Chapter 2 eee 2.1-1 Let us denote the signal in question by g(t) and its energy by By. For parts (a) and (b) te rc sintrdt= 3 fo ae cos 2dt=r+0=m 2), 2 ©) ee [ sinttut= 2 [at-} [costae rt0=% ae Bhp Se 2° eee (a) ef casin 1? ar = a3 ek [corte] = ae soln ae lo 2 Io 2 to Sign change and time shift do not affect the signal energy. Doubling the signal quadruples its energy. In the same way we can show that the energy of kg(t) is k*Ey, 2ae2 (a) Kes [Zar 2. Ey fe(uae+ f2(-1rae = 2 (2)?dt=4. Ex [rans “2 5 0/2 f pas f tora f (ytare o ia m2 a0/2 a Eey= [tars [ores [Cota ee 0 f f Therefore Eazy = Ex ~ Ey (bE, = [ors f 7 Similarly. we can show that Ezy = 44 Therefore Exsy = E. + Ey. We are tempted to conclude that Ext = E,—£y in general. Let us see. o ee [Corns [cntaer ee [ataes seve [ores [omer torn [ome [cote ‘Therefore. in general Exsy # Ex + Ey Lf cheatin snare Sf ec reed [Meteor ona & [econ sana "hh I, dt+ [, cou(aaet +204] = Fete +0) F 2.1+4 This problem is identical to Example 2.26. except that 1 # w2. In this ease, the third integral in Py (see p. 19 is not zero. This integral is given by 2 y= gi, 2 [contort + O:}c08 ait + Ha) at at rie a a eed if onl; —@)dt +f cos(2art +01 +62) dt T Ura “ra " jinn, SY 11 eoaity ~ M9)} + O-= C.Ca costs ~ 2) 121-5 21-6 21-8 2.21 Therefore F + crcecosins - #2) 2 Py [eras cur ®) port (yar = 64/7 a aj a [ (ct Pat = 6407/7 wr=5 f (2t*yPat = 4(64/7) = 256/7 — (€) Pea = Sign change of a signal does not affect its power. Multiplication of a signal by ® constant © increases the power by a factor ©? ani aaron Lf7ss dima ert Poa gin tf nine cars gin $f" SY parrot ra 119 on tam ‘The integrals of the cross:product terms (when & # r) are fnite because the integrands are periodic signals rinade up of sinusoids), ‘These terms. when divided by T — oc. yield zero. The remaining cerms (k=) yield ue * yy um $f Data = De? 71a bem fd (a) Power of a sinusoid of amplitude C is C7/2 Bq, (2.6a)] regardless of its frequency (o # 0) and phase Therefore. in this case P ~ (10)*/2 = 50. (b) Power ofa sum of sinusoid is equal to the sum of the powers of the sinusoids [Eq. (2.6b)}. Therefore, this case P = OME 4 UES = 178. (€) (10 +2 sin 31) cos 101 = 10¢os 101 = sin 13¢ — sin 3¢. Hence from Ea, (2.60) P ay p+ $= 51 (d) 10cos 5¢ cos 10t = 5(cos 5t + cos 18¢. Hence from Eq, (2.6b) @ + SF a8, (e) 10sin 5¢ cos 101 = S(sin 15/ ~ sin 8¢. Hence from Eq, (2.6b) P of 4 opt mas, fetorse 4 ==). Using the esl in Prob. 21-7, we obtain P= (1/4) 40/4) = 17 (f) 0/9! cos.wat = For a real a - cof. rp tim & tonT Jurpe For imaginary a. let a= jr. Then pea nineand [emcee ad [atm 1 ~ 2J ttre) Fig. $2.5-2 Clearly, if is real. ¢7* is neither energy not power signal. However, fa is imaginary, itis @ power signal with power 1 malt) = alt 1) rant 2). galt) salt 1) tout +1). galt) = alt — 08) + sit +05) “The signal qs(1) can be obtained by (i) delaying g(t) by 1 second (replace t with f— 1). (i) then time-expanding by a factor 2 {replace f with #2). (il) then muttiply with 1.5, Thus as(t) = 1.595 2.32 All the signals are shown in Fig. $2.3-2 2.3-3 All the signals are shown in Fig, $2.3-3 4 Ba-4> 4 9 Gs) 3 A = 2 A ° oe -6 t 3 , gars) 4 gu-d 2 4 2 o = Fig. 82.3-3 Fae [anra= [eon e, Exon [cota f° dena ty bane [writen [Pours te Baum [ietarns [Pov 2.34 F y/o bnone [ner oPaaE [doen Bue. Baio [etinieine [doi eat of Pudn ets Een = f fag(t)j? ar 24-1 Using the fact that q(7)6(r) = 9(0)4(7). we have (a) 0 (b) FA(u) Ce) FMA) (4) $A) phate +3) (1) kO(w) (use L’ Hépital’s rule) 12.2 In these problems remember that impulse A(x) is located at x = 0, Thus. an impulse 4(¢—r) is located at and so on. (a) The impulse