(#) through an ideal low-pass filter of bandwidth B, defined as one-half the bandwidth of the narrowband noise n(f), we obtain Sy) = | Sy(f) for -BS f0 oe 20° ~ far) = | (oo otherwise To evaluate the variance 0°, we note that the power spectral density of n (t) or n_(t) is as follows zr Q 39Sy (fs, (£) Tr q Since the mean of n(t) is zero, we find that 2 os 2NB ‘Therefore, ROM = (op. Tue,mean value of the envelope is equal to VaNGB, and its variance is equal to «858 NOB. 0Problem 1.32 Autocorrelation of a Sinusoidal Wave Plus White Gaussian Noise In this computer experiment, we study the statistical characterization of a random process X(t) consisting of a sinusoidal wave component Acos(2nf.t + @) and a white Gaussian noise process W() of zero mean and power spectral density No/2. That is, we have X(1) = Acos2nf,t+@+ W(r) 0) where © is a uniformly distributed random variable over the interval(-at,1). Clearly, the two components of the process X(t) are independent. The autocorrelation function of X(0) is therefore the sum of the individual autocorrelation functions of the signal (sinusoidal wave) component and the noise component, as shown by 2 N Ry(t) = Acos(any,2) +528(t) This equation shows that for |t] > 0, the autocorrelation function Ry(t) has the same sinusoidal waveform as the signal component. We may generalize this result by stating that the presence of a periodic signal component corrupted by additive white noise can be detected by computing the autocorrelation function of the composite process X(t). ‘The purpose of the experiment described here is to perform this computation using two different methods: (a) ensemble averaging, and (b) time averaging. The signal of interest consists of a sinusoidal signal of frequency f, = 0.002 and phase 0 = - 1/2, truncated to a finite duration T = 1000; the amplitude A of the sinusoidal signal is set to V2 to give unit average power. A particular realization x(¢) of the random process X(2) consists of this sinusoidal signal and additive white Gaussian noise; the power spectral density of the noise for this realization is (No/2) = 1000. The original sinusoidal is barely recognizable in x(0), (a) For ensemble-average computation of the autocorrelation function, we may proceed as follows: + Compute the product x(t + t)x(¢) for some fixed time t and specified time shift t, where x(t) is a particular realization of the random process X(#). + Repeat the computation of the product x(t + t)x(1) for M independent realizations (i.c., sample functions) of the random process X(t). + Compute the average of these computations over M. + Repeat this sequence of computations for different values of t. The results of this computation are plotted in Fig. 1 for M = SO realizations. The picture portrayed here is in perfect agreement with theory defined by Eq. (2). The important point to note here is that the ensemble-averaging process yields a clean estimate of the true 4autocorrelation function Ry(t) of the random process X(t). Moreover, the presence of the sinusoidal signal is clearly visible in the plot of R(t) versus t. (b) For the time-average estimation of the autocorrelation function of the process X(t), we invoke ergodicity and use the formula Rea) = lim R(t, 7) @) where R,(t,7) is the time-averaged autocorrelation function "x(t t)xttyar (4) 1 ‘The x(1) in Eq. (4) is a particular realization of the process X(1), and 27 is the total observation interval. Define the time-windowed function ap(t) = [0 -TSIST 5 mh 0, otherwise ©) We may then rewrite Eq, (4) as nl RUG T) = spf arle4 t)ap(tpat © For a specified time shift t, we may compute R,(t,7) directly using Eq. (6). However, from a computational viewpoint, it is more efficient to use an indirect method based on Fourier transformation. First, we note From Eq. (6) that the time-averaged autocorrelation function R,(t,T) may be viewed as a scaled form of convolution in the t-domain as follows: RAT) Af ao x(t) ) where the star denotes convolution and x/(t) is simply the time-windowed function x7(#) with replaced by 7. Let X7(f) denote the Fourier transform x7(t); note that X;(/) is the same as the Fourier transform X(f,7). Since convolution in the t-domain is transformed into multiplication in the frequency domain, we have the Fourier-transform pair: RT) = Ef [xP ®The parameter |X7()|?/27 is recognized as the periodogram of the process X(t). Equation (8) is a mathematical description of the correlation theorem, which may be formally stated as follows: The time-averaged autocorrelation function of a sample function pertaining to a random process and its periodogram, based on that sample function, constitute a Fourier- transform pair. We are now ready to describe the indirect method for computing the time-averaged autocorrelation function Ry(t,7) + Compute the Fourier transform X7(f) of time-windowed function x7(1). + Compute the periodogram |X7(f/?/2T. + Compute the inverse Fourier transform of |X()/?/27- To perform these calculations on a digital computer, the customary procedure is to use the fast Fourier transform (FFT) algorithm, With x(t) uniformly sampled, the computational procedure described herein yields the desired values of | R,(t,7)_ for 1 = 0,4,2A,-,(N-1)A where A is the sampling period and N is the total number of samples used in the computation. Figure 2 presents the results obtained in the time-averaging approach of “estimating” the autocorrelation function Ry(t) using the indirect method for the same set of parameters as those used for the ensemble-averaged results of Fig. 1. The symbol Rx(t) is used to emphasize the fact that the computation described here results in an “estimate” of the autocorrelation function Ry(t). The results presented in Fig. 2 are for a signal-to-noise ratio of + 10dB, which is defined by A AT No/QT) No (9) On the basis of the results presented in Figures 1 and 2 we may make the following observations: + The ensemble-averaging and time-averaging approaches yield similar results for the autocorrelation function Ry(t), signifying the fact that the random process X(¢) described herein is indeed ergodic. + The indirect time-averaging approach, based on the FFT algorithm, provides an efficient method for the estimation of Ry(t) using a digital computer. + As the SNR is increased, the numerical accuracy of the estimation is improved, which is intuitively satisfying, 431 Problem 1.32 Matlab codes 4% Problem 1.32a CS: Haykin 4% Ensemble average autocorrelation %M. Sellathurai clear all Arsqrt (2); % signal s=cos(2epitf_cxt); Ynoise snr = 10°(SNR4b/10); ‘randn(1, Length(s) ))/sqrt(snr)/sqrt(2); gna plus noise autocorrelation [e_corrs]=en_corr(s,s, N} averaged autocorrelation -corrt_tte_corrt} wprints plot (~600:500-t,¢_corrt_t/M); xlabel(?(\tau)?) ylabel R_K(\tau) ") 44% Problem 1.32b 0S: Haykin 4 time-averaged estimation of autocorrelation %M. Seliethurai clear all % signal s=cos(24piet_cet) :Ynoise Ynoise snr = 10°(SNRdb/10); wn = (randn(1,2length(s)))/sqrt (snr)/eqrt(2); Ysignal plus noise sestwn; 4% time -averaged autocorrelation [e.corrf]=time_corr(s,N); Yprints plot (~500:500-1,e_corrt); xlabel(?(\tau)') ylabel (?R_X(\tau)’) 45function [eorr#} % ensemble average 4% used in problem 1.32, cS: Haykin %M. Sellathurai, 10 june 1999. ncorr(s, v, NY funtion to compute the autocorreation/ crose-correlatic nax_cross_cors teslength(u); tor shitted_u=(u(mti:tt) a(i:m)]; corr (m+1)=(sum(v.*shitved_u))/(N/2); if (abs (corr)>max_cross_corr) max_cross_corr=abs(corr); end end 46function [corrf]=time_corr(s,N) 4% tuntion to compute the autocorreation/ cross-correlation % time average 4% used in problem 1.32, CS: Haykin %M. Sellathurai, 10 june 1999, (5); Aa=tttshift ((abs(X).°2)/(N/2)); corrf=(tftshift (abs (ifft(X1)))); 47Answer to Problem 1.32 Figure 1: Ensemble averaging eas aa! ats mats — Mats ap ate ates — ats its ab Figure 2: Time averaging, 48Problem 1.33 Matlab codes % Problem 1.33 CS: Haykin % multipath channel %M. Sellethurai clear all 5% initializing counters 10000; % number of samples act; % line of sight component component for i=1:P Assqrt (randn(W,M).°2 + randn(N,M).-2); +cos(cos(rand(N,M)*2epi) + rand(N,W)*24pi); % inphase epmponent -#8in(cos(rand(w,M)*2epi) + rand(N,")*2*pi); quadrature phase component sum(xi?)); ssum(xq’))} x¥ rarsqrt((xita).~2+ xq.°2) ; % rayleigh, rician fading Gh X]=hist (ra,60); % print plot (Kt ,Nf/(sum(Xf .#Nt)/20)) xlabel(’v’) ylabel(*£_v(v)*) 49Answer to Problem 1.33 Figure 1 Rician distribution 50Problem 2, (a) Let the input voltage v; consist of a sinusoidal wave of frequency $f, (i.e., half the desired carrier frequency) and the message signal m(t): vy = A, cos(nf t)+m(t) Then, the output current 1, is : 3 tot ayy tag vy aj{A,cos(af,t)+n(t) ]easth costa ft)+m(t)]> ay(Ajcostnt t)mt)] + Jagh3 [eos(ief,t)+3eosnf,t)] + Bagt2 w(erttecostantgt)] + 3ahjcostnt tart) + age) Assume that m(t) occupies the frequency interval -W 20 a . 20 4 aa SLFrom this diagran we see that in order to extract a DSBSC wave, with carrier frequency f, from i,, we need a band-pass filter with mid-band frequency f, and bandwidth 2, which satisfy the requirement: f fw >See that is, f, > Gt Therefore, to use the given nonlinear device as a product modulator, we may use the following configuration: | device BEE ° 3. 42 7 43 Al m(t) cos (2ne_t) A cos (net) m(t) (b) To generate an AM wave with carrier frequency f, we require a sinusoidal component of frequency £, to be added to the DSBSC generated in the manner described above. To achieve this requirement, we may use the following configuration involving a pair of the nonlinear devices and a pair of identical band-pass filters. Nonlinear BPF device A cos (7) &) AM wave Nonlinear BPF device 52The resulting AM wave is therefore $a, AaCAgem(t)leos(2rf,t). Thus, the choice of the de level Ag at the input of the lower branch controls the percentage modulation of the AM wave. Problem 2.2 Consider the square-law characteristic: volt) = ayvy(t) + agvzit) a where a, and a, are constants. Let v(t) = Agcos(2nfgt) + m(t) @ ‘Therefore substitutingEq. (2) into (1), and expanding terms: volt) = ayAj1 + 222 mit) | cos(2nfyt) LOR : @) + aym(t) + apm %(t) + agA2cos*(2nf,t) ‘The first term in Eq. (3) is the desired AM signal with k, = 2ay/a,. The remaining three terms are unwanted terms that are removed by filtering. Let the modulating wave m(t) be limited to the band -W < f< W, as in Fig. 1(a). Then, from Eq. (3) we find that the amplitude spectrum |V,(f)| is as shown in Fig. 1(b). It follows therefore that the unwanted terms may be removed from vz(t) by designing the tuned filter at the modulator output of Fig. P2.2 to have a mid-band frequency f, and bandwidth 2W, which satisfy the requirement that f, > 3W. Wvs00 won KA A A A WO WF 2h “haw WOW Wh mh he (@) (b) aw Figure 1 33Problem 2.3 The generation of an AM wave may be accomplished using various devices; here we describe one such device called a switching modulator. Details of this modulator are shown in Fig. P2.3a, where it is assumed that the carrier wave c(t) applied to the diode is large in amplitude, so that it swings right across the characteristic curve of the diode. We assume that the diode acts as an ideal switch, that is, it presents zero impedance when it is forward-biased [corresponding to c(t) > 0}. We may thus approximate the transfer characteristic of the diode-load resistor combination by a piecewise-linear characteristic, as shown in Fig. P2.3b. Accordingly, for an input voltage v(t) consisting of the sum of the carrier and the message signal: v,(0) = A,cos(2n ft) +m(t) a) where |m(2)| << A,, the resulting load voltage v9(0) is, v(t), o(t)>0 0, e(t)<0 vl) = @ ‘That is, the load voltage v(t) varies periodically between the values v;(1) and zero at a rate equal to the carrier frequency /,. In this way, by assuming a modulating wave that is weak compared with the carrier wave, we have effectively replaced the nonlinear behavior of the diode by an approximately equivalent piecewise-linear time-varying operation, We may express Eq. (2) mathematically as v(t) = A, cos(2nf1) + m(Ngr,(0) (3) where g7,(#) is a periodic pulse train of duty cycle equal to one-half, and period Ty = Mf, as in Fig. I. Representing this g,(#) by its Fourier series, we have 1,2 (—1 bri = 542 cos[2mf,t(2n-1)] 4) ‘Therefore, substituting Eq, (4) in (3), we find that the load voltage v2(7) consists of the sum of two components: 1. The component 1m] 0s (27, f,.t) 34which is the desired AM wave with amplitude sensitivity k, = 4m4,. The switching modulator is therefore made more sensitive by reducing the carrier amplitude A,; however, it must be ‘maintained large enough to make the diode act like an ideal switch. 2. Unwanted components, the spectrum of which contains delta functions at 0, +2/. +4fe. and so on, and which occupy frequency intervals of width 2W centered at 0, 43f,.. 45f,, and so on, where W is the message bandwidth. aryl i To te : why wt Fig, |: Periodic pulse train The unwanted terms are removed from the load voltage v2(*) by means of a band-pass filter with mid-band frequency f, and bandwidth 2W, provided that f, > 2W. This latter condition ensures that the frequency separations between the desired AM wave the unwanted components are large enough for the band-pass filter to suppress the unwanted components.