is located at + and gf) at r =f is a(t). Therefore 32.43 251 25-2 2.83 2.4 f a(r)a(t = r)dr = alt) (b) The impulse A(z) is at 7 =O and g(t ~ 7) at 7 = 0 is g(t). Therefore [i Moae-nar=ato Lning similar arguments. we obtain (e)1 (a0 Ce) (A) 5 (we) 9-1) Ch) << Letting at = 7, we obtain (for a > 0) [stomona= Similarly for a < 0, we show that this integral is -40(0). Therefore ~ gtZyavayde = Lo(0) 1 ™~ 1 [evmanae 20 = 3 [some Therefore 1 sat) = Fr Trivial. Take the derivative of je:? with respect to « and equate it to zero. (a) In this case Ey = fi dt = 1. and [ ainrtoa =} f tar =05 = 0.8 over (0 <1 < 1), and zero outside this intervai, Also Ey (b) Thus, aff) = 0.52(¢). and the error e(t) = and Ee (the energy of the error) are E [vow f Pare1/3 and B= [lu-osra= ina lo fo lo The ervor (1 ~ 0.5) is orthogonal to r(1) because [u-onmase Note that Ey = (2£s + Es. To explain these results in terms of vector concepts we observe from Fig, 218 Ae ihe enice vector e is orthogonal to the component ¢x. Because of this orthogonality, the length-square of fg energy of 9(0) i equal to the sum of the square of the lengths of ox and ¢ (surn of the energies of -7(1) and (| In this case Ey = f2q2(Ndt = [2 fdt = afirmars “thus. 2(t) © 1.S9(f). and the error e(¢) = (1) ~ 1.59(1) = 1 1.5f over (0 <1 < 1). and zero outside this ‘Also Ey (the energy of the error) is Ex = fo (1 ~ 1.8t)?dt = 1/4 mare (a) In thin cose By = [sin 2xtat = 08, and altya(t)at = 1 tsin Qetdt = —1/e (b) Thus. git) = ~(1/n)r(0). and the ervor e(t) = t+ (1/x)sin 2xt over (0 < tS 1). and zero outside this interval. Also Ey and E¢ (the energy of the error) are 425-8 2.5-6 5 1 1 [roa [ Pat=1/3 and Ee 37 i ‘The error [t + (1/n)sin 2nt} is orthogonal to x(t) because [ sin 2nt|t + (1/x)sin 2xf}dt = 0 Note that Ep = «?Ex + Be. To explain these resulte in terms of vector concepts we observe from Fig, 2.15 that sere e vector e ls orthogonal to the component cx. Because of this orthogonality, the length of f (energy of Sin) is equal to the sum ef the square of the lengths of cx and e [sum of the energies of ex(t) and c(t) (a) If +(1) and y(t) are orthogonal. then we can show the energy of x(t) + y(t) is Ex + Fy £ reo sunrar= [~ veto [~ moras [~ sor oae [” ronne a = £ n(n dt + £ iu PRae ® ‘The last result follows from the fact that because of orthogonality, the two integrals of the cross products Tnywt) and 2° (F)y{t) are zero [ore Eq, (2.40)]. Thus the energy of 7(*) + v(t) is equal to that of 2(1) ~ vie) if (0) and 1/(1) are orthogonal. {h) Using similar argument. we can show that the energy of e17(t) +ean(t) is equal to that of et) ~ aut) if x(t) and yi!) are orthogonal. This energy is given by lea|?Ex + lea Ey. (e) If sit) = 7(0) £ u{0), then it follows from Eq, (1) in the above derivation thet Ey = Ex + Ey (Exy + Eve) 1(2.1), ga(1.2). g3(0.—2). g4(2-2). e9(2.1), and aa(3,0). From Fig, $2.6-6. we see that pairs (R-66) {e1-Bs) and (B2-s) ate orthogonal. We can verify this also analytically. aig, $2.5-6 gy-gs = (0x3) +(-2%0)=0 gree = (2x1) + (-12) = (-1x2)4(2x1)=0 We can show that the corresponding signa! pairs are also orthogonal. [stone [° crstnise ia = f atten J erie ~ zane) +210 =0 f atpostnar= J t-n(-+2estnitr +2a09)dt = 0In deriving these results, we used the fact that J™ 2?ut = f™ 23(e)dt-= 1 and J ru(t}za(t)dt = 0 2.6-1 We shall compute rn using Eq. (2.48) for each of the 4 cases. Let us first compute the energies of all the signals. [ sin?artdt = 05 E, In the same way we find E, Using Eq. (2.48), the correlat Eq, = 05. n coefficients for four cases are found as a ates sin Detsin Ant dt = 0 (2) yous [ (sin 2n1)(~sin Int) at , * @) satan [ oorsin 2rtur=0 (4) ganon [ff o.07 sin area [on areas] Laide Signals z(t) and g2(?) provide the maximum protection against noise. 