(a) The envelope detector output is v(t) = All1+ woos (anf t) | which is illustrated below for the case when p= We see that v(t) is periodic with a period equal to f,, and an even function of t, and so we may express v(t) in the form: v(t) = ag +2 Ea, cos(2nn t,t) nl Vet, art vit)dt mo where 13 q 1fy = At s (142 cos(arf tyidt + 2hf. s (-1-2008(arf,t)1dt Gh, ‘ f) orm ae m ™ © sin(E am Vet, = ats v(t) cos(2nm ft) dt ny im 1438, # Abeta J (142c0s(2nf,,t) Jeos(2nr ft) dt 561peE, n + Att (-1-2cos(2rf,.t) Jeos(2nx ft) dt 1/3Ey 2 29 ¢2 otnc22%) = atntond] + eS (2 stnt2EKowty] = stats (mt 91) = FB 2 sin@% ~ sine + ate En te {2 int 2 1)] sin(n (n=1)]) (2) + Gete 2 sinl2h(n-1)) ~ sinta (ne For nz0, Eq. (2) reduces to that shown in Eq. (1). (b) For n=1, Eq. (2) yields Gs} an * D For n=2, it yields 1 an Therefore, the ratio of second-harmonic amplitude to fundamental amplitude in v(t) is 2 38 = 0.452 By nSProblem 2.5 (a) The demodulation of an AM wave can be accomplished using various devices; here, we describe a simple and yet highly effective device known as the envelope detector. Some version of this demodulator is used in almost all commercial AM radio receivers. For it to function properly, however, the AM wave has to be narrow-band, which requires that the carrier frequency be large compared to the message bandwidth, Moreover, the percentage ‘modulation must be less than 100 percent. An envelope detector of the series type is shown in Fig. P2.5, which consists of a diode and a resistor-capacitor (RC) filter. The operation of this envelope detector is as follows. On a positive half-cycle of the input signal, the diode is forward-biased and the capacitor C charges up rapidly to the peak value of the input signal. When the input signal falls below this value, the diode becomes reverse-biased and the capacitor C discharges slowly through the load resistor R). The discharging process continues until the next positive half-cycle. When the input signal becomes greater than the voltage across the capacitor, the diode conducts again and the process is repeated. We assume that the diode is ideal, presenting resistance r; to current flow in the forward-biased region and infinite resistance in the reverse-biased region. We further assume that the AM wave applied to the envelope detector is supplied by a voltage source of internal impedance R,. The charging time constant (rr + R,) C must be short compared with the carrier period I/f, that is (ry + RICA a) so that the capacitor C charges rapidly and thereby follows the applied voltage up to the positive peak when the diode is conducting. (b) The discharging time constant R;C must be long enough to ensure that the capacitor discharges slowly through the load resistor R; between positive peaks of the carrier wave, but not so long that the capacitor voltage will not discharge at the maximum rate of change of the modulating wave, that is 1 1 FeRCeg Q) where W is the message bandwidth, The result is that the capacitor voltage or detector output, is nearly the same as the envelope of the AM wave. 58vy(t) = Ag1 + kgmi(t)]cos(2nf,t) (2) Then the output of the square-law device is va(t) = ayvy(t) + agv f(t) = aA{1 + kgm(t)]eos(2nf,t) + peokzt + 2kymi(t) + k2m%t)] [1 + cos(4nf.t)] (b) ‘The desired signal, namely ajA,k,m(t), is due to the av} (t) - hence, the name "square-law detection”. This component can be extracted by means of a low-pass filter. This is not the only contribution within the baseband spectrum, because the term 1/2 a)A,"k,”m%(t) will give rise to a plurality of similar frequency components. The ratio of wanted signal to distortion is 2/k,m(t). To sake this ratio large, the percentage modulation, that i, Tiym(F should be kept small compared with unity.Problem Tne squarer output is arg 2 cos? vy(t) = AS Cokgmt)]? cos? (are, t) 1% Tsk m(t) +n? Ct) IE eeosCHet st) 1 The amplitude spectrum of v,(t) is therefore as follows fi the interval Wife: |! + assuming that m(t) is limited to vce] Since f, > 21, we find that 2f,-2W > 2. ‘Therefore, by choosing the cutoff frequency of the low-pass filter greater than 2W, but less than 2f,-2H, we obtain the output vat) Cakegm(t) 7? le’ Hence, the square-rooter output is a c v(t) = 7 Ciakegnte)) de whieh, excent for the de component — , 1s proportional to the message signal a(t). 2 Problem 2.8 (a) For f, = 1.25 klz, the spectra of the message signal m(t), the product modulator output s(t), and the coherent detector output v(t) are as follows, respectively: ©vee) KK (ate) = 0 Fy (bv) For the case when f, = 0.75, the respective spectra are as follows: M(f) - £ (kHz) s(t) = 5° mt) cos(2nart) Thus the output signal is the message signal modulated by a sinusoid of frequency Af. (b) If m(t) = cos(2nr.t), A then sa(t) = 5° cos(2mf it) cos(2natt) 4 Ah obl 2, (a) y(t) = s(t = A? cos®(2at yar (t) 2 = 5f [4cos( aft) Im? (t) 2 ce Therefore, the spectrum of the multiplier output is 2 eS) az ow o £ MOONEE adda + ge LF MOUMCE2E adda +S MOOMCE42f,-2)da) Y(f) = “o’ where M(f) = Fmt) J. (b) at 8 fy We haveae iw =e Yer) #t FMOOM(2E,~ Maa a © © + UL NOOM- da eS MOONCHE apd A] Le a e Since M(=A) = M#(A), we may write a2 c YQE) = BS MONM(2F,-A)ad. LINGO aa ef MOONCE,~ aan) a 2 et F + With m(t) limited to -W < f < Wand f, > W, we find that the first and third integrals reduce to zero, and so we may simplify Eq. (1) as follows 2 Ao oe 2 yer sais imcofPaa where Eis the signal energy (by Rayleigh's energy theoren). Similarly, we find that > Y(-2f,) = GE The band-pass filter output, in the frequency domain, is therefore defined by 2 A Vif) = GPE afl e(e2t.) + 6(fe2r,)) Hence, > 7 v(t) le E af cos(4xt,t) oyProblem 2.12 The multiplexed signal is s(t) = A my(t) cos(2mf.t) + Ay my(t) sin(2nft) Therefore, > 4, +58 ee 1 % SCH) = 3PM Cet dM jCC40,)] + 35 Mol Pf IM t4f,)) where M,(f) = Flm(t)] and Mo(f) = Flm(t)], The spectrum of the received signal is therefore RCE) H(f)S(f) 1 (2) MCE day (Pef Ie FMS = 5 5 M(tet)I To recover m,(t), we multiply r(t), the inverse Fourier transform of R(f), by cos(2mft) and then pass the resulting output through a low-pass filter, producing a signal with fhe following spectrum FUr(tycos(2aft)] = F (R(faf,)4R(f40,)] A ce NCEE OM (20) + MCE) + FMolter,) - $Me) A + TE HCPeE DMC) + My (te2e.) + F Ma(t) 5 Mylt42r,)1 a The condition H(f,+f) = H#(f,-f) is equivalent to H(f+f,)=H(f-£,)s this follows from the fact that for a real-valued impulse response h(t), we have H(-f)-H#(f), Hence, substituting this condition in Eq. (1), we get A c Fr(t)eos(2nf,t)] = 5° H(f-£,)M,(£) ry ce 1 1 + Pe HLM (P20) + F Np (fF, )4M (F420, - 5 Mp(f42f,)1 The low-pass filter output, therefore, has a spectrum equal to (A,/2) H(f-f,)M,(f). Similarly, to recover m(t), we multiply r(t) by sin(2mf.t), and then pass the resulting signal through a low-pass filter. In this case, we get an output with a spectrum equal to (A,/2) H(f-£,)M)(f). a66 Problem 2.13, When the local carriers have a phase error 4, ve may write cos(2af,t + 4) = cos(2nf.ticos ¢- sin(2xf,t) sin ¢ In this case, we find that by multiplying the received signal r(t) by cos(2af.t+4)s and passing the resulting output through a low-pass filter, the corresponding low-pass filter output in the receiver has a spectrim equal to (A./2) H(f-£,) [cos M,(f) - sing NQ(f)]. This indicates that there is cross-talk at the demodulator outputs. 66Problem 2.14 ‘The transmitted signal is given by 3() = Apm,(t)cos (2 ft) + Apma(1)sin(2n ft) = A(LVo + mj(t) + m,(t)]cos(2nf,t) + A[m,(t) ~m, (1) ]sin (2m f,2) (a) The envelope detection of s(0) yields Yilt) = Agal(Vo +m (4) +m,(1))° + (omit) = m,(1)° mt) — m,(t) \2 = A(Vg + m,(t)+m,(0) 1 aa) To minimize the distortion in the envelope detector output due to the quadrature component, we choose the DC offset Vo to be large. We may then approximate yy(1) as y(t) = A(Vo + m1) +m,(0)) which, except for the DC component A, Vo, is proportional to the sum m(t) + m,(t). (b) For coherent detection at the receiver, we need a replica of the carrier A,cos(2n/,t). This requirement can be satisfied by passing the received signal s(¢) through a narrow-band filter of mid-band frequency f.. However, to extract the difference m,(t) - m,(2), we need sin(2rf,f), which is obtained by passing the narrow-band filter output through a 90°-phase shifter. Then, multiplying (0) by sin(2ryt) andt low-pass filtering, we obtain a signal proportional to mt) - m,(0) (©) To recover the original loudspeaker signals m1) and m,(t), we proceed as follows: + Equalize the outputs of the envelope detector and coherent detector. + Pass the equalized outputs through an audio demixer to produce mt) and m,(0), o7Problem 2.15 (a) s(t) = A.C + kym(1))cos (2m ft) (1, te id 5 |cos(2m ft) \ ler) To ensure 50 percent modulation, ka = I, in which case we get s(t) = alt+s 1 )eos(2nf.) +P (b) s(t) = Agn(t)cos(2nf,.t) A, S cos(2n ft) Ler Ac, (c) s(t) = q Lam(z)cos (2m ft) — Hi(r)sin(27f.t)) 7 cos(2mf,.t)- 1 sin(2nf,1) Sram aerate sy = Sf 1 cos(2me ft) + sin y_0)] thier? ier As an aid to the sketching of the modulated signals in (c) and (d), the envelope of either SSB wave is _if?ser 1 fa 00 aaa Wie? Plots of the modulated signals in (a) to (d) are presented in Fig. 1 on the next page. 68= 10a) 6 4 ~2 0 2 4 6 8 10 69Problem 2.16 Consider first the modulated signal s(t) = fim(t)cos(2m ft) ~ sin(t) sin (2m f,.1) a Let S(f) = F[s()], Mi) = Flm(2)], and M(f) = f[sa(t)] where r(r) is the Hilbert transform of the message signal m(j). Then applying the Fourier transform to Eq. (1), we obtain SU) = MEF) # MU + fo)I= LMS F-) MUS + £2] @ From the definition of the Hilbert transform, we have Mf) = ~jsgn(f)M(f) where sgn({) is the signum function. Equivalently, we may write HF f= sen f-FMU-F.) HHS) = sen + SIMS +S (i) From the definition of the signum function, we note the following for f> 0 = and f> fi: sen(f- fo) sen(f+f,) = +1 Correspondingly, Eq. (2) reduces to SCA) = HMC.) 4 MCF ef 1+ IMC -F)-MF +f] 1 = 5MUE-f) In words, we may thus state that, except for a scaling factor, the spectrum of the modulated signal (¢) defined in Eq. (1) is the same as that of the DSB-SC modulated signal for f > f,. Gi) For f> 0 and f< f,, we have 70sen(f-f,) = -l sen(f+f,) = +1 Correspondingly, Eq. (2) reduces to S(f) GIMS-L¢ MF + f+ LMU fe) - MEF] =0 In words, we may now state that for / 0. This result also holds for f< 0, the proof for which is left as an exercise for the reader. Following a procedure similar to that described above, we may show that the modulated signal s(t) = (eos (2n ft) + (0) sin(2nf.1) @) represents a single sideband modulated signal containing only the lower sideband. nNProblem 2.17 ‘An error Af in the frequency of the local oscillator in the demodulation of an SSB signal, measured with respect to the carrier frequency f,, gives rise to distortion in the demodulated signal. Let the local oscillator output be denoted by A, cos(2n(f, + Aft). The resulting demodulated signal is given by (for the case when the upper sideband only is transmitted) volt) = ; A,A, [m(t)cos(2nAft) + m(t)sin(2nAft)] ‘This demodulated signal represents an SSB wave corresponding to a carrier frequency Af. ‘The effect of frequency error Af in the local oscillator may be interpreted as follows: @ (b) If the SSB wave a(t) contains the upper sideband and the frequency error Af is positive, or equivalently if s(t) contains the lower sideband and Af is negative, then the frequency components of the demodulated signal v,(t) are shifted inward by the amount Af compared with the baseband signal m(t), as illustrated in Fig. 16). If the incoming SSB wave s(t) contains the lower sideband and the frequency error Af is Positive, or equivalently if s(t) contains the upper sideband and Af is negative, then the frequency components of the demodulated signal v,(t) are shifted outward by the amount Af, compared with the baseband signal m(t). This is illustrated in Fig. 1¢ for the case of a baseband signal (e.g., voice signal) with an energy gap occupying the interval -f, _# The output of the lower first product modulator has the spectrum: J Gat E-§) EMEP A i The output of the upper low pass filter has the spectrum: aM, G+f) 14The output of the lower low pass filter has the spectrum: ~ => M_(f. ay MF) Sth f OL ~4.£ mo “ ! 0 “3 M$) Tne output of the upper second product modulator has the spectrum: iM, Ge fed) 4 1 io, (44-2) fl . 4 Fog tM G-£-4) ME Sad) ‘ @ et i aa oo <_—~I___+ oF The output of the lower second product modulator. has the spectrum: 2 4f-£) Mg bef) 4 Ma fh au, §-4) Adding the two second product modulator outputs, their upper sidebands add constructively while their lower sidebands cancel each other. (c) To modify the modulator to transmit only the lower sideband, a single sign change is required in one of the channels. For example, the lower first product modulator could multiply the message signal by -sin(2nf.t). Then, the upper sideband would be cancelled and the lower one transmitted. 15Problem 2.19 m(t) Product El sa High-pass 2} Product: 3 (or /OW~Pass i) modulator filter [--=|modulator filter cos (2nf t) cos [2m (f+, )t] (a) The first product modulator output is v(t) = m(t) cos(2xf,t) The second product modulator output is v3(t) = volt) cosl2mt +f.) t] The amplitude spectra of mt), v4(t), vp(t), v3(t) and s(t) are illustrated on the next page: . We may express the voice signal mt) as mt) = Fim (t) + m¢t)1 where m,(t) is the pre-envelope of m(t), and m_(t) * m,(t) is its complex conjugate. The Fourier transforms of m,(t) and m(t) are defined by (see Appendix 2) ae), £>0 M(t) = oO, £0 Mf) = a ar), feo Comparing the spectrum of s(t) with that of m(t), we see that s(t) may be expressed in terms of m,(t) and m(t) as follows: 1 1 s(t) = gm (toexp(-Jenf,t)+ g m (texpl jex,t) Bo f= of Lint )+ 580 t)Jexp(-Jenf t+ Bim t)-5ACt dLexpl Jen) n(teos(2aft+ yf A(t) sin(2af,t) (>) With s(t) as input, the first product modulator output is v(t) = s(t) cos(2af,t) 16Ince) | Ince | 1 81 C4Problem 2.20 (a) Consider the system described in Fig. 1a, where u(t) denotes the product modulator output, as shown by u(t) = Agm(t)cos(2n ft) ane product | (| Bandpass Modulated a modulator titer signa! ae HO 1) A005 (28) ) Modulates psu signal Product |_| Lowpass Demeaistes 0) modulator ‘iter “A Aye0s (nf) o Figure 1: (a) Filtering scheme for processing sidebands. (b) Coherent detector for recovering the message signal. Let H(f) denote the transfer function of the filter following the product modulator. The spectrum of the modulated signal s(7) produced by passing u(t) through the filter is given by S(f) = U(NH(f) A = ZIMF-f+MUF + SHY) wm where M(f) is the Fourier transform of the message signal m(?). The problem we wish to address is to determine the particular H(f) required to produce a modulated signal s(¢) with desired spectral characteristics, such that the original message signal m(t) may be recovered from s(t) by coherent detection, The first step in the coherent detection process involves multiplying the modulated signal s(t) by a locally generated sinusoidal wave A’,cos(2m/,t), which is synchronous with the carrier wave A.cos(2n/,f), in both frequency and phase as in Fig. 1b. We may thus write 78vn) cos (2mf,.t)s(t) ‘Transforming this relation into the frequency domain gives the Fourier transform of v(1) as av VP) = SASF. + SF + F1 @ Therefore, substitution of Eq. (1) in (2) yields AA’, VP) = SMA -F.) +H + Fd IM(f-2F )H(f -f.)+M(f + 2f HS + £21 @) (b) The high-frequency components of v(t) represented by the second term in Eq, (3) are removed by the low-pass filter in Fig. 1b to produce an output v,(z), the spectrum of which is given by the remaining components: AA Vif) = AM NLAS -F.) + HF +f) @) For a distortionless reproduction of the original baseband signal m(#) at the coherent detector output, we require V,(/) to be a scaled version of M(/). This means, therefore, that the transfer function H(f) must satisfy the condition AS-f)+H(F +f.) = 2H(f) (6) where H(,), the value of H(f) at f= /,, is a constant. When the message (baseband) spectrum M(f) is zero outside the frequency range -W < f'< W, we need only satisfy Eq. (5) for values of f in this interval. Also, to simplify the exposition, we set H(f,) = 1/2. We thus require that H1(f) satisfies the condition: Hf-f+H(f +f) -wsf> 1, we may express S(f) as follows : 3 O(tof,) + F sinett(t-r,-0)1 = Esinettft.y), > 0 sf) = A 2 O(f0,) + F sineltitetyeae)) = Z sinelt(ff,)1, Problem 2.27 For SSB modulation, the modulated wave is A s(t) = [m(t) cos(2nt.t) + f(t) sin(2nrt)], the minus sign applying when transmitting the upper sideband and the plus sign applying when transmitting the lower one. Regardless of the sign, the envelope is A a(t) = 527 me(ty + w2cey (a) For upper sideband transmission, the angle, O(t) = anfyt + tan” The instantaneous frequency is, 1 ft) £,(t) = 7 SD sf e MUD ANE) — ACE) atte) © ar(m@(t) + Ce) where ' denotes time derivative. (b) For lower sideband transmission, we have fit). mt) * d Q(t) = ant + tan’ and (ty et.» BUED at{e) = ate) f(t) 2x (m@(t) + H(t) 88Problem 2.28 (a) The envelope of the FM wave s(t) is Syn? sin® 1p ¥ te sin“(ant,t) The maximum value of the envelope is a(t) and its minimum value is nin ‘Therefore, a st = / 148 ‘min This ratio is shown plotted below for 0 < 8 < 0,3: Qmae nin (>) Expressing s(t) in terms of its frequency components: s(t) = A, cos(2nf.t) +36 A, coslan(fi+f dt] - 38 A, coslen(f-f,)t] 89The mean power of s(t) is therefore 2 42,242 ag aS | BAS 2*o +3 2 AS cs ‘c ea+hy The mean power of the unmodulated carrier is which is shown plotted below for 0 <8 < 0.3: 90(e) The angle 6,(t), expressed in terms of the in-phase component, s,(t), and the quadrature component, s(t), ist 05 (t) = anft ji = ft + (esin(2rf,t)1 Since tan""(x) = x= x9/3 4.0.5 ag(t) = anf + pain(ant.t) — & stn3anet) 00) = Sri ot + Bsintentt) ~ 3 7 The harmonic distortion is the power ratio of the third and first harmonics: For 8 = 0.3, Dy = 0.09% n Problem 2.29 (a) The phase-modulated wave is s(t) = A, coslarf.t + kA, cos(2nfgt)] A, coslanf.t + 8, cos(2nf,t)] ¢ COS(2nf gt) coslB, cos(2af,t)] - A, sin(2nfgt) sins, cos(2nt,t)) SD)If 8, $0.5, then costa, cos(anf,t)] = 1 Sin{B, cos(2nf,t)] = 8, cos(2nt,t) Hence, we may rewrite Eq. (1) as s(t) = Ay cos(2nf .t) - 8 Ay sin(2nf .t) cos(2nft) 1 = AS cos(2nft) “2 8D Ay sinl2n(f,+f,)t] 1 ~ 38, A, sinter (t-f,)t] (2) The spectrum of s(t) is therefore ar S(f) = FALLS (ff,) + 6 (#4801 ~ Ty Bp AglOUtty-t,) - Siar sf,)] | 1 ~ Tp Bp Mel (Eat yef,) = S(tef -£,)1 (b) The phasor diagram for s(t) is deduced from Eq. (2) to be as follows: Lower side-frequency carrier anf t m <—— Upper side-frequency The corresponding phasor diagram for the narrow-band FM wave is as follows: 92carrier a, Lower side~frequency Comparing these two phasor diagrams, we see that, except.for a phase difference, the narrow-band PM and FM waves are of exactly the sane form. Problem 2.30 The phase-modulated wave is s(t) = AY coslanf t + a cos(2nf t)) The complex envelope of s(t) is Bt) = Ay expl 8, cos(2nf,t)] Expressing S(t) in the form of a complex Fourier series, we have Blt) =e, exp( j2mnt,t) news where Vey. che ty! 3(t) exp(-j2mnf,t) at nM aye, . vet, FAL ag APL Sp COBC2E GE) — J2mf,t] dt a” ” Let aft = 1/2 -¢. Then, we may rewrite Eq. (1) as A =f exp- A -1/2 £ expt J, sin(g) + ing] do a /2 93The integrand is periodic with respect to ¢ with a period of 2. Hence, we may rewrite this expression as 2 exp(- IE) 5 expt js, sin(g) + jmp] dp Boe p However, from the definition of the Bessel function of the first kind of order n, we have « jo ek £ expld x sing - nie) a6 Therefore, ine ig exp(- AB) JB) We may thus express the PM wave s(t) as s(t) = Re[S(t) expt janf.t)] A, Rel T° J_n(8p) exp(- 9B expt jeentyt) exp( j2nfgt)) ns» lan (Bp) cosl2n(foenf Dt = 5E The band-pass filter only passes the carrier, the first upper side-frequency, and the first lower side-frequency, so that the resulting output is Bolt) = A, Jo(B,) cos(2nt.t) +A, J_1(8,) coslan(f +f, It = 2 1 +A, Jy(B) coslen(t Fyre + 5 = AQ IolB,) cos(2nf gt) +A, J_{(8,) sink2n (tf, )t] A, J4(8,) sinlan(f ft] But 146,) = J4@,) Therefore, y(t) eA JB) cos(2nf t) -Ay J4@p) {sinlan(f +f, )t) + sinl2n(f,-f,)t]) eA, Jo(By) cos(2nft) = 2 A, 946) cos(nf t) sin(arf.t) The envelope of s,(t) equals o4a(t) / 2, +4 (8) 408.) cos“(2nf t) ec” Yo! The phase of s(t) is 2 34(8,) Jol) 1 ot) = tan"! £ cos(2xft)) The instantaneous frequency of s,(t) is ef. + Le dott) nen ete 4 a 2 J (Bp) 5,6,) sin(2nf t) SIC) + 4958, cos“C2nr gt) Problem 2.31 (a) From Table Au.1, we find (by interpolation) that Jp(B) is zero for B= 2.44, 8 = 5.52, B = 8.65, B = 11,8, and so on. (b) The modulation index is oat, etn tn Therefore, Bf, n keg n Since Jg(8) = 0 for the first time when B = 2.44, we deduce that 2.uu x 103 2 = 1.22 x 103 hertz/volt Next, we note that Jo(B8) = 0 for the second time when 8 = 5.52. Hence, the corresponding value of A, for which the carrier component is reduced to zero is 951.22 x 103 5.52 x 103 = 4,52 volts Problem 2.32 For 8 Ig) 340) JQ) Therefore, s9ct) = 1, we have = 0.765 = 0.4m = 0.115 the band-pass filter output is (assuming a carrier amplitude of 1 volt) = 0.765 cos(2nf,t) + 0.44 {coslan(f.+f,)t] - cosl2n(f,-f,)t]) + 0,115 {coslén(f,+2f,)t] + coslan(f,-2f,)t]) » and the amplitude spectrum (for positive frequencies) is [So a 96Problem 2.33 (2) The frequency deviation is Af = kp Ay = 25 x 103 x 2025 x 10° Hz ‘The corresponding value of the modulation index is af | 5 x 10% 10 B= 5 ‘The transmission bandwidth of the FM wave, using Carson's rule, is therefore By = 2f,(148) = 2x100 (145) = 1200 kHz = 1,2 MHz. (>) Using the universal curve of Fig. 336 we find that for a ra Therefore, By = 3x500 = 1500 kHz = 1.5 MHz (c) If the amplitude of the modulating wave is doubled, we find that Af = 1 Miz and p = 10 Thus, using Carson's rule we obtain By = 2x100 (1410) = 2200 xHz = 2,2 MHz Using the universal curve of Fig. 3:36, we get By apt 25 and By = 2.75 Miz. (4) If £, is doubled, g 2.5. Then, using Carson's rule, By = 1.4 Miz. universal curve, By/af and By = 4af = 2 Miz. 7 Using theProblem 2.34 (a) The angle of the PM wave is o(t) Aafgt + ky mt) arf gt + ky A,cos(2nf,t) ant.t + 8, cos(2rf,t) where B, =k, Aye The instantaneous frequency of the PM wave is therefore (0) ale ar at f. - 8, fy sin(2nf,t) We see that the maximum frequency deviation in a PM wave varies linearly with the modulation frequency f,. Using Carson's rule, we find that the transmission bandwidth of the PM wave is approximately (for the case when 6, >> 1) By = 2 + 8p fq) = 2y(1 +8,) = 2, 8 Tais shows that By varies linearly with f,. (>) In an FM wave, the transmission bandwidth By is approximately equal to 2af, if the modulation index 8 >> 1. Therefore, for an FH wave, By is effectively independent of the modulation frequency f,. Problem 2.35 ‘The filter input is v(t) = g(t) s(t) g(t) cos(2nt,t = set?) The complex envelope of v,(t) “is F(t) = g(t) explagekt®) ‘The impulse response h(t) of the filter is defined in terms of the complex impulse response h(t) as follows h(t) = Reh(t) exp( jarft)] with h(t) = costanf st + wkt™), we have Ree) = expt jnit) aof the The complex envelope¥filter output is therefore (see Appendix 2) Volt) =gRceoee V(t) a(t) exp(-jnkr®) expl jak(ter)]°dr via f 1 2a B exp jake) f g(t) exp(-Jenktr) dr 1 2. B exp(inkt) G¢Kt) Hence, Wigton =F tect This shows that the envelope of the filter output is, except for the scale factor of 1/2, equal to the magnitude of the Fourier transform of the input signal g(t), with kt playing the role of frequency f. Problem 2.36 The overall frequency multiplication ratio is n= 2x3 = 6 Assume that the instantaneous frequency of the FM wave at the input of the first frequency multiplier is fyy(t) = fy + af cos(2sf gt) The instantaneous frequency of the resulting FM wave at the output of the second frequency multiplier is therefore fy p(t) = nf, + ndf cos(2nft) Thus, the frequency deviation of this FM wave is equal to nAf = 6x10 = 60 kHz and its modulation index is equal to nat _ 60 nat . 0 2 a2 t, 3 The frequency separation of the aijacent side-frequencies of this FM wave is unchanged at f, = 5 kHz. ™ 99Problem 2.37, (a) Figure 1 shows the simplified block diagram of a typical FM transmitter (based on the indirect method) used to transmit audio signals containing frequencies in the range 100 Hz to 15 KHz. The narrow-band phase modulator is supplied with a carrier signal of frequency f; = 0.2 MHz by a crystal-controlled oscillator. The desired FM signal at the transmitter output is to have a carrier frequency f= 100 MHz and a minimum frequency deviation Af'= 75 kHz. In order to limit the harmonic distortion produced by the narrow-band phase modulator, we restrict the modulation index Bj of this modulator to a maximum value of 0.3 radians. Consider then the value i; = 0.2 radians, which certainly satisfies this requirement. The lowest modulation frequencies of 100 Hz produce a frequency deviation of Af, = 20 Hz at the narrow-band phase modulator output, whereas the highest modulation frequencies of 15 KHz produce a frequency deviation of Af, = 3 kHz. The lowest modulation frequencies are therefore of immediate concern, as they produce a much lower frequency deviation than the highest modulation frequencies. The requirement is therefore to ensure that the frequency deviation produced by the lowest modulation frequencies of 100 Hz is raised to 75 KHz. aera ro aowbona] | Freaerey Fraqueney | soa Hd ingatoe Lol ne Lol fattet Lol ier Loy moter bo meta [| at ey cyst cry Figure 1 To produce a frequency deviation of Af = 75 kHz at the FM transmitter output, the use of frequency multiplication is obviously required. Specifically, with Af, = 20 Hz and Af = 75 kHz, ‘we require a total frequency multiplication ratio of 3750, However, using a straight frequency multiplication equal to this value would produce a much higher carrier frequency at the transmitter output than the desired value of 100 MHz. To generate an FM signal having both the desired frequency deviation and carrier frequency, we therefore need to use a two-stage frequency ‘multiplier with an intermediate stage of frequency translation as illustrated in Fig. 1. Let m and ny denote the respective frequency multiplication ratios, so that Af _ 73000 af 20 nyny = 3750 a 100‘The carrier frequency nf; at the first frequency multiplier output is translated downward to (J - nyfi) by mixing it with a sinusoidal wave of frequency f, = 95 MHz, which is supplied by a second crystal-controlled oscillator. However, the carrier frequency at the input of the second frequency multiplier is required to equal f./ny. Equating these two frequencies, we thus get Soom fy = E ny Hence, with fy = 0.1 MHz, fy = 9.5 MHz, and f, = 100 MHz, we have 9.5-0.1n, = 1 2) nm Solving Eqs. (1) and (2) for n, and 9, we obtain ny =75 ny =50 (b) Using these frequency multiplication ratios, we get the set of values indicated in the table below: Table -Values of Carrier Frequency and Frequency Deviation at the Various Points in the Wide-band Frequency Modulator of Fig. 1 At the First ‘At the Second At the Phase Frequency Frequency Modulator Multiplier | Atthe Mixer | Multiplier Output Output Output Output Carrier 0.1 MHz 7.5 MHz 2.0 MHz, 100 MHz frequency Frequency 20 He 1.5 kHz 1.5 kz 15 kHz deviation 101Jem 2.3: (a) Let L denote the inductive component, ¢ the capacitive component, and Cy the capacitance of each varactor diode due to the bias voltage V, acting alone. Then, we have 7 -12 Cy = 100 vo'/? pF and the corresponding frequency of oscillation is awl (Cat 72) Therefore, 25/200 x 10 Solving for Vp, we get Vp = 3-52 volts (b) The frequency multiplication ratio is 64. Therefore, the modulation index of the FM Wave at the frequency multiplier input is 0,078 Tris indicates that the FM wave produced by the combination of L, C and the varactor diodes is a narrow-band one, which in turn means that the amplitude A, of the modulating wave is snall compared to V,. We may thus express the instantaneous fFequency of this FM wave as follows: 12 1224 172 ft) = & 200 x 106 (100 x 1071? + 50 x 107"? 