28-1 Here To = 2. s0 that wo = 2n/2 mand lt) = 00+ Srancosnat +qsinnst 1S 4S 1 acy" cosnat dt = Therefore {-1)" i) be SM cose wists 25+ SL GP cont lees Figure $2.61 shows git) = 1? for all and the corresponding Fourier series representing ait) over ‘1. 1). ge) fu) a “iL t> => 4 5%t> Fig. $2.61 ‘The power of a(t) is. Pz = 0.19837. which is greater than 0.198 For N = 1. Pe Thus. N = 3. 2.8-2 Here To = 2n. so that wo = 2n/2x = 1. and 0.1111; for N = 2. Py = 0.19323, For N = a(t) = 40+ San cosnt + by where ans mere a i “saeo, eed [toners Thevfre . gen Svat ests Figure $2.82 shows a(t) = ¢ for allt and the corresponding Fourier se ar) to represent g(t) over (—, 7) eu) = yi Fig. $2.6-2 ‘The power of alt) is 1 tae nok [uted Moreover, from Parseval’s theorem (Eq, (2.90)} myer Sodan eens The power Pe is required to be 080 x 32 = 026?, Therefore a4 Lee Pz = 2.927, which is less than 0.30, For N= 6. Pe = Pee For N= 1, Pe = 2; for N= 2. Py = 25. for N 2.9625. which is greater than 0.3n?. Thus. N = 6. 2.8-3 Revall that FF a 1a) wna] foe (0 ay ond f g(t) cos nwot dt {1b} rena rant el Fle a(t) sin neat dt ue)Deed Recall also that cos nent is an even function and sin nut is an odd function oft. If g(t) is an even function of Fetch at!) cos next is also an even function and a(@)sin nut is an odd function of f- Therefore (see hint) a fr mond f attae (2) pen Oo af a(t)cos nuot dt (2) bea ro) similarly if (t) is an odd function oft, then g(t) cos nugt is an odd function of t and g(t)sin nwot is an even function of t. Therefore a = 4 =0 (3a) 4 RP b= [ a(t) sin nwotdt (ab) fon required to compute the coefficients need be performed over Observe that, because of symmetry. the integré only half the period. (a) To = 4. wo = $f = F Because of even symmetry, all sine terms are 280 att) = 404 5 ances (Ze) 0 (by inspection) Bat Dog ME Leos HE) 275 2 7 2 and we allow Ci 10 take negative values, Figure §2.8-4a shows the plot of Cu {E Because of even symmetry. all the sine terms are 2670 Here bn (b) To = 10.40 = F ait) = 204 590s (21) + basin (31) ok mle) ee dO Se) mae f “sn (Be) et = 0 (ntgrand isan odd function of #) Here by = 0, and we allow Cx to take negative values. Note that Cy = am for m= 0;1,2.3,:-~. Figure $2.8-th shows the plot of Cn (¢) To = 22. wo = 1 git) = 00+ Syanconnt + besinnt with 9 = 05 (by inspection) wae Lyme ! eact [" homnire teed [" Lanntdt=- 2 [ greomnar=o b [fannie —z and Lanne tonee! (sors Hanars Lana omer) ) + Seon (30+ Fvo “STEEL cc 12> %5 6 we 123456 © t t c oa} 5” the an ol} |e 7 (e) oy. ar om ae cf? Pig. $2.84 ‘The reason for vanishing of the cosines terms is that when 0.5 (the de component) is subtracted from 9/). the qemaining function has edd symmetry. Hence, the Fourier series would contain de and sine terms only. Figure $2.8-4c shows the plot of Cy and 4». (4) To = ¥. wo = 2 and y(t) = 4h ao=0 (by inspection). an =0 (n>0) because of odd symmetry. 4 ear 4 ye 4 1 a(t) = Apsin 9 2 singt — go sin 6t — gain t+Figure $2.8-4d shows the plot of Cy and 4p. (0) To = 3. wo = 25/3. Therefore Co = } and oe at (f) To = 6. +0 = 7/3. 