13.52 + AL silent ty 7 An _ 10 1/2 ,-1/2 = (1 + 0,266 (1 +3755 sin(ant tyy7/*) even 7 A ~ 10 m See cee arson ame “1/2 = 10° (1 ~ 0,03 A, sin¢amrt) 37/2 = 10° (1 + 0,015 A, sin(ent,t)) 102js “ith @ modulation index of 0.078, the corresponding value of the frequency deviation 8 ate ee, 0.078 x 104 He Therefore, 0.015 A, x 10° = 0,078 x 10" where A, is in volts. Solving for 4,, we get Ay = 52 x 1079 volts. in Problem 2.39 ‘The transfer function of the RC filter is non) = ae If 2xfCR << 1 for all frequencies of interest, then we may approximate H(f) as H(f) = j2nfCR However, multiplication by j2xf in the frequency domain is equivalent to differentiation in the time domain. Therefore, denoting the RC filter output as v,(t), we may write = or S86) v(t) = oR SE A t = OR Ge (A, cosl2net + 2k, a mt) dt}} t WOR AL2nf, + 2tkpm(t)] sinl2nt.t + 2tky is mt) dt] The corresponding envelope detector output is k f volt) = 2xfcr A,|1 + Fae] Since k,{m(t)] < , for all t, then Ke vat) = 2nfcR ALL + pe mt} Which shows that, except for a de bias, the output is proportional to the modulating signal m(t). 103Problem 2.40 The envelope detector input is v(t) = s(t) = s(t-T) FA, cosl2af te $(t)] - A, cosl2nf (t-T) + ¢(t-T)) nf (2-7) + g(t) + Mt) 2af T + o(t) - g(t-T) = -2A, sinl- J sin J a e 2 2 where o(t) = 8 sin(2et,t) The phase difference g(t) - g(t-T) is ot) ~ (tT) = B sin(2af,t) - 8 sinl2nf,(t-T)) @lsin(2nf,t) - sin(2nf,t) cos(2mf,T) + cos(2nf,t) sin(2rf,T)] Blsin(2uet) - sin(2mf,t) + 2nf,T cos(2mf.t)) 2nA€T cos(2nft) where af: af. Therefore, noting that 2uf.T = 1/2, we may write ant T + Ht) - tT) sint- z J = sinlef.T + xAfT cos(2nft)) = sinl] + "eT cos(2ar,t)] 1 ¥2 cost rafT cos(2nft)] + v2 sin[ wAfT cos(2nf,t)] V2 + V2 wAeT cos(2nt,t) where we have made use of the fact that nAfT << 1, We may therefore rewrite Eq. (1) as v(t) = -2/2 ALC1 + wAfT cos(2afgt)) sintaf,(2t-7) + #0) + oCt-T) Accordingly, the envelope detector output is a(t) = 2 72 A [i+nafT cos(2af,t)] Which, except for a bias term, is proportional to the modulating wave. 104Problem 2.41 (a) In the time interval t-(T4/2) to t+(T4/2), assume there are n zero crossings. ‘The phase difference is 0,(t4T,/2) - 0(t-T,/2) = nt t O((t) = 2nf t + 2mky vs mt) dt. Also, the angle of an FM wave is Since m(t) is assumed constant, equal to my, 0j(t) = 2nf jt + 21kpmyt. Therefore, O,(b4T 4/2) ~ OAT 4/2) = (2, + 2mkym,) [t+T,/2 = (t=T,/2)]. = (xf, + 2rk—my) T,. But 48, (t) f(t) 2 Fp = amr, + Ankeny. Thus, 6, (aT 4/2) - Oy (t-T4/2) = f(t) Tye But this phase difference also equals nt. So, f(t) T, = on and f(t) = n/t, (b) For a repetitive ramp @s the modulating wave, we have the following set of waveformsLimiter output Pulse outputProblem 2.42, The complex envelope of the modulated wave s(t) is Blt) = alt) expljg(t)] Since a(t) is slowly varying compared to exp{jp(t)], the complex envelope S(t) is restricted effectively to the frequency band - B,/2 < f < B,/2. An ideal frequency discriminator consists of a differentiator followed by an envelope detector. The output of the differentiator, in response to 8(t), is F(t) = Ge He) = & fale) expljg(t)]) = SBD cor sgcty] + 3 28 acty explig(t)) a(t) expt so(t)) Catgy S202 + 3 $8) Since a(t) is slowly varying compared to #(t), we have 1 da(t) | > tetey SSE agit) I at Accordingly, we may approximate 7,(t) as ~ =G do(t) T(t) = 9 at) SE expr ject] However, by definition t Ht) = 2mke J mt) dt ° Therefore, T(t) = somky a(t) mt) expl jo(t)] Hence, the envelope detector output is proportional to a(t) m(t) as shown by I(t) 1 = 2nke a(t) mt) Problem 2.43 (a) The limiter output is 2(t) = sgnfa(t) cosl2nf.t + 9(t)]) 107Since a(t) is of positive amplitude, we have 2(t) = sgnfeost2nt.t + o(t)1) Let Vit) = 2afgt + g(t) Then, we may write senfeos ¥] = = ¢, exp(jny) ni ® Fy F santeos ¥ expl-sny) dv 1 272 1 FL G1 expleinpay + Ly (41) expl-gny) ay a 7 7 a ane 7 + £1 expl-iny ay ar ye eg = Faagay (exp Bb sexp( nnd sexp (8%) expat) ~exp(-jnndeexp(Hl85} Lire sinc®h sin(nt)] 2 (n-1)/2 | cn ’ n odd 9, n even If nz0, we find from Eq. (1) that e,20. Therefore, senteos yl = 2 2 2 cat)? exocjnyy eo n odd mre 4 (-1) e- £ cost p(2k+1)] F ly Beet We may thus express the limiter output as 108yk ae a(t) = 4 + eo cosl2nf .t(2k+1) + o(t) (2k+1)] @) (b) Consider the term cosl2nr t(2k+1) + O(t) (2k+1)] = Relexpl Jnr t(2k+1)Jexpl 4(t) (2k+1)]} 241) Re(expl J2nt,t(2e+1) Iexptj4(t)) 1 The function exp[{j¢(t)], representing the complex envelope of the FM wave with unit amplitude, is effectively low-pass in nature. Therefore, this term represents a band-pass Signal centered about #f,(2k+1). Furthermore, the Fourier transform of (explo(t)]}**! is equal to that of expljq(t)] convolved with itself 2k times. Therefore, assuming ee explj@(t)] is limited to the interval -B,/2 < f < By/2, we find that (exotsoces} “te limited to the interval (By /2) (+1) aaa (By/2) (241). Assuming that f > Bor as is usually the case, we find that none of the terms corresponding to values of k greater than zero will overlap the spectrum of the term corresponding to ks0, Thus, if the limiter output is applied to a band-pass filter of bandwidth By and midband frequency f,, all terms, except the term corresponding to k=0 in Eq. (2), are removed by the filter. ‘The resulting filter output is therefore vit) =F costantst + t)) We thus see that by using the amplitude limiter followed by a band-pass filter, the effect of amplitude variation, represented by a(t) in the modulated wave s(t), is completely removed. Problem 2.44 (a) Let the FM wave be defined by t s(t) = Ay cosl2nf.t + 2nky a mt) dt] Assuming that f, is large compared to the bandwidth of s(t), we may express the complex envelope of s(t) as 7 t s(t) = AS expl j2mk, 4 mt) dt] But, by definition, the pre-envelope of s(t) is © Rppencliz 2) 109s(t) = S(t) exp(j2nt,t) s(t) + 5 8(t) where 8(t) is the Hilbert transform of s(t). ‘Therefore, t s(t) + J8(t) = Ay explj2mk, J, me) dtd expt sentt t t Aleosl2nf.t + ark, Fyate) dtl + J sinlant,t + 2aky oo Equating real and imaginary parts, we deduce that t sinl2nf t+ 2nke f mt) dt] a) ce a Blt) = (>) For the case of sinusoidal modulation, we have m(t) =A, cos(2af.t) The corresponding FM wave is s(t) cosl2nf.t + 8 sin(2af,t)] where B= Ke A, Expanding s(t) in the form of a Fourier series, we get ae 4,(8) coslan(f,+nf,t] s(t) c Noting that the Hilbert transform of cos{2(f.+nf,,)t] is equal to sint2m(f.snf_)t], and using the linearity property of the Hilbert transform, we find that the Hilbert Eransform of s(t) is TE J,(8) sinlaw(r ent.) t] B(t) = ni =A, sinf2ne t + 6 sin(2nf_t)] This is exactly the same result as that obtained by using Eq. (1). In the case of sinusoidal modulation, therefore, there is no error involved in using Eq. (1) to evaluate ‘the Hilbert transform of the corresponding FM wave. Problem (a) The modulated wave s(t) is s(t) = expl-@(t)] coslanf.t + ¢(t)] 110Refexp(-9t)] expljanf.t + j9(t)]) Refexpl s2nfgt + JC g(t) + 59(t))]) Relexpl j2af.t + J9,(t)1) a where ¢,(t) is the pre-envelope of the phase function g(t), that is, $, CE) = g(t) + 39Ct Expanding the exponential function exp{jo,(t)] in the form of an infinite series: expl jo, (t)) @ Taking the Fourier transform of both sides of this relation, we may write n Flexpljg,(t)]) = 2 J erg%cty) + Bente For n>2, we may express ¢7(t) as the product of ¢,(t) and g@7"(t). Hence, Fatty] = @ COMEFLt)D where 6,(t)= @ (f), and %denotes convolution. Since @(f) = 0 for f < 0, it follows that for all n > 0, FLAC) = 0, for f <0 Hence, Flexplje,(t)]} = 0 © for £< 0 By using the frequency-shifting property of the Fourier transform, it follows that Flexplse,(t)1 exp(J2negty} = 0 for f < f, @) From Eq. (1), © a(t) = d fexplsoetgt + 39,(t1 + expl=g2nt gt = j¢n(t)) where g(t) is the complex conjugate of 4,(t). Therefore, Fls(t)] = 3 Flenpl iar gt + 39,(t))) + 3 Flexpl-s2ngt - Soet) Applying the conjugate-function property of the Fourier transform to £q. (3), we find that Flexpl-jemt.t - j¢ht)]) = 0, for f> ~ ilus Hence, it follows that the spectrum of s(t) is zero for -f, < f < f,. However, this Spectrum is of infinite extent, because the expansion of s(t) contains an infinite number of terms, as in eq. (2). (b) With $(t) = 8 sin(2af,t), we find that (t) = -8 cos(2nf,t) Therefore, 4,(0) = B sin(2at,t) = 38 cos(2at,t) zo JBloos(2nft) +j sin(2nft)] = = J8 expl nf.) Hence » expl J¢,(t)] = explB exp( j2nf,t)] = gn a ee am: . ae ar exP( Semnf, t) ‘The modulated wave s(t) is therefore s(t) = Relexp( j2nf.t) expl Jo, (t)1} 2 an = Relexp( janf.t) oe = oon = Ret 2 BF expljen(tent,)t]} n=0 cosl2n(f nf, )t]Problem 2.46 Alter passing the received signal through a narrow-band filter of bandwidth 8kH1z centered on Sfe= 200kHz, we get x(t) = Agm(t)cos(2mf,t) + n’(t) = Agn(t)cos(2nf 1) +1’ (t)cos(2nf,t)—n' (d)sin(2nf.1) Agn(t) + ,(1))cos(2nf 1) —n’g(0)sin(2nf,2) where n’(1) is the narrow-band noise produced at the filter output, and n’,(r) and n“9(t) are its in-phase and quadrature components. Coherent detection of x(t) yields the output v(t) = Agm(t) +n’ (0) The average power of the modulated wave is, AP 10W where P is the average power of m(2). To calculate the average power of the in-phase noise component n’,(1), we refer to the spectra shown in Fig. 1 Part (a) of Fig. 1 shows the power spectral density of the noise n(), and a superposition of the frequency response of the narrow-band filter. Part (b) shows the power spectral density of the noise n’,(t) produced at the filter output. + Part (c) shows the power spectral density of the in-phase component n’,() of n’(t) Note that since the bandwidth of the filter is small compared to the carrier frequency f,, we have approximated the spectral characteristic of n’(r) to be flat at the level of 0.5 x 10% watts/FHz. Hence, the average power of n’,(1) is (from Fig. Ic): (10° watts/Hz) (8 x 10) = 0.008 watts The output signal-to-noise ratio (SNR) is therefore 10 = 12 o.o08 = 125° Expressing this result in decibels, we have an output SNR of 31 dB. M3Problem 2.47 From Problen 5. 38, we note that the quadrature components of a narrow-band noise have autocorrelations : Ry (4) = Ry (4) = Ryle) cos(2nfit) + Ry(t) sin(2nf.r) T Q where Ry(t) is the autocorrelation of the narrow-band noise, Ry(t) is the Hilbert transform of Ry(t), and f, is the band center. The cross-correlations of the quadrature components are Ry y (= NN, 18 (a) For a DSBSC system, NN = Ry(t) sin(2nfyt) ~ Ry(t) cos(2nf,t) By (0) = Ry (0) = Ryle) cosCent) + Rye) sin(2xegn) Ry y (1) = Ry y Ce) = Ryle) sin(2net) - Ry(t) cos(2nf,t) ra Qt _ Where f, is the carrier frequency, and Ry(t) is the autocorrelation function of the narrow-band noise on the interval f,-W eALIT + kamtI1) < 645 then, with a probability greater than 1-6, we may say that yO) = ETAL + AS ey met) + mct)]2)72 That is, the probability that the quadrature component n,(t) is negligibly small is greater than 1 - 54. (b) Next, we note that if ke m(t) < -1, then we get overmodulation, so that even in the absence of noise, the envelope detector output is badly distorted. Therefore, in order to avoid overmodutation, we assume that k, is adjusted relative to the message signal m(t) such that the probability PAL + AL ky mt) + ny(t) <0) Then, the probability of the event yt) = ASET +k, mt) + nt) for any value of t, is greater than (1 = 6)(1 - 82). (c) When 6, and 6, are both small compared with unity, we find that the probability of the event, y(t) = ALL1 + ky m(t)] + no (t) for any value of t, is very close to unity. Then, the output of the envelope detector is approximately the same as the corresponding output of a coherent detector. Pr 2.52 The received signal is 123x(t) = A, cos(2nf.t) + n(t) A, cos(2mf.t) +n (t) cos(2mf.t) - ng(t) sin(2ef.t) (a, +n (t)] cos(2nf.t) ~ n.(t) sin(2nf.t) ‘The envelope detector output is therefore a(t) 2g 24) 172 {Ag + ney}? n’ct)) For the case when the carrier-to-noise ratio is high, we may approximate this result as a(t) =A, + ny (t) The term A, represents the useful signal component. ‘The output signal power is thus A®. ‘The Power spectral densities of n(t) and n{(t) are as shown below: s,(£) 2 t_ {Xo = o WW The output noise power is iW. The output signal-to-noise ratio is therefore (SNR)g = 124Problem 2.53 (a) From Section 1.12 of the textbook we recall that the envelope r(¢) of the narrow-band noise n(t) is Rayleigh distributed; that is fel) = fon] oy 20% 2 2 where of is the variance of the noise n(f). For an AM system, the variance oy is 2WNo. Therefore, the probability of the event that the envelope R of the narrow-band noise n(f) is large compared to the carrier amplitude A, is defined by P(R2A,) = f Sp(r)ar " 0) Define the carrier to noise ratio as average carriet_power bandwidth of the modul average noise por d message signal Since the bandwidth of the AM signal is 21, the average noise power in this bandwidth is 2WNo. The average power of the carrier is A2/2. The carrier-to-noise ratio is therefore Ar ° = TWN, ® (b) We may now use this definition to rewrite Eq, (1) in the compact form P(R2A,) = exp(-p) 4) Solving P(R > A,) = 0.5 for p, we get 125p= log? = 0.69 Similarly, for P(R > A. 0.01, we get p=logl00=4.6 ‘Thus with a carrier-to-noise ratio 10log 90.69 = -1.6 dB, the envelope detector is expected to be well into the threshold region, whereas with a carrier-to-noise ratio 10logi94.6 = 6.6 dB, the detector is expected to be operating satisfactorily. We ordinarily need a signal-to-noise ratio considerably greater than 6.6 dB for satisfactory intelligibility, and therefore threshold effects are seldom of great importance in AM receivers using envelope detection. Prol 54 (a) Following a procedure similar to that described for the case of an FM system, we find that the input of the phase detector is v(t) =A, costanf.t + o(t)] where nat) at) = ky mt) + c with ng(t) denoting the quadrature noise component. The output of the phase discriminator is therefore, nig(t) yt) = ky mt) + c The message signal component of yt) is equal to k, m(t). Hence, the average output signal power is Ko P, where P is the message signal power. With the post detection low-pass filter following the phase detector restricted to 126the message bandwidth W, we find that the average output noise power is 2HNy/A2, Hence, the output signal-to-noise ratio of the Pi system is x? pa? e AW (bv) The channel signal-to-noise ratio of the PM system is the same as that of the corres- ponding FM system. That is, (SNR)y = a (SNR )g = ago 0 Tne figure of merit of the PM system is therefore equal to ky P. For the case of sinusoidal modulation, we have m(t) = A, cos(2nf,t) Hence , a2 peat z The corresponding value of the figure of merit for a PM system is thus equal to 3 sy Where 8, = ky Aq. On the other hand, the figure of merit for an FM system with sinusoidal 3 2 modulation is equal to $ 6°. We see therefore that for a specified phase deviation, the FM systen is 3 times as good as the PM system. Problem 2.55 (a) The power spectral densities of the original message signal, and the signal and noise components at the frequency discriminator output (for positive frequencies) are illustrated below: Spectral densi of message Signa £ (kHz) Spectral density | ot signal component ot discriminator out ek £ (kHz) 4 8 32 ie 20 24 28 32 36 40 44 48 127Spectr density eb neve component of diserim: nator | i i 1 ube 1 ! 