00 0.5 (by inspection). Even symmetry; bn = 0. ap. ar nnd f ston than git) 208 ~ Observe that even harmonics vanish. The reason is that if the de (0.5) is subtracted from a(t), the resulting function has half-wave symmetry, (See Prob. 28-6). Figure $2.8-4f shows the plot of Cn 2.8-5 An even function get) and an odd function gg(f) have the property that ~g9e(-1) ay Gelt) = gel—t) and aol) Every signal g(t) can be exptessed as a sum of even and odd components because a(t) = 4 fate) + af—t)} +4 alt) - 9(-) wearer From the definitions in Eq. (1). it can be seen that the first component on the right-hand side is an even function. while the second component is odd. This is readily seen from the fact that replacing f by ~f in the first compotient yields the same function. The same maneuver in the second component yields the negative of that component. To find the odd and the even components of 9(t) = u(t). we have a(t) = gelt) + got) where {from Eq. (1)] wlth = $y emia d and golt) = } u(t) —n(-] = pee “The even and odd components of the signal u(t) are shown Similarly. to find the odd and the even components of ¢(!) a(t) = a(t) + volt) where alt) = 3 [et ulty + etu(~1)] ond 10(cp) Fig. $2.85 elt) = $ [om*tu(t) = eu(—1)] ‘The even and odd components of the signal ¢~*'n(t) are shown in Fig. $2.8-5b, For g(t) =". we have a gel!) taal!) where elt) = 4 [OH and antt) = 4 fet =o] isin t 2.8-6 (a) For half wave symmetry at) = -9 (1+ 8) and nt mek [anenattnn df arcanatars [” gisnate Let + = 1 - To/2 in the second integral. This gives Zl [ nrcnarars [°"o(e+) conn (e+ 3) «] m2 To) -2[f acncosnaata f -a()[-coonanrd] 28 [ [comma] Ina similar way we can show that 4 02 Bhs rsinneatat (b) () To = 8. «0 = F. 00 = 0 (by inspection). Half wave symmetry, Hence uTherefore Similarly and t= Yo ances Et + basin elas. (8) Tp = 27. wo = 1. ao = 0 (by inspection). Half wave symmetry. Hence alt)= So ancosnt + besinnt neta. (n_ odd) and 1° sin melt 2 [sept coason nena” (n oda) 2.91 (a): To = 4.wo = 4/2. Also Do = 0 (by inspection). nf el dts Zain = Inf>1 (b) To On. wo = 2e/10r = 1/5 a= 3 De. wherefF) Fig. $2.0-1 Oy a(t) = Do+ > Dae'™. where. by inspection Do = 05 . F on>o eeitar= 2, sothat [Daim gi, and 2De= 4? |, 2m an Onn sono and Dn =0 amt where Dro’2.02 (e) To = 3.40 ae Smee wns Daa bf te P ate ras [EP +) a] ant Therefore 3_| (oe hen in en oa (Bee cos a=| a4 SE — too BGP - SSP ain P| and Do = tan (AS) \D-! (f) To = 6. 20 = 7/3 Do = 05 For a compact trigonometric form. all terms must have cosine form and amplitudes must be positive For this reason, we rewrite g(f) as ait) = Seost + con (se Z~ £) +2008(8'- 5-7) meant con (se = 26) +200 Figure $2.9-2a shows amplitude and phase spectra. (b) By inspection of the trigonometric spectra in Fig. $2.9-2a, we plot the exponential spectra as shown in Fig, §2,9-2b, By inspection of exponential spectra in Fig, $2.9-2s, we obtain aye Bet eer [er sess [ost at gst] + (get) om (eo) e+ Set (Got) ome (em 142.9-3 (a) att) = 2+ Boat =m) + co(st- $) = 2 2cos 2t + sin 3t (b) The exponential spectra are shown in Fig, $2.9-3. (c) By inspection of exponential spectra auiye ae fer vn Ef HON] 224 2eont-0) 400 (st~ 5) (a) Observe that the two expreasions (trigonometric and exponential Fourier series) are equivalent Fig, 82.93 2.964 To)? nen [[ riers [Nosema Tel? 2 If ott) is even. the second term on the right-hand side is zero because its integrand is an odd function of f Hence, Da is real In contrast. if g(f) is odd. the first term on the right-hand side is zero because its integrand is an odd function of ¢. Hence, Dx is imaginary,Chapter 3 a 3.161 Ba-2 31-3 ews fr aye mann [7 gpeoatar—j f at0sinsta If ol") is an even function of. a(t)sinat ip an odd function of t, and the second integral vanishes, Morea", video ie an even function oft. and the frst integral is twice the integral over the interval Oto oe. Thus when ai!) is even ors) =2 f a(t) cost dt a) Similar argument shows that when g(t) is odd Cle) 25 J sensinatar @) If gl?) is also real (in addition to being even), the integral (1) is real. Moreover from (1) ai-w)=2 | atycosseat = Glu) Heuce Gx) is real and even function of 2. Similar arguments can be used to prove the rest of the properties f Gly d= zf- Glaieteee dhe 5 + it iacaniostr + 6terite i f (otasiniot + eho] since 'Glu){ is an oven funetion and 8,(w) is an odd function of w. the integrand in the second integial is an Since 10 (=)1 i an a therefore vanishes. Moreover the integrand in the fist integral is an even function of and therefore a= 2 [icteric + Gl) de ty (t), G(w) = phy. Therefore |G(w)| = 1/VaT a? and 6,(w) = — tan™*(¥). Hence Cotee be Ole [i awemra ohw)= f ™ pene ae = f ayer 16 For 9(!) Gi Therefore and ol8.164 (a) ) 7 7 Ge) f[ te at = f[ lo lo 3.15 (a) 1 2 1 — 908 Gu) forts f arm gt Ader a te lo : je (e) auy= fo temans [tomar Zeleoner + wrsinwr ~ 1) ‘This result could also be derived by observing that g(t) is an even function. Therefore from the result in Prob. 3-1 ow) = 2f eon stat m 2p fconur + ur sinwr — 1) 3.1-6 (a) pot \* 21? — 2) sin wot + 2wot cos. f etace LO 2 ajiat ea] = WBE a Bsns Beet cot ost = Gry 2n a G0? =r fet Fig. $3.16 (b) The derivation ean be simplified by observing that G() can be expressed as a sum of two gate functions Gylw) and Ga(w) as shown in Fig. $3.1-6. Therefore avez fiisw +Gala\ do = {fines otauh = sat 3-7 (a) me) + 5 ae Glu) Application of duality property yields 1 miity+ ee aenl-s) ma wooo) 194 Ee] ee ule (+ Tae (~~) Application of Eq. (3.28} yields [uot mae 1 + uw) 1) +) ue) (») 08 wot =o T|6(w + wo) + Mw — #0)] wo Gor Application of duality property yields [A(t + 00) + A(t — w)] => 2x cos (wou) = 2 £08 (wos!) oa “ao Setting wo = T yields A(t + T)+6(t-T) <=> 200s Tw () gin wot, = jr[Bw + wo) - dw ~ 20)} con Application of duality property yields jit + 0) — A(F ~ s0)] => 2esin(—wow) = -2n sin(wow) one en Bea) Setting wo = T yields 19Olt + T) — (t= 7) ee Bi sin Tw 3.3-2 Fig. (b) o1(t) = 9(-t) and Gilw) = G-w) I 4 je — 1) 1 a Pig. (c) g2(t) = o(t — 1) + oi(t ~ 1). Therefore Galo) = (Gla) + Gi (w]e = [Glw) + G(-w)}e" 2 (coo + wins ~ 1) Fig. (4) an(t) = it~ 1) + al +1) Gao) = Glu)" + G(-w)e™” 1 4 2 “ = Fle - 2eos.e} = Soin? 5 = sine? (3) Fig. (e) ga(t) = o(t - 3) + oult + $), and Gale) = Glw)e™? + (wer? Fig. (1) ostt) can be obtained in three stepe: (i) time-expanding g(t) by a factor 2 (i) then delaying it by 2 ‘seconds. (il) and multiplying it by 1.5 [we may interchange the sequence for steps (i) and (li), The first sten (time-expansion by a factor 2) yields ‘ oye de® 20-4) £(5) 26 (2a) = Fag — JD = 1) Second step of time delay of 2 secs. yields = jdwe!™ = 1) ‘The third step of multiplying the resulting signal by 1.5 yields sty astt) = 155 (43?) os G0 te ae? a Sane) Ue? cost uit) “cos (tL) ute) = sintuin @> qb Fig. $3.33 3.33 (a) syn (2522) ~ mn (3.4 and Glu) = Tsine (2) [oot eT 2 arsine (22) sin ~#o (4) (b) From Fig. $3.3-3b we verify that a(t) = sint u(t) + sin(t ~ x)u(t - ") Note that sin(t —x)u(t ~ =) is sin f(t) delayed by m. Now. sintu(t) ome $[A(w— 1) — (s+ 1)] + gear and sin(t = n)u(t = 2) = {Flo -1)- bw + D)+ poole Therefore Glo) = LZ lMe = 1) ~ Ble +D)+ deer) T-3 Recall that g(r)#(1 x0) = a(ro)A(z ~ 70). Therefore A(w + 1)(1 +677") = 0. and (c) From Fig. $3.3-3¢ we verify that att) = cost [ut ~ 0 (r= §)] = cosrmte ~ cost But sinit - $) =~ cos? Therefore att) = cosru(t) +sin (rE) u (+ Gla) = Fite - $Me D+ ic Also because g(r)4(7 ~ 70) = alzo)A(z — 70). Blast Ie I? 2 alot eI? = 2 j0(4 1) Therefore (d) alt) = eo" fult) — u(t = 7) Typ TY eT From time-shifting property sy g(t T) > Glu)e Therefore gt ET) + 9(t — T) =o Glue + Gluye*T = 2G(w) cos T We can use this result to derive