1 £ (kHz) ° 48 (b) Each SSB modulated wave contains only the lower sideband. Let A, and kf) denote the amplitude and frequency of the carrier used to generate the kth modulated wave, where — 4 kHz, and k = 1, 2, ..., 12. Then, we find that the kth modulated wave occupies the frequency interval (k - 1)f < If] < kfg. We may define this modulated wave by A My sy(t) = 5 m(t) cos(2nkf gt) + 5 A(t) sim(2nkf gt) where m(t) is the original message signal, and fi(t) is its Hilbert transform. Therefore, the average power of 5,(t) is A? P/u, where P is the mean power of mt). We may express the output signal-to-noise ratio ; for the kth SSB modulated wave as follows: ee BAL keCAD P/4) (SNR), of 3p fl 2N(ktS = (k - 19°85] 2 4242 Bae aR KEP 303x2 BN of 5 (3k - 3k + 1) where A, is the carrier amplitude of the FM wave. For equal signal-to-noise ratios, we must therefore choose the A, so as to satisfy the condition 2 K Re - eaT a constant for k = 1, 2 see, 12.Problem 2.56 The envelope r(t) and phase y(t) of the narrow-band noise n(t) are defined by x0 = nro + nd eee [ess] aye) For a positive-going click to occur, we therefore require the following: ny(t) = - Ae ng(t) has a small positive value 4 tant 2 |g at a Correspondingly, for a negative-going click to occur, we require nyt) ~ Ag ng(t) has a small negative value 8 tant [MA] 9 ad Ay)Problém 2.57 c V4) ex Mul! Let H(f) be Vi (fVy,(f, or the transfer function of the filter. At low frequencies, the capacitor behaves as an open circuit, Then, R_ LR HOD) # gt Thus, the low frequencies of the input are frequency-modulated. At high frequencies, the capacitor behaves as a short circuit in relation to the resistor. Then, H(t) = s ance, B+ Soare and Vour(t) = RC gS vy Ct) Frequency modulating the derivative of a waveform is equivalent to phase modulating the waveform. ‘Thus, the high frequencies of the input are phase modulated. Problem 2.58 (a) For the average power of the emphasized signal to be the same as the average power of the original message signal, we must choose the transfer function H,,(f) of the pre~ emphasis filter so as to satisfy the relation - 2 LS s Fel? Sy(nar With r s, — tcc 1+ (£/t9)2 o, elsewhere. 5 + £ Hoo(f) = kG + 2) we have w r af Ws (ify 130Solving for k, we get a (b) The improvement in output signal-to-noise ratio obtained by using pre-emphasis in the transmitter and de-enphasis in the receiver is defined by the ratio aw W 2 2 3s f° tH (At? at wee aw? fiat eK? 1 + (4/25) = —W>2 3. 2 3 k°(W/ fy) 7 £ (2) 3U(W/ty) ~ tan™"(W/E9)1 Substituting Eq. (1) in (2), we get 2 at (W/f))° tan” (W/fp) eee eee (3) BECW/fy) ~ tan™'(W/E5)I This result applies to the case when the rms bandwidth of the FM system is maintained the same with or without pre-emphasis. When, however, there is no such constraint, we find from Example 4 of Chapter 6 that the corresponding value of Dis 3 De (Wty) 3L(w/fy) = tan™"(W/fy)I a) In the diagram below, we have plotted the improvement D (expressed in decibels) versus the ratio Wf, for the two cases; when there is a transmission bandwidth constraint and when there is no such constraint: 131Ots> a decibels Problem 2,59 W/hy 1 3.16 10 31.6 100) In a PM system, the power spectral density of the noise at the phase discriminator output (in the absence of pre-emphasis and de-emphasis) is approximately constant. Therefore, the improvement in output signal-to-noise ratio obtained by using pre-emphasis in the transmitter and de-enphasis in the receiver of a PM systen is given by W fat 0 W 2 a Tg) Mae With the transfer function Hy9(f) of the de-enphasis filter defined by Hge(f) = 4 de’ T+ Gt) ' We find that the corresponding value of Dis w i a 0 1+ (£/f9) D= Wy ant tan™'(W/f,) For the case when W = 15 kz, f) = 2.1 kHz, we find that D = 5, or 7 4B, The corresponding value of the improvement ratio D for an FM system is equal to 13 dB (see Example 4 of Chapter 5), Therefore, the improvement obtained by using pre-emphasis and de-emphasis in a PM systen is smaller by an anount equal to 6 dB. 132Problem 2.60 Matlab codes 4 amplitude demodulation Yproblem 2.60, CS: Haykin % Mathini Sellathurai % message signal +t=[0:260] +10-8; ncein(2epiein, +: plot(t, m) xlabel('time (s)") ylabel (Amplitude? ) pause % amplitude modulated signal UFAM_wod(mue,m,ts,f¢); plot(t,u) xlabel (time (s)’) ylabel( Amplitude’) Pause 4% demodulated signal Let, demt]=AM_demod(me,u,ts,te) ; plot(tists, dent) xlabel(’tine (s)’) ylabel(’Amplitude’) axis([0 2.5e-3 0 21) 133function u=AM_mod(mue,m,ts,f¢) % amplitude modulation Yused in problem 2.60, CS: Haykin 4% Mathini Seliathurai % t=[0:1ength(m)-1] ts; cxcos(24pivtc.+t); m_n=n/nax(abs(m)); u=(14muetmn).#¢; 134function [t, env]=AM_denod(mue,m,ts,£c) % Amplitude demodulation Yused in problem 2.60, CS: Haykin % Mathini Sellathurai % ts=1/ts; fsofc=round(ts/tc); n2=Length(m); -zeros(1,round(n2/fsote)); % initializing the envelope (000; % load R €=0.010-6; % capacitor Yadenodulate the envelope 1=0; v(1)=m(1); for sofc:n2-fsofe 1142; v(l)=n(k)sexp(-ts/(R_L#C)/tsofe); % discharging v(LH)=m(keteote); “charging end 4% envelope t =0:fs0fc/2: (Length(v)-1)*£s0f¢/2; 135Answer to Problem 2.60 \A tine () Figure 1: Message signal 136ne ime =) Figure). Amplitude modulated signal me 6) Figureg: Demodulated signal 137Problem 2.61 Matlab codes 4% Problem 2.61 CS: Haykin 4% phase lock loop and cycle slipping %M. Sellathuras Y time interval t0=0;tf=28; % frequency step =0.125 Hz delt=0.125; u0=[0 ~delz+2¥pi]; [e,u)=ode23('1in’ [to tf] ,u0); plot (t,u(:,2)/2/pitdelf); xlabel('Tine (8)") ylabel(t_4 (t), (He)’) pause 4% frequency step =0.51 Hz delf=0.5; u0=[0 -deltepiv2]; Ce ul=ode2a(*1in’, [to t#],uo); plob(t,u(:,2)/2/pitdels); xlabel (Time (s)’); ylabel (fi (t), (Bz)"); pause; 4% frequency step =7/12 liz el#=7/12; 0 ~deltepiv2} "5 Ce ul=ode23(°2in’ , [eo e£],u0); plot (t,u(:,2)/2/pitdelt); xlabel(’Time (s)’}; ylabel(?£i (t), (Hz); pause; % frequency step =2/3 Hz delf=2/3; 0 ~delfepie2)'; [e,ulsode23(?1in’, [to t#],u0); plot (e,u(:,2)/2/pitdels); xlabel(’Time (s)'); ylabel(’fi (t), (Rz)'); 138function uprim =1in(t,u) % used in Problem 2,61, CS: Haykin PLL Transfer function (1+as)/(1+bs), gain K=50/2/pi, natural frequency 1/2/pi damping 0.707 Mathini Sellathurai uprim(2)=-(1/60+1 .2883¢cos(u(1)))eu(2)-sin(a(t)); uprim=uprim’ ;Answer to Problem 2.61 ) Figure (: Variation in the instantaneous frequency of the PLL’s voltage con- trolled oscillator for varying frequency step A f. (a) A f= 0.125 He 1401 “Tone te) Pigure 22 (b) A f= 0.5 Hz mee) Figures: (b) Af = 7/12 He 1412 Figure 4: (b) Af = 2/3 Hz 142CHAPTER 3 Pulse Modulation Problem 3.1 Let 2W denote the bandwidth of a narrowband signal with carrier frequency f,. The in-phase and quadrature components of this signal are both low-pass signals with a common bandwidth of W. According to the sampling theorem, there is no information loss if the in-phase and quadrature ‘components are sampled at a rate higher than 2W. For the problem at hand, we have fe = 100 kHz 2W= 10 kHz Hence, W = 5 kHz, and the minimum rate at which it is permissible to sample the in-phase and quadrature components is 10 kHz. From the sampling theorem, we also know that a physical waveform can be represented over the interval -0 << 00 by 8 = YD 4,0, (0) co) where {@,(f)} is a set of orthogonal functions defined as sin{nf,(¢-n/f,)} 8) = FWA where 1 is an integer and f, is the sampling frequency. If g(#) is a low-pass signal band-limited to W Hz, and f, > 2W, then the coefficient a, can be shown to equal g(n/f,). That is, for f, > 2W, the orthogonal coefficients are simply the values of the waveform that are obtained when the waveform is sampled every L/f, second, As already mentioned, the narrowband signal is two-dimensional, consisting of in-phase and quadrature components. In light of Eq. (1), we may represent them as follows, respectively: gilt) = DY an/F.)0,(2) 143solt) = DY soln/F)0,(0) Hence, given the in-phase samples af +) and quadrature samples sf?) . We may reconstruct oi fe the narrowband signal g() as follows: 8(1) = g(t)cos(27f,t) — go(t)sin(2nf,0) => [s{F) cos(2nf,t) - go sin(2nf0)]0,(8) where f, = 100 kHz and f, > 10 kHz, and where the same set of orthonormal basis functions is used for reconstructing both the in-phase and quadrature components. 144Problem 3.2 (a) Consider a periodic train o(f) of rectangular pulses, each of duration T. The Fourier series expansion of et) (assuming that a pulse of the train is centered on the origin) is given by ct) = YO f, sindinf, TexpGzmnf,t) where f, is the repetition frequency, and the amplitude of a rectangular pulse is assumed to be 1/T (ie., each pulse has unit area). The assumption that f,T>>1 means that the spectral lines (ie., harmonics) of the periodic pulse train e(t) are well separated from each other. Multiplying a message signal g(t) by c(t) yields s(t) = o(t)g(t) = Ef, sine(nf,T) g(t) expG2enf,0) os ‘Taking the Fourier transform of both sides of Eq.. (1) and using the frequency-shifting property of the Fourier transform: EG | oy eee ra @ nese where Gif) = Flg(t)]. Thus, the spectrum S(f) consists of frequency-shifted replicas of the original spectrum G(f), with the nthreplica being scaled in amplitude by the factor f,sine(nf,T). 145() In accordance with the sampling theorem, let it be assumed that « The signal g(t) is band-limited with Gf) =0 for -W2W Then, the different frequency-shifted replicas of G(f) involved in the construction of S(f) will not overlap. Under the conditions described herein, the original spectrum G(f), and therefore the signal g(t), can be recovered exactly (except for a trivial amplitude scaling) by passing s(t) through a low- pass filter of bandwidth W. Problem 3.3, (a) g(t) = sine(200¢) This sinc pulse corresponds to a bandwidth W = 100 liz. Hence, the Nyquist rate is 200 liz, and the Nyquist interval is 1/200 seconds. (b) g(e) = sine” (200t? This signal may be viewed as the product of the sinc pulse sinc(200t}’ with itself. Since multiplication in the time domain corresponds to convolution in the frequency domain, ve find that the signal g(t) has a bandwidth equal to twice that of the sinc pulse sin(200t}; that is, 200 Hz. The Nyquist rate of g(t) is therefore 400 Ha, and the Nyquist interval is 1/400 seconds. (e) g(t) = sine(200e), + sinc”(200e) The bandwidth of g(t) is determined by the highest frequency component of sinc(200t) or sinc2(200t}, whichever one is the largest. With the bandwidth (ive., highest frequency component of) the sinc pulse sinc(200 t) equal to 100 Hz and that of the squared sinc pulse sinc2(200t}) equal to 200 Hz, it follows that the bandwidth of g(€) is 200 Hz. Correspondingly, the Nyquist rate of g(t) is 400 Ilz, and its Nyquist interval is 1/400 seconds. 146Problem 3.4 (a) The PAM wave is s(t) = E (1 + um! (nT) Je(tent,), where g(t) is the pulse shape, and mt(t) = m(t)/A, = cos(2mf,t). The PAM wave is equivalent to the convolution of the instantaneously sampled [1 + um'(t)] and the pulse shape g(t): s(t) = { E [taum'(nt)] Stent.) } xe at) nese = {Ct4nm'(t)1 8 (tnt) } Ye @(t) Tne spectrum of the PAM wave is, SCA) = {CCC some) ye or ocr = By} acer 5 1 2 2 ett) E c6cr = By sume By] 1; m-° Ts Ts For a rectangular pulse g(t) of duration T:0.45s, and with AT = we have: G(f) = AT sine(fT) = sinc(0.45f) For m'(t) = cos(2™fmt), and with ty = 0.25 Hz, we have: mice) = 4 6¢#-0.25) + 8(140.25)) For T, = 18, the ideally sampled spectrum is $,(f) = [6(fom) + uMt(fom)J. $a) 3 me a =1as” hd =, 35 9 The actual sampled spectrum is 147s(t) sine(0.U5f)[6(f-m) + pM! (f-m)] Ss) ors 07 0.63% i outs onttys estes outSta Ae *s t t f 1 | ae) 2 lO 01s =O GO 0.35 07 VO Lae (b) The ideal reconstruction filter would retain the centre 3 delta functions of S(f) or: f z — sas 0 bas With no aperture effect, the two outer delta functions would have amplitude 3. Aperture effect distorts the reconstructed signal by attenuating the high frequency portion of the message signal. Problem 3.5 The spectrum of the flat-top pulses is given by H(f) = Tsinc ( fT)exp(—jnfT) = 10™*sinc (10 f)exp(—jmf 10) Let s(f) denote the sequence of flat-top pulses: s(t) = SY m(nTh(t-n7,) ‘The spectrum Sf) = Fls(#)] is as follows: Sf) =f MUP - KH) kaw = fH) DY MUP EF) no ‘The magnitude spectrum |S(f)| is thus as shown in Fig. Ic. 1484 \rol -1/T a (a) wT f 1 (Me fs Ww f, (b) u Sif) 7 0 Wt 1/T © f VT = 10,000Hz f, = 1,000Hz W =400Hz Figure 1 149Problem 3.6 At f= 1/27, which corresponds to the highest frequency component of the message signal ae sampling. rate equal to the Nyquist rate, we find from Eq. (6-19) that the amplitude respor sine (O5777,) lea cndiion 1.0) 7 re $s 34 35 38 buy eve 17, Figure 1 of the equalizer normalized to that at zero frequency is equal to 1 (n/2XT/T) sine(O.ST/T,)_sin[(n/2NT/T,)] where the ratio T/T, is equal to the duty cycle of the sampling pulses. In Fig. 1, this result is plotted as a function of T/T,. Ideally, it should be equal to one for all values of 7/T,. For a duty cycle of 10 percent, it is equal to 1.0041. It follows therefore that for duty cycles of less than 10 percent, the aperture effect becomes negligible, and the need for equalization may be omitted altogether. 