transforms of signals in Fig, P3.3-4 (a) Here g(t) is a gate pulse as shown in Fig, $3.3-4a. 2ain) = re (£) a aint) Also T = 3. The signal in Fig. P3.3-4a is off +3) + 9(t 3), and g(t +3) +9(t 3) > Asine(u) cond (b) Here 9(t) is a triangular pulse shown in Fig. $3.3-4b. From the Table 3.1 (pair 19) at = (5) sn? (2) Also T = 3. The signal in Fig. P3.3-4b is o(t +3) + 9(¢ ~ 3), and 41149) + (t= 8) me Bein’ (8) conse t) 1 X > t> Fig. $3.5-4 3.35. Frequency-shifting property states that alte**! exe Glo = 40) Therefore alt) sin oo Fie" 4 alten] = Flew = a0) + Glu = 20)} Time-shifting property states that a(t 7) a9 G(e)e2™7 Therefore git =T) = oft 7) e=9 Glu)? = Glee 7 = 2G lw) sin oT and wit + T) = olt - T)) => Glv)sinTw The signal in Fig, P3.3-5 is a(f +3) ~ g(t - 3) where a(t) = rect (5) ee 2sinc() Therefore a(t +3) - off — 8) eo 2j{2sine(s) sin Bu} = 4 sine(w)sin 3s 38-6 Fig. (a) The signal 9) in this case isa triangle pulse A( f,) (Fig. $33-6) multiplied by cos 10¢ att) = A (<-) cos 108 [Also from Table 3.1 (pair 19) A(&) e= © sine?(3#) From the modulation property (3.35), it follows that ct =a (ge) cor tor om § fete FSI] ie [=32}} TThe Fourier transform in this case is a real function and we need only the amplitude spectrum in this case as shown in Fig, $3.3-6a. Fig. (b) The signal g(t) here is the same as the signal in Fig. its Fourier transform is the same as in part (a) multiplied by « delayed by 2x. From time shifting property. 2%) Therefore 223.3-7 7 - (® Fig. $3.3-6 5 fant [25 ane [82220] oe The Fourier transform in this case is the same as that in part (a) multiplied by ¢~/2*~. This multiplying {xctor represents a linear phase spectrum —2rw. Thus we have an amplitude spectrum (same as in part (a); as well as linear phase spectrum 2G(s) = ~2rw as shown in Fig. $3.3-6b. the amplitude spectrum in this case as shown in Fig $3.3-6b Note: In the above solution. we first multiplied the triangle pulse A(-) by cos 101 and then delayed the result be 2m. This means the signal in Fig, (b) is expressed as A(432#) cos 10(t - 2). We could have interchanged the operation in this particular case. thet is, the triangle pulse A( 5) is first delayed be 2r and then the result is multiplied by cos 10t. In this alternate procedure. the signal in Fig, (b) is expressed as M4) cos 101 This inverchange of operation is permissible here only because the sinusoid cos 10t executes integral number of eveles in the interval 2r Because of this both the expressions are equivalent since cos 10(t ~ 2x) = cos 16" Fig. (c) In this case the signal is identical to that in Fig. b. except that the basic pulse is rect (3) instead of a triangle pulse A{zé). Now Cle 1) os oe see (gb) ae Bein) sing me eame argument a for pat (0) bein Glw) = w{sine|x(w + 10)] + sine|r(w - 10)]}e“ (a) Gta) = rect (254) + ree (2$4) Also 2sine(t) =e rect (3) Therefore a(t) = sine(t)eosat (b) Also 23Lancy 4 (2) Therefore a(t) = Ssinc?(t) cost 3.3-6 From the frequency convolution property. we obtain F(t) = LGw)+ Gl) 2 ‘The width property of convolution states that if c1(z) ca(z) = w(z), then the width of y(z) is equal to the sum of the widths of r(r) and r2(x). Hence, the width of G(u) + G(w) is twice the width of G(). Repeated application of this argument shows that the bandwidth of ¢”(t) is nB Hz (n times the bandwidth of 9(t)) 33-0 (a) ows f wean [eit arm Bi con at= Hast (F) (bo) aed Gl) = Tsine (4) [ete — @o 07/2 w inane (22) = 400 () (e) = A(t 7) ~ 26(0) + 90-7) ‘The Fourier transform of this equation yields eT ab MP aft cont] = tein? (22 jwG(w) = eT 24 21 | = ~4sin? (22) ‘Therefore 3.3-10 A basic demodulator is shown in Fig. $3.3-10a. The product of the modulated signal 9(t)coswot with 2cos of yields a(t) coswot * 208 wot = 2a(t) cos? wot = 9(t)[1 + cos 2uot] = 9(t} + 9(t) cos 2u0t ‘The product contains the desired 9(f) (whose spectrum is centered at w = 0) and the unwanted signal g(t) cos 2o! with spectrum 3{G(u-+2v0]+G(u~20o!. which is centered at +2. The two spertra are nonoverlapping because 2434-1 gltjeosa,t glt> GAO) 2goreT Ww -29500T Fig. $3.4-1 \s < wo (See Fig, $3.3-10b). We can suppress the unwanted signal by passing the product through lowpass filter as shown in Fig, $3.3-102, Gifs) = sine gp) and Gale Figure $3.4-1 shows Gi(«). G2(w). Hi(w) and Hi). Now Vale) = Gao) Ai(w) Ya(w) = Galw) Halu) ‘The spectra Yi(w) and Ys(w) are also shown in Fig. $3.4-1. Because v(t) = 1n(t)va(t), the frequency convalbtion property yields ¥(a) = Yi(u) + Ya(w). From the width property of convolution, it follows that the bandwidth of Ye) is the sum of bandwidths of ¥i(w) and Ya(w). Because the bandwidths of Yi(w) and Ya(w) are 10 kHz. > kiz. respectively. the bandwidth of ¥(w) is 15 kHz. Ho) = ie Using pair 22 (Table 3.1) and time-shifting property, we get centeP4e MO) = Jae ‘This is noncausal. Hence the filter is unrealizable. Also ™ fini (wt) ~ hk? WnlA Gl 2 [~ BY a = 00 f at Of oeT 235-3 3.54 Figure $3.5-1 Hence the filter is noncausal and therefore unrealizable. Since h(t) ia a Gaussian fonction delayed by to it looks serveur in the adjacent figure. Choosing to = 3V2K. h(0) = e~*® = 0.011 or 1.1% ofits peak value, Hence tow 3V3E is a reasonable choice to make the filter approximately realizable 2x10 -juto e108 From pair 3, Table 3.1 and time-shifting property, we get Hw)= A(t) = enters The impulse response is noncausal. and the filter is unrealizable. Figure $3.5-2 The exponential delays to 1.8% at 4 times constants. Hence to = 4/a = 4.x 10°* = 40us is @ reasonable choice to make this filter approximately realizable. From the results in Example 3.16 a 1 Los WON= yee = RC 1. Hence if w; is the frequency where the amplitude response drops to 0.95. then weni= Jats 107 Moreover. the time delay is given by (see Example 3.16) Also Hi = 0.95 => ws = 328.684 0-8 telw) = Seige tO) = i = If.ay is the frequency where the time delay drops to 0.98% of its value at «= 0. then 108 -« = 7. taw2) = qe = 0.98 x 1078 => wy = 142,857 We select the amaller of w and w2. that is w = 142,857, where both the specifications are satisfied. This vields 1a frequency of 22.7364 Hz. ‘There is a typo in this example. The time delay tolerance should be 4% instead of 1%. sc SFI 2000 centered at w = 10° represents the frequency range from 0.99 x 10° to 1.01 x 10°, Let us consider the gains and the time delays at the band edges. From Example 3.16 rr 1a tal At the edges of the band 261H (0.99 x 10° = Seeger = 10 «107%, and (0.01 « 10°)| = Jepgedigrraes = 9:90) «107 “The gain vasiation over the band is only 1.99%. Similarly, we find the time delays atthe band edges as 1410.99 x 10°) = ooerlrsrer = oor and ta(1.01 10°) = orci sre = ronbaor “The time delay variation over the band is 4%. Hence, the transmission may be considered distortionless, The Zignal is tranemitted with a gain and time delay at the center of the band, thet ig © 10°, We also find TH(0*) = 0.01 and t4(10°) = 17. Hence, if g(t) is the input. the corresponding output i= u(t) = 0.01 a(t ~ 107 git) t> Fig. 83.6-1 3.61 ¥ (a) = G(w) rect (3)° detotkomeT) x Gla) rect (72) Ib ~ sk sinsT} ‘This follows fiom the fart that ef & 1-47 when + € 1, Moreover, Glu)rect (gp) = OCH) hecause GI) is bandlimited to B He. Hence ¥ (a) = Gla) 2"? = jkG(w) sin wT oS Moreover, we can show