150Problem 3.7 Consider the full-load test tone A cos(2tf,t). Denoting the kth sample amplitude of this signal by A), we find that the transmitted pulse is Ay g(t), where g(t) is defined by the spectrum: co If < Bp Gf) = otherwise The mean value of the transmitted signal power is 1 its, it 2, Eins sy Ft ray g(ty iat) Lre UTS our, L 2 2 ALA, eo(t) dt) LT, ELLA S © e@tyat ba -LT, where T, is the sampling period. However, ELAAL] = o, otherusie Therefore, e(tyat Using Rayleigh's energy theorem, we may write L gtat =f peceyi2ar 151Therefore, ae The average signal power at the receiver output is A°/2. He th to-noise ratio is given by pass ness She output signal By choosing B,=1/2T,, we get (SNR)g This shows that PAM and baseband signal transmission have the same signal-to-noise ratio for the same average transmitted power, with ‘additive white Gaussian noise, and assuming the use of the minimum transmission bandwidth possible. Problem 3.8 (a) The sampling interval is T, = 125 ¥s, There are 24 channels and 1 syne pulse, so the time alloted to each channel i8 1, = T,/25 = 5 Us. The pulse duration is 1 Ms, so the time between pulses is 4 Us. (b) If sampled at the nyquist rate, 6.8 kHz, then T, = 147 vs, T, = 6.68 Us, and the time between pulses is 5.68 us. Problem 3.9 (a) The bandwidth required for each single sideband channel is 10 kHz. The total bandwidth for 12 channels is 120 kHz. (>) The Nyquist rate for each signal is 20 kHz. For 12 TDM signals, the total data rate is 2N0 kliz, By using a sinc pulse whose amplitude varies in accordance with the modulation, and with zero crossings at multiples of (1/240) ms, we need a minimum bandwidth of 120 KHz. 152Problem 3.10 (a) The Nyquist rate for s,(t) and s.(t) is 160 Hz. Therefore, at must be greater than 160, and the maximun R is 3: (b) With R= 3, we may use the following signal format to multiplex the signals s,(t) and 8,(t) into a new signal, and then multiplex s,(t) and 8,(t) and 9,(t) including markers for synchronization: Macker s eA poytid Time SS 535 535 Se SSS SSS S, @ zero sampas Based on this signal format, we may develop the following multiplexing system: 2400 Hz. Clock, FL - is Delay Dato baencett [serete} ae samt fe 5,ce (e) M [828 [e poten] v x Mutbixploud [signa xcz a Sampler, 153Problem 3.11 In general, a line code can be represented as N s(t) = YS a,g(t—n,) na Let g(t) = G(f). We may then define the Fourier transform of s(t) as N SN = DY a,GNe nN -jont, a jon’ =a) Sage nN where «= 2zf. The power spectral density of s(¢) is See | n= SA) = lim [How 5 Ii T+ a ae i(m=n)oT, =IG()P lim # Y Flajagle” ' ro n= om=N where T is the duration of the binary data sequence, and E denotes the statistical expectation operator. Define the autocorrelation of the binary data sequence as R(k) = Ela,a,.,) Indy + By letting m =n +k and T= (2N + 1)T,, we may write ig? SA) = I6(Al inde ; No keNen ar 8,7) = GDP im I T, N-=| 2N41 154- IeanP oly “T Zwe @ where 1 RU) = Eldan gl = Sanne .dP; @ in where p; is the probability of getting the product (dy dn4,); and there are J possible values for the 4, dng product. G(/) is the spectrum of the pulse-shaping signal for representing a digital symbol. Eqs. (1) and (2) provide the basis for evaluating the spectra of the specified line codes. (a) Unipolar NRZ signaling For rectangular NRZ pulse shapes, the Fourier-transform pair is a= Aree x) FG(f) = AT,sinc(fT,) For unipolar NRZ signaling, the possible levels for a’s are +4 and 0. For equiprobable symbols, we have the following autocorrelation values: 20) = tare to =A/2 4 RU) = SnAg 2):Pi ia 2 2 Aan OMOMOEAg = Fafa de Ga 4 for dl>o Thus Rk) =f A fork = 0 GB) 74 for k#0Therefore, the power spectral density for unipolar NRZ signals, using formulas (1) and (3), is 3e i ae som t ral TA 4 is But, where 8(f) is a delta function in the frequency domain. Hence, AT, 2 SCA) = sino Ty) ia] We also note that sinc(f7,) = 0 at f = +,n#0; we thus get AT, sue teine(sTy)[ +5 (b) Polat Non-return-to-zero Signaling For polar NRZ signaling, the possible values for a’s are +A and -A, Assuming equiprobable symbols, we have RO) = Say4,),P; ist For k #0, we have 156‘ RE) = YA ntg 0):Pi i=l 5 CA) 5 =0 Thus, R(k) = | A for k= 0 4) 0 for k#0 ‘The power spectral density for this case, using formulas (1) and (4), is S(f) = A’T,sine*(fT,) (c) Return-to-zero Signaling The pulse shape used for return-to-zero signaling is given by (x ) We therefore have T,/2 GY) spPsine(f1,/2) The autocorrelation for this case is the same as that for unipolar NRZ signaling. Therefore, the power spectral density of RZ signals is 2 aT, Sf) = sine’ gancTall+ (d) Bipolar Signals The permitted values of level a for bipolar signals are +A, -A, and 0, where binary symbol 1 is represented alternately by +4 and -A, and binary 0 is represented by level zero. We thus have the following autocorrelation function values:Fork>1, RW) = Day. Thus, as for k = ae AS for (a | ° 0 for |k| > 1 ‘The pulse duration for this case is equal to T,/2. Hence, Gin = Bai ine 22) © Using Equations (1), (5) and (6), the power spectral density of bipolar signals is jot, -joT,, be | sine ZA 1 -cos(2n/T,)] 158(ec) Manchester Code, The permitted values of a’s in the Manchester code are +A and -A. Hence, latte ata lta leg? R(O) = GAP HAY + GCAY? + 34?) =a For k#0, 4 RE) = Daya, dis = =0 Thus, 2 mua 4 for k = 0 0 for k#0 U The pulse shape of Manchester signaling is given by tT y/4 (t-T,/4y Hn = 0S) (FH) The pulse spectrum is therefore Gf) fT) ~ioty/4 sine) 3 srind p22) 159Therefore, the power spectral density of Manchester NRZ has the form Problem 3.12 Power spectral density of a binary data stream will not be affected by the use of differential encoding. The reason for this statement is that differential encoding uses the same pulse shaping functions as ordinary encoding methods. If the number of bits is high, then the probability of a symbol one and symbol zero are the same for both cases, Problem 3.13 (a) | (mt) Ty oT (b) an = ): 0, otherwise Equivalently, we may write a(t) = cos( FF Aree) where rect(®) is @ rectangular function of unit amplitude and unit duration. The Fourier transform of g(t) is given by Gf) = FS r-A rez] sinc(fT,) where A denotes the pulse amplitude and * denotes convolution in the frequency domain. 160Using the replication property of the delta function 8(/), we get + sine (rls + asine(n(/-F)) sine (7 or J 7) a 2 Note that the two spectral components sine (Ti f 2) and sine (rf fe 7) overlap in 0 the frequency interval -(1/T,) < f < (I/T;), hence the presence of cross-product terms in Eq. (). Figure 1 plots the normalized power spectral density S()/(A77)/4) versus the normalized frequency fT), The interesting point to note in this figure is the significant reduction in the power spectrum of the puls shaped data stream x(t) in the interval -L/T, << Ty (©) The power spectral density of the standard form of polar NRZ signaling is S(f) = AT ysine*(fT,) Comparing this expression with that of Eq. (1), we observe the following differences: SS Polar NRZ ing | Polar NRZ signals using cosine pulses rectangular pulses feo 0 “7, f= 22/T, ATy4 0 161 @)rk 162 14 12h = = © wt ws” oatProblem 3.14 (4 Ch a TELE = FR Ppl ps © HILAL 163roblem 3.15(a) PrProblem 3.15(b) cay" dn tt boo tot 165Problem 3.16 ‘The minimum number of bits per sample is 7 for a signal-to-quantization noise ratio of 40 dB. Hence, (™* number of sampl in a duration of 10s ) = 8000 10 8x10" samples The minimum storage is therefore =7x8x 10* = 5.6.x 105 = 560 kbits 166Problem 3.17 Suppose that baseband signal m(t) is modeled as the sample function of a Gaussian random process of zero mean, and that the amplitude range of m(t) at the quantizer input extends from 4Armg t0 4Arm,: We find that samples of the signal m(t) will fall outside the amplitude range 8Aymg with a probability of overload that is less than 1 in 10* If we further assume the use of a binary code with each code word having a length n, so that the number of quantizing levels is 2", we find that the resulting quantizer step size is 3 = SAnms @ aR Substituting Eq. (1) to the formula for the output signal-to-quantization noise ratio, we get (SNR), = son @ Expressing the signal-to-noise ratio in decibels: 10logg(SNR) = GR - 7.2 (3) This formula states that each bit in the code word of a PCM system contributes 6dB to the signal- to-noise ratio. It gives a good description of the noise performance of a PCM system, provided that the following conditions are satisfied: 1, The system operates with an average signal power above the error threshold, so that the effect of transmission noise is made negligible, and performance is thereby limited essentially by quantizing noise alone. 2. ‘The quantizing error is uniformly distributed. 3. The quantization is fine enough (say R > 6) to prevent signal-correlated patterns in the quantizing error waveform. 4. The quantizer is aligned with the amplitude range from -4Armg 0 4Armg- In general, conditions (1) through (3) are true of toll quality voice signals. However, when demands on voice quality are not severe, we may use a coarse quantizer corresponding to R < 6. In such a case, degradation in system performance is reflected not only by a lower signal-to-noise ratio, but also by an undesirable presence of signal-dependent patterns in the waveform of quantizing error. 167Problem 3.18 (a) Let the message bandwidth be W. Then, sampling the message signal at its Nyquist rate, and using an R-bit code to represent each sample of the message signal, we find that the bit duration is le * aR The bit rate is 1 wt BR % ‘The maximum value of message bandwidth is therefore 10° 7 7) Ynax = “2 x x = 3.57 x 10° te (b) The output signal-to-quantizing noise ratio is given by (see Example 2): 10 10819 (SNR) = 1.8 + OR 1.8467 43,8 dB Problem 3.19 Let a signal amplitude lying in the range 1 1 HO sr iy tg hy be represented by the quantized amplitude x,. The instantaneous square value of the error is (exy)*, Let the probability density function of the input signal be f(x). If the step size 6; is small in relation to the input signal excursion, then fy(x) varies little within the quantum step and may be approximated by fy(x. )+ Then, the mean-square value of the error due to signals falling within this quantum is ? tyra 168a (2) Therefore, eliminating fy(x,) between Eqs. (1) and (2), we get 2,22 EIQf] = + The total mean-square value of the quantizing error is the sum of that contributed by each of the several quanta. Hence, 169Problem 3.20 (a) = = Voltage Quaticn output Quadkzer eutpok Voltane Tyla. a Tone es ey it 83 8% * . Quoted one anpor (>) 10Problem 3 The quantizer has the following input-output curve: Ouspur At the sampling instants we havet t m(t) code 3/8 2 0011 -1/8 v2 0011 1/8 3v2 1100 43/8 3v2 1100 And the coded waveform is (assuming on-off signaling): a pouty Time (Seconds) + 3 8 7 roblem 3.2: The transmitted code words are: t/t, code 1 001 010 3 on 4 100 5 101 6 110 mtThe sampled analog signal is Problem (a) The probability p, of any binary symbol being inverted by transmission through the system is usually quite small, so that the probability of error after n regenerations in the system is very nearly equal ton p,. For very large n, the probability of more than one inversion must be taken into account. Let p, denote the probability that a binary symbol is in error after transmission through the Complete system. ‘Then, p, is also the probability of an odd mumber of errors, since an even number of errors"restores the original value. Counting zero as an even number, the probability of an even number of errors is 1-p,. Hence Pp (1-P4)+ (1-1 Poet Py C2 PLHP, This is a linear difference equation of the first order. Its solution is 1 a Py =e (1-1-2, (bv) If p, is very small and n is not too large, then (1-2) 9" = 1-2) yn and 1 Py = gli-(-2p,n)] = pyr 172Problem 3.24 - Regenerative repeater for PCM Three basic functions are performed by regenerative repeaters: equalization, timing and decision- making. Equalization: The equalizer shapes the incoming pulses so as to compensate for the effects of amplitude and phase distortion produced by the imperfect transmission characteristics of the channel. ‘Timing: The timing circuitry provides a periodic pulse train, derived from the received pulses, for sampling the equalized pulses at the instants of time where the signal-to-noise ratio is maximum. Decision-making: The extracted samples are compared to a predetermined threshold to make decisions. In each bit interval, a decision is made whether the received symbol is 1 or 0 on the basis of whether the threshold is exceeded or not. Problem 3.25 m(1) = Atanh(Br) To avoid slope overload, we require |dm(1)| ml | w am) = 4 Bsech?(Br) Q) dt Hence, using Eq. (2) in (1): A> max(ABsech*(Br)) xT, @) Since sech(Br) = —1. cosh (Bi) Bi it follows that the maximum value of sech(fr) is 1, which occurs at time 1 = 0, Hence, from Eq. (3) we find that A > ABT, 173Problem 3.26 The modulating wave is mt) = A, cos(2nf,t) The slope of m(t) is an(t) at W2nf hy sin(2nft) The maximum slope of m(t) is equal to 2nf,A, The maximum average slope of the approximating signal m,(t) produced by the delta modulator is 6/T,, where 6 is the step size and T, is the sampling period. The limiting value of 4, is therefore given by é anf, > S mn ? TS or 6 A> sq in? Bef, Ty Assuming a load of 1 ohm, the transmitted power is AC/2, Therefore, the maximum power that may be transmitted without slopeoverload distortion is equal to 6°/8x"r272, 174Problem 3.27 Is= Wxyquist Frxyquist = 6-8 KHz f= 10x 6.8 x 10° = 6.8 x 10* Hz. A 2.>max| idm(1)) dt For the sinusoidal signal m(t) = A,,Sin(2rtfyf), we have POs raf pA yC08(2%F yf) dt Hence, |dm(t)| 2: ri em = Pha or, equivalently, 4 T, Ti Arb max Therefore, Andean = oto mlmax ~ TX 20x Fy ~ AL: 2th m 01x68 x 10° 2nx 10° u 1osv 175Problem 3.28 (a) From the solution to Problem 3.27, we have a) ‘The average signal power = £ =e) With slope overload avoided, the only source of quantization of noise is granular noise Replacing A/2 for PCM with A for delta modulation, we find that the average quantization noise power is A7/3; for more details, see the solution to part (b) of Problem 3.30. The waveform of the reconstruction error (i.e., granular quantization noise) is a pattern of bipolar binary pulses characterized by (1) duration = T, = I/f,, and (2) average power = A/3. Hence, the autocorrelation function of the quantization noise is triangular in shape with a peak value of A?/3 and base 27,, as shown in Fig. 1: Rol) > v3 Fig. 1 From random process theory, we recall that SoAjao = [Rola which, for the problem at hand, yields ‘Typically, in delta modulation the sampling rate f, is very large compared to the highest frequency component of the original message signal. We may therefore approximate the power spectral density of the granular quantization noise as 1762 Soif=) A3F, “WS SSW 0, otherwise where W is the bandwidth of the reconstruction filter at the demodulator output. Hence, the average quantization noise power is 20°W 3f5 w N= J So(fdf = Q) -w Substituting Eq. (2) into (1), we get of 2 LmAY? W wa (Say _ 8x fw af (b) Correspondingly, output signal-to-noise ratio is (3) SNR = —— >, (81° f,A°W)/3f? ah lor Ww Problem 3.29 @as dhs 2 hn 172xmx 10x af 0) (SNR) = SS ax’ fw 3 fe e3e (50x13) 16m? 10°x5x 10° = 475 In decibels, (SNR) oq, = LOlog 49475 = 268 4B Problem 3.30 (@) For linear delta modulation, the maximum amplitude of a sinusoidal test signal that can be used without slope-overload distortion is Af, A= aap, 0.1.x 60x 10° 3 —veo ff, = 2x3x10° 2nx 1x10 = 0.95V (b) i) Under the pre-filtered condition, it is reasonable to assume that the granular quantization noise is uniformly distributed between -A and +A. Hence, the variance of the quantization noise is 178(SNR), prefiltered = _ 3x 0.95? Sieean ik 2x0. = 135 = 21.3 dB (ii)The signal-to-noise ratio under the post-filtered condition is 3 (3) =F (Weestneea ” Ten? fa (60 (1y’x3 = 1367 = 313 dB ‘The filtering gain in signal-to-noise ratio due to the use of a reconstruction filter at the demodulator output is therefore 31.3 - 21,3 = 10 dB. 179Problem 3.31 Let the sinusoidal signal m(t) = Asinaog, where @p = 2xfo The autocorrelation of the signal is 2 R(t) = mn c0s(Wot) A 2a _ ( 1 R,(L) = 5 °°5( %0%* ia) 2 a = F.e0s(0.1) For this problem, we thus have Ry, = (Ry, (O)) (R,(1)] (a) The optimum solution is given by Wo = Rie, 2 Fos (0.1) = 2 = 00s(0.1) ae 2 = 0.995 (©) Jig = Ry(0) = TRG Ey 2 2 2 5 4-4 .08(0.1) x 4-cos(0.1)/(A?/2) 180(1 cos?(0.1)) np = 0.00547 Problem 3.32 1 08 06] = |08 1 08) 0.608 1 = [os, 06, 04)” (a) Wo = RY -1 1 08 06] [os = |08 1 0.8} 0.6) 0.608 1} [o4| 0.875 =l 0 -0.125 (©) Jin = R,(0)— 4, Ri 'r, r, = R,(O0)—1 Wo 0.875 08,06, 0.4]| 0 —0.125] (0.8 x0.875- 04x 0.125) 0.7 + 0.05 0.35 181Problem 3.33 RK =|! 08 08 1 (@) wo = Rr = | 0.8889 -O.1LIL () Imin = R,(O)— ERY 10.6444 = 0.3556 which is slightly worse than the result obtained with a linear predictor using three unit delays (ic., three coefficients). This result is intuitively satisfying. Problem 3.34 Input signal variance = R,(0) The normalized autocorrelation of the input signal for a lag of one sample interval is, = 0.75 Error variance = R,(0)-R,(1)Rz!(0)R,(1) = R,(0)(1-p2(1)) 182RO) Processing gain. = ———*—__ RO) ~P-(1)) 2.2857 Expressing the processing gain in dB, we have 3.59 dB 10log 19(2.2857) Problem 3.35 Processing gain (a) Three-tap predictor: Processing gain = 2.8571 4.56 dB (b) Two-tap predictor: Processing gain = 2.8715 4.49 dB Therefore, the use of a three-tap predictor in the DPCM system results an improvement of 4.56 - 4.49 = 0.07 dB over the corresponding system using a two-tap predictor. Problem 3. (a) For DPCM, we have 10log9(SNR)o = 01 + 6n dB For PCM, we have 10logyg(SNR)g = 4.77 + 6n - 20log;o(log(1 + 11)) where n is the number of quantization levels SNR of DPCMSNR = ce + 6n, where -3<0< 15 For n=8, the SNR is in the range of 45 to 63 dBs. SNR of Pt SNR = 4.77 + 6n - 20logyo(log(2.56)) 4.77 +48 - 14.8783 =38dB Therefore, the SNR improvement resulting from the use of DPCM is in the range of 7 to 25, 4B. (b) Let us assume that m) bits/sample are used for DPCM and » bits/sample for PCM If = 15 dB, then we have 15 +6m, = 6n- 10.0 10+15, 6 Rearranging: (n ~ nj) = 4.18 which, in effect, represents a saving of about 4 bits/sample due to the use of DPCM. If, on the other hand, we choose -3 dB, we have -3 +6n, = 6n - 10 Rearranging: (n-n,) = 12-3 = 101 which represents a saving of about 1 bit/sample due to the use of DPCM. 184,Problem 3.37 The transmitting prediction filter operates on exact samples of the signal, whereas the receiving prediction filter operates on quantized samples. Problem 3.38 Matlab codes 4% Problem 3.38, CS: Haykin Ylat-topped PAM signal Yand magnitude spectrum o Mathini Sellathurai Yaata £3=8000; % sample frequency ts=1.260-4; Yit/ts pulse_duration=5e-S; Ypulse duration % sinusoidal sgnai; 260-6; Ysampling frequency of signal £4=80000; t= (0:td: 100¥td); m=10000; sesin(fmet); % PAM signal generation pan_s=PAM(s,td, ts, pulse figure(1);hold on 185plot(t,s,’—-"); plot(t(1:length(pam.s)),pam_s); xlabel('tine’) ylabel (‘magnitude’) Legend ‘signal’, *PAN-signal’); % Computing magnitude spectrum S(#) of the signal a=((abs(t2t(pam.s)).-2)); a=a/nax(a); f=t50(ts/24:ts%(t5/ta) : CLength(a))+£a¢ (£2/24); figure (2) plot(t,a); xlabel(’frequency’); ylabel (‘magnitude’) 4% finding the zeros index=tind(a ost > Output 1 Fitter matched > to 59(0) > Output 2 — input Filter matched - to 54(0) Output Fig. 4 202Problem 4.3 Ideal low-pass filter with variable bandwidth. The transfer function of the matched filter for a rectangular pulse of duration t and amplitude A is given by Hopt(®) = sine(fT)exp(-jnfT) @ The amplitude response |H,,,(f)| of the matched filter is plotted in Fig. 1(a). We wish to approximate this amplitude response with an ideal low-pass filter of bandwidth B. The amplitude response of this approximating filter is shown in Fig. 1(b). The requirement is to determine the particular value of bandwidth B that will provide the best approximation to the matched filter. We recall that the maximum value of the output signal, produced by an ideal low-pass filter in response to the rectangular pulse occurs at t = T/2 for BT < 1. This maximum value, expressed in terms of the sine integral, is equal to (2A/x)Si(nBT). The average noise power at the output of the ideal low-pass filter is equal to BNy. The maximum output signal-to-noise ratio of the ideal low-pass filter is therefore = (2A/n)?Si2(nBT) (2) (SNR), a 0 Thus, using Eqs. (1) and (2), and assuming that AT = 1, we get (SNR, 2 _2_ 5 2¢¢8T) GNR, BT This ratio is plotted in Fig. 2 as a function of the time-bandwidth product BT. The peak value on this curve occurs for BT = 0.685, for which we find that the maximum signal-to-noise ratio of the ideal low-pass filter is 0.84 dB below that of the true matched filter. Therefore, the " best” value for the bandwidth of the ideal low-pass filter characteristic of Fig. 1(b) is B = 0.685/T. 203Hope (fl 10 ae 2 a r r (a) a a \H(f)l 1.0 f -B O B (b) Figure 1 10 Matched filter eos ores hie s\e B/S os ee a 0.2 Figure 2 204Problem 4.4 output of the low-pass RC filter, produced by a rectangular pulse of amplitude A and duration T, is as shown below: Soft) A(L-exp(-2nfT)) | — 0 T t The peak value of the output pulse power is. Poy = ALL ~ exp(-20 fT) where fy is the 3-dB cutoff frequency of the RC filter. The average output noise power is No Nout = om — Noto “8 The corresponding value of the output signal-to-noise ratio is therefore Nenfglt ~ HP2RF7I Differentiating (SNR)p with respect to foT and setting the result equal to zero, we find that (SNR)ouy attains its maximum value at 0.2 fo ‘The comesponding maximum value of (SNR)oyx isol - exp(-0.40) 7 For a perfect matched filter, the output signal-to-noise ratio is 2E (SNR)o matched = No os No Hence, we find that the transmitted energy must be increased by the ratio 2/1.62, that is, by 0.92 4B so that the low-pass RC filter with fy = 0.2/T realizes the same performance as a perfectly matched filter. Problem 4.5 ©) Po>Pr The transmitted symbol is more likely to be 0. Hence, the average probability of symbol error is smaller when a 0 is transmitted than when a 1 is transmitted, In such a situation, the threshold A in Figs. 4.5(a) and (b) in the textbook is moved to the right. Gi) Py > Po ‘The transmitted symbol is more likely to be 1. Hence, the average probability of symbol error is smaller when a 1 is transmitted than when a 0 is transmitted. In this second situation, the threshold din Figs. 4.5(a) and (b) in the textbook is moved to the left. 206Problem 4.6 ‘The average probability of error is Pe = pr {% fyly Idx + po f° fyty lnndx @ An optimum choice of ) corresponds to minimum P.,. Differentiating Eq. (1) with respect to A, we get: ap, x = py fy IL) - pofy lo) ap, Setting —* = 0, wo got the following condition for the optimum value off: £yQope I) _ Po FyMopt 10) Pi which is the desired result. 207Problem 4.7 Ina binary PCM system, with NRZ signaling, the average probability of error is P, = d erk fe 2 0 The signal energy per bit is Ey = A?T) where A is the pulse amplitude and 7, is the bit (pulse) duration. If the signaling rate is doubled, the bit duration Ty, is reduced by half. Correspondingly, Ey, is reduced by half. Let u = ¥E,/No. We may then set P, = 10-6 = 3 erfe(u) Solving for u, we get u=33 When the signaling rate is doubled, the new value of P, is P, = 2 erfe| & og ote = 2 erte(2.33) 2 = 10°. 208where Ey, = AT. We may rewrite this formula as, P, = 1 erfe(A @ 2 o where A is the pulse amplitude at o = VNoT,. We may view 0” as playing the role of noise variance at the decision device input. Let al> We are given that o? = 10 volts, 6 = 0.1 volt P, = 10° Since P, is quite small, we may approximate it as follows: erfe(u) = 2xe(-u?) vru‘We may thus rewrite Eq. (1) as (with P, = 10%) en Nir = 108 Solving this equation for u, we get u = 3.97 ‘The corresponding value of the pulse amplitude is A=ou=0.1x 3.97 = 0.397volts (b) Let 6%, denote the combined variance due to noise and interference; that is 2 tag? Baotedl where o” is due to noise and 6%, is due to the interference. The new value of the average probability of error is 10°. Hence 106 = t we(] 2 oT (2) = + erfe(u) where een Or 210Equation (2) may be approximated as (with P, = 10°) exp(-up) 2a up 106 Solving for up, we get up = 3.37 ‘The corresponding value of 02, is 2 o8 = ‘AN. (0-397 F ~ 0.0138 volts? uy) (3.37 ‘The variance of the interference is therefore 2 of = op - 0? = 0.0138 - 0.01 = 0.0038 volts”Problem 4.9 Consider the performance of a binary PCM system in the presence of channel noise; the receiver is depicted in Fig. 1. We do so by evaluating the average probability of error for such a system under the following assumptions: 1. The PCM system uses an on-off format, in which symbol 1 is represented by A volts and symbol 0 by zero volt. 2, The symbols 1 and 0 occur with equal probability. 3. The channel noise w(t) is white and Gaussian with zero mean and power spectral density Ny2. To determine the average probability of error, we consider the two possible kinds of error separately. We begin by considering the first kind of error that occurs when symbol 0 is sent and the receiver chooses symbol 1. In this case, the probability of error is just the probability that the correlator output in Fig. 1 will exceed the threshold 4 owing to the presence of noise, so the transmitted symbol 0 is mistaken for symbol 1. Since the a priori probabilities of symbols 1 and O are equal, we havep, = p, Correspondingly, the expression for the threshold A simplifies as follows: raat @ 2 where TT, is the bit duration, and A", is the signal energy consumed in the transmission of symbol 1. Let y denote the correlator output: y = £7 stoxtat (2) Under hypothesis Ho, corresponding to the transmission of symbol 0, the received signal x(t) equals the channel noise w(t). Under this hypothesis we may therefore describe the correlator output as Hoy =A ie w(t)dt @) Since the white noise w(t) has zero mean, the correlator output under hypothesis Hy also has zero mean, In such a situation, we speak of a conditional mean, which (for the situation at hand) we Aladescribe by writing No = ELY [Hol = 5| ic wood | Co) where the random variable Y represents the correlator output with y as its sample value and W(t) is a white-noise process with wit) as its sample function. The subscript 0 in the conditional mean Hp tefers to the condition that hypothesis Hy is true. Correspondingly, let oy denote the conditional variance of the correlator output, given that hypothesis Hy is true. We may therefore write 08 = ELY? [Hol 6) : 3| G7 L7 wenwepraty dt, The double integration in Eq. (5) accounts for the squaring of the correlator output. Interchanging the order of integration and expectation in Eq. (5), we may write of = £7 [7 EIWit)Wty)ldty dt, © = fT Ch Ruts - teddy dt The parameter Ry(ty - t,) is the ensemble-averaged autocorrelation function of the white-noise process W(t). From random process theory, it is recognized that the autocorrelation function and power spectral density of a random process form a Fourier transform pair. Since the white-noise process W(t) is assumed to have a constant power spectral density of Ny/2, it follows that the autocorrelation function of such a process consists of a delta function weighted by No/2. Specifically, we may write Ry - ) = Mae - 4 +) @ Substituting Eq. (7) in (6), and using the property that the total area under the Dirac delta function &(¢ - t, + t,) is unity, we got 2 of = NoToA ® 2 2 213The statistical characterization of the correlator output is compizted by noting that it is Gaussian distributed, since the white noise at the correlator input is itself Gaussian (by assumption). In summary, we may state that under hypothesis Hy the correlator output is a Gaussian random variable with zero mean and variance NgT}A7/2, as shown by fly) = 2 exp| - =, (9) IRNoT, A NoTpA where the subscript in f,(y) signifies the condition that symbol 0 was sent. Figure 2(a) shows the bell-shaped curve for the probability density function of the correlator output, given that symbol 0 was transmitted. The probability of the receiver deciding in favor of symbol 1 is given by the area shown shaded in Fig. 2(a). The part of the y-axis covered by this area corresponds to the condition that the correlator output y is in excess of the threshold A defined by Eq, (1). Let Pyg denote the conditional probability of error, given that symbol 0 was sent. Hence, we may write Pro = f° fo) ay 1g A (10) 7 y = exp| - —>—_. |dy |mNoTy a 7? { NoTA? Define ze an We may then rewrite Eq. (10) in terms of the new variable z as 1 fe 2) Pio = exp(-2?) da (a2) vn “VAAN which may be reformulated in terms of 214complementary error function 2 fe 2 erfe(u) = — exp(-z7) dz (13) Ie 5 Accordingly, we may redefine the conditional probability of error Pg os A*T, “AN (4) 1 =f ert dapat Consider next the second kind of error that occurs when symbol 1 is sent and the receiver chooses symbol 0. Under this condition, corresponding to hypothesis H,, the correlator input consists of a rectangular pulse of amplitude A and duration Ty, plus the channel noise w(t). We may thus apply Eq. (2) to write Hy:y=A ( [A + wit)] dt (15) ‘The fixed quantity A in the integrand of E q. (15) serves to shift the correlator output from a mean value of zero volt under hypothesis Hy to a mean value of A*T,, under hypothesis H,. However, the conditional variance of the correlator output under hypothesis H, has the same value as that under hypothesis Hy. Moreover, the correlator output is Gaussian distributed as before. In summary, the correlator output under hypothesis Hy is a Gaussian random variable with mean AD, and variance NoT},2/2, as depicted in Fig. 2(b), which corresponds to those values of the correlator output less than the threshold A set at A*I\/2, From the symmetric nature of the Gaussian density function, it is clear that For * Be ae Note that this statement is only true when the a priori probabilities of binary symbols 0 and 1 are equal; this assumption was made in calculating the threshold 2. To determine the average probability of error of the PCM receiver, we note that the two possible kinds of error just considered are mutually exclusive events. Thus, with the a priori probability of transmitting a 0 equal to p, and the a priori probability of transmitting a 1 equal top ,we find 215that the average probability of error, P,, is given by + . 17) PIO” PP “ Since 7, = pyg and p,+p, = 1, Ha. (17) simplifies as Pe = Pro = Poy or A’, 1 P, = A erfe| 1 b as) 2 24 No Choose H, it 5 Nisexcedded xy —)—H] Jy ae L-»| Decision} g Otherwise, t choose Hy s(t) » Figure 1 foo! Bo 0 5a ° Ao 1 i i i ' 7 An, * Figure 2 216Problem 4.10 For unipolar RZ signaling, we have Binary symbol 1: s(t) = 4A for 0 <1< 712 and s(t) =0 for 7/2 A, as shown byP(error]0) = J” Fyr|0ay We then have P(error|0) = Define z = and so write 1 2 P y=4f -2")de (error| 1) Flim exp(-z") NoP(error|1) The average probability of error is therefore P(1)P(error|1) + P(error|0)P(0) 1 = ert 5 E,7N,) z Leste 1 [4 | pees aK, | w The average probability of error for on-off (i.e., unipolar NRZ) type of encoded signals is dosed AT) B34 Ne | Comparing this result with that of Eq. (1) for the unipolar RZ type of encoded signals, we immediately see that, for a prescribed noise spectral density No, the symbol energy in unipolar RZ. signaling has to be doubled in order to achieve the same average probability of error as in unipolar NRZ signaling. Problem 4.11 Probability of error for bipolar NRZ signal A Binary symbol 1 : s(t) Binary symbol 0: s(¢) = 0 Energy of symbol 1 = E, = A77), 219‘The absolute value of the threshold is A= Refering to Fig. 1 on the next page, we may write P(errorls exp| [* + a ie = af{5 2) «(3 | P(error|s = -A) Similarly, P(error|s = +A) P(error|s = 0) = Fehon Fa (1 [Eo «aiff The average probability of error is therefore P P(s=+A)P(error|s= +A) + P(s=0)P(error|s = 0) The conditional probability density functions of symbols 1 and 0 are given in Fig. 1:Fuly|s=0) pr _| Srols=+A) WE, Figure 1 224Problem 4.12 ‘The rectangular pulse given in Fig, P4.12 is defined by (0) = rec(WT) The Fourier transform of g(t) is given by 1/2 GA =f pa SPCP2np dt = Tsinc (fT) We thus have the Fourier-transform pair rec(t/T) # T sine (fT) ‘The magnitude spectrum |G(f/T is plotted as the solid line in Fig. 1, shown on the next page. Consider next a Nyquist pulse (raised cosine pulse with a rolloff factor of zero). The magnitude spectrum of this second pulse is a rectangular function of frequency, as shown by the dashed curve in Fig. 1. Comparing the two spectral characteristics of Fig. 1, we may say that the rectangular pulse of Fig. 4,12 provides a crude approximation to the Nyquist pulse. 2221 T 7 | oof ggeist poles Wil, zaco vallefe f j~— { 1 ost o7 ler | os oa a2 oz o4 i) -¥T ~2/r Vem hr 9 ag VE ar afr Frequna , f Figen I Spechel chancheei shies 223Problem 4.13 Since P(f) is an even real function, its inverse Fourier transform equals pit) = 2 [ PCD cost2ntt) df @ The P(f) is itself defined by Eq. (7.60) which is reproduced here in the form w o< kl<& : f @ Pe af vo Fl- If. f 2w-f, Hence, using Eq. (2) in (1): -1 fh Ne cos MED Pit) = = Gh coscanty af ae, ht one | feosenm or , [sinenty PY 4nwWt |, ew-f, : 5) PW sinlonty + 74D sinfantt - 75-5. + lw 2Wo. eee awe te ant + W2Wa | 4w Qnt — W2Wo I = Sin(2nft) sin[2xt(2W-f,)] Fewt 4nWt 1 sin(2nfyt) + sinf2n2W-f)] | sin(@nfyt) + sinl2nt(2W-f,t)] aw Ont — mWa 2nt — W2Wo 1). . 1 = 2 [sin(antyt) + sintam2w-fJ] | - ___™ __. Milman »)] ; * =| (ant)? 224223 = 7 fein(enwe)eos(2naw)) |——~ @2Wo? __ w Ant [(2nt)? - (W/2Wa)* i 1 = sinc(2Wt) cos(2naWt) | —__-__ 16 o?W? al Problem 4.14 The minimum bandwidth, Br, is equal to 1/2T, where Tis the pulse duration. For 64 quantization levels, log,64 = 6 bits are required. Problem 4.15, The effect of a linear phase response in the channel is simply to introduce a constant delay + into the pulse p(t). The delay t is defined as -1/2n)times the slope of the phase response; see Eq. 2.171. 225 WyProblem 4.16 The Bandwidth B of a raised cosine pulse spectrun is 2W - where W vat, and W(1-a), Thus B= W(t+x), For a data rate of 56 kilobits per second, W = 28 kz. (a) For a= 0.25, B= 28 kHz x 1.25 = 35 kHz (b) 28 kz x 1.5 = 42 kHz (eo) B= 49 kHz (a) B= 56 kiz Problem 4.17 The use of eight amplitude levels ensures that 3 bits can be transmitted per pulse. The symbol period can be increased by a factor of 3. All four bandwidths in problem 7./2 will be reduced to 1/3 of their binary PAM values. Problem 4.18 (a) For a unity rolloff, raised cosine pulse spectrum, the bandwidth B equals 1/T, where T is the pulse length. Therefore, T in this case is 1/12kHz. Quarternary PAM ensures 2 bits per pulse, so the rate of information is 2M , 2y kxiooits per seoond, -(b) For 128 quantizing levels, 7 bits are required to transmit an amplitude. The additional bit for synchronization makes each code word 8 bits. The signal is transmitted at 2u kilobits/s, so it must be sampled at 2u_kbits/s Bvits/sanpie ~ 3 KHz. The maximum possible value for the signal's highest frequency component is 1.5 kHz, in order to avoid aliasing. 226Problem 4.19 The raised cosine pulse bandwidth B= 2W - f,, where W/= 1/2T,, For this channel, B= 75 kHz. For the given bit duration, W= 50 kiiz, Then, f,22W-B = 25 kz = 1- t,/8, 20.5 Problem 4.20 The duobinary technique has correlated digits, while the other two methods have independent digits. : 227Problem 4.2 (a) binary sequence by o Oo 1 1 oO 1 oo 1 omen ee a duobinary coder output % eee ete s ee ee output binary sequence o Oo 1 1 O 1 oo 1 (b) receiver input Sere eee een oO receiver output 1 V1 1-1 1-1-1 1 output binary sequence 0 1 oO 1 O 1 oo 1 We see that not only is the second digit in error, but also the error propagates. Problem 4.22 (a) binary sequence by o 0 1 1 0 1 0 0 1 coded sequence dj 1 ett Ont leet) O10] 0) a polar representation se duobinary coder output o, 2 2 0 0 2 0 -2 -2 0 receiver output o 0 1 41 0 41 0 0 14 (b) receiver input 20 00 2 0 ° receiver output o 1 4 41 0 1 001 In this case we see that only the second digit is in error, and there is no error Propagation. Problem 4.23 (a) The correlative coder has output Fn =n 7 Ynat ItS impulse response is otherwise. The frequency response is 228H(t) = hy exp(—J2nfkT,) z kz: = 1 = exp(-J20fT,) (>) Let the input to the differential encoder be x,, the input to the correlative coder be qs and the output of the correlative coder be z, Then, for the sequence 010001101 in its on-off form, we have Then z, has the following waveform The sequence z, is a bipolar representation of the input sequence x,+ Problem 4. (a) The output symbols of the modulo-2 adder are independent because: ie 2. the input sequence to the adder has independent symbols, and therefore knowing the previous value of the adder does not improve prediction of the present value, i.e. feyatyy 4) = £0q) > where y, is the value of the adder output at time nT,» The adder output sequence is another on-off binary wave with independent symbols. Such a wave has the power spectral density (from problem 4/0) , a 2, oes Sy(f) = Of) + sino*(rT,) . 229The correlative coder has the transfer function Hf) = 1 = exp(-J2xfT,)» Hence, the output wave has the power spectral density S7(f) mice? sy (1 = exp(-g2ee7,)] (1 - exp(jentT, I sy(0) [2 - 2 cos(2mfT,)] Sy(f) 2, 4 sin®(xtT,) SYP) 2, 2 AT, 4 sin®(net,y CA ace) + a? sino*(eT,)1 2, 2, ‘Ty sin*(xfT,) sinc“(fT,) In the last line we have used the fact that sin(afT, Oat f= 0. (b) gp Note that the bipolar wave has no de component. (Note: The power spectral density of a bipolar signal derived in part (a) assumes the use of a pulse of full duration 7, On the other hand, the result derived for a bipolar signal in part (d) of Problem 3.11 assumes the use of a pulse of half symbol duration T;,) 230Problem 4.25 The modified duobinary receiver estimate is 4, = ¢ (a) binary sequence a, bipolar representation modified duobinary 0, receiver output 8, output binary sequence (b) receiver input receiver output output binary sequence 0 -1 “1 1 1 1 1 Here we see that not only is the third digit in error, Problem 4.26 (a) binary sequence by coded sequence a, polar representation -1 -1 modified duobinary &% receiver output b, = loyl output binary sequence (>) receiver input receiver output output binary sequence ole) <1 ot 2 <1 -1 oo 2 3-1 oo but also a) lee 1-1 oo oo 0 0 oo a) oo O @ 104 o 0 14 o 4 an) the error propagates. -1 -2 This time we find that only the third digit is in error, “propagation. 231 o 2 o 2 4 o 2 o 2 o 4 and there is no errorProblem 4.27 (a) Polar Signalling (M=2) In this case, we have t mt) = EA, sinetg - n) where A, = * A/2. Digits 0 and 1 are thus represented by -A/2 and +A/2, respectively. The Fourier transform of m(t) is MCE) t 2A, Flsine(y - n)] Trect(fT) Z A, exp(-jemer) Therefore, m(t) is passed through the ideal low-pass filter with no distortion. The noise appearing at the low-pass filter output has a variance given by N ar Suppose we transmit digit 1, Then, at the sampling instant, we obtain a random variable at the input of the decision device, defined by A XeZeNn where N denotes the contribution due to noise. The decision level is 0 volts. If x > 0, the decision device chooses symbol 1, which is a correct decision. If X < 0, it chooses symbol 0, which is in error, The probability of making an error is ° PCKCO) =F fyCx) dx The expected value of X is A/2, and its variance is o°. Hence,fyi) 2 P(x<0) xo $ ertec 2v20 Similarly, if we transmit symbol 0, an error is made when X > 0, and the probability of this error is A P(x0) = F erfet 2ve0 ) Since the symbols 1 and 0 are equally probable, we find that the average probability of error is = $ P(x<0 | transmit 1) + $ PGOO | transmit 0) 2veo (b) Polar ternary signaling In this case we have t w(t) = EA, since n) The 3 digits are defined as follows Digit Level ° “A 1 0 2 +A Suppose we transmit digit 2, which, at the input of the decision device, yields the random variable Xa Aan 233The probability density function of X is 2 1 (ay £00 = xp (= ) 0 Vimo 20% The decision levels are set at -A/2 and A/2 volts. Hence, the probability of choosing digit 1 is we 2 1 expt- UA) ax -2 Yano 20° pedcx ch =} Lerte(—4A_) ~ erto( 34} ao ao Next, the probability of choosing digit 0 is If we transmit digit 1, the random variable at the input of the decision device is X=N The probability density function of X is therefore The probability of choosing digit 2 is A wee Pa > & = d ert) The probability of choosing digit 0 is px <4 1 B erfot ave o Next, suppose we transmit digit 0. Then, the random variable at the input of the decision device is Xe-kan The probability density function of X is therefore 2342 ext Gp Yano 20° £00 = The probability of choosing digit 1 is Ay ~ erzoc—3h Pe Scx< veo av2o 2 lerfe ¢ 0] The probability of choosing digit 2 is 3A) 2720 eee PK > 5) = 5 erfet Assuming that digits 0, 1, and 2 are equally probable, the average probability of error is 1 A A 3A 1 1 3A Poze erfo(: = 5 erfeol- J1+ 4° 5 lerfe(——)] 3 avBo 2 ao 3 2 2v8 o 1.1 A 1 1 A +3 °° > lerfe(——)] + = * 5 [erfet- 1 3 e vBo «3 2 28 o + 1 1 A 3A 1 1 3A gz ° 5 lerfe(——>) - erfet- J] +5 ° 5 erfel- ) ate 2v2 o ao 3 ? 2vB 6 ao 3 wo (c) Polar quaternary signaling In this case, we have Suppose we transmit digit 3, which, at the input of the decision device, yields the random variable: 235