that (see Prob. 3.3-5) Plott +7) ~ ake TH Ge Hence ut) = oft = to) + Eigtt = to 7) ~ alt= 19+ 7) Figure $3.6-1 shows 9(f) and u(t) 4.62 Recall that the transfer function of an ideal time delay of 7 seconds is ¢"/“7. Hence, the transfer function of the equalizer in Fig. P3.6-2 is Heq(w) = 90+ aye age Pet ane 2meOt Ideally, we require the equalizer to have 1 Wal laesves = TER wang gta eB gh PMB Lt ‘The equalizer in Fig. P3.6-2 approximates this expression if we select ao = 1 a1 = —a. 02 (-1)"o" 23.71 [Ufone Letting £ = dy and consequently dt = yd egg [rte Ee Ey = Va Sagas = Fave Also from pair 22 (Table 3.1) Aut Gu) £ [awe a Lerting au: = Jp and consequently da = hyde Bl are vie} El. a= Tava OVE 3.7-2 Consider a signal alt) =sinc(kt) and Glu) = Zrect (#) sine?(ki) dt = 3.763 Recall that [eater ae and f alter dt = Or) Therefore f au(tgalt) dt = af a(t) oe «| dt = [om if ate «4 dan [eut-epoaterae Interchanging the roles of g:(t) and a(t) in the above development. we can show that [atone = & f_overanaree 3.7-4 In the generalized Parseval's theorem in Prob. 37-3, if we identify ox(t) = sine (2#Bt — mz) and on(t) = sinc (2eBt - nz), then Gala) = Ayreet (qq) HR. and Gate f{ au(thar(tydt = | < 2B. and is 0 otherwise, Hence Therefore am [b= (aa) OF But rect (zp) = 1 for 28[-ncooxnar= ae [ee fem OE In evaluating the integral. we used the fact that #7" = 1 when k is an integer. 3.7-5 Application of duality property (Eq. (3.24)] to pair 3 (Table 3.1) yields weg ‘The signal energy is given by & if fare ™ ho The energy contained within the band (0 to W’ ) is Bente [ede BHM If Ew = 0.996y. then 2.3028 001 me W = 28 aie 0.366 He 3.76 14 92) ea A(w). then the output (2) = A(c)H (w). where H(w) is the lowpass filter transfer function (Fig. $3 7-6). Because this filter band Af — 0, we may express it as an impulse function of area 4nA/. Thus. [Ars ]Ae) and ¥ (w) & [4m A(w)As]O(e) = A ACO}ALIN) Here we used the property 9(7)8(r) = 9(0)4(2) {Eq. (2.23a)]. This yields ylt)= 2A(O)AS Next. because q2(1) <=> A(z). we have Atey= [O o%ner"ar ao that a= [~ Pinan es, Hence. »(t) = 2E,Af. a Rao> Tag $/2D rap = af o> al o> Fig. 83.7-6 3.81 Let glt) = ou(t) + aa(t). Then 7. Ry(r) = firm if [ault) + gatt)jion(t +7) + galt + r)}ut am T Sena 191 (7) + Regal 7) + Roaga(T) + Reass (7) where . ym Rested = gin 3 [stones a 293.82 If we let gi(t) = Cr cos(wit + 41) and ga(t) = C2cos(wat + 42). then 17a Ramtri jin # [ cxcreatnt + yer +t + fet Tm -7/2 According tothe argument uted in Example22b. the integral on the right-hand side sae, Hence Ryin(7) = 0 Sing the same argument, we have Ryyy(*) = 0. Therefore Ralt) ‘This result can be extended to a sum of any number of sinusoids as long as the frequency of each sinusoid is distinct. hence. if Rost) + Ryle) = Gecorurr + Peosuat olt) = So Cn cosinst +8) then Ralr) = eG coomvor Moreover, for go(t) = Co, Reo(T) = C3, and ranici= din, # [Tceicaten suerte ‘Thus. we can generalize the resuit as follows. If alt) =Cot Si Cacostnsot +6) then Ral) 8+. F coenor and Syl) = 2 CBA(e) + FS CHo(w — nwo) + Ole + nve)] Eat Figuve $3.8-20 shows the waveforms 1(t) and z(1~ 1) for 7 < To/2. Let T = NTs. On the scrate there are An ulate in the waveform of duration 7. The area under the product 7(t)z(t ~ 7) is N/2 times (Ba ryes shown in Fig. $3.8-2b. Therefore vn 3 [arene toa T Sra For 3 < jr] < To. there is no overlap between pulses, and R.(r) = 0. For Ty < |ri S 3B. pulses again overlap. Bee erage. only half pulses overlap. Hence. Rq(r) repeats every Tp seconds, bur only half the Pragnitude, ax shown in Fig. S28-2¢. We can express Rx(r) a8 a sum of two componen's. shown in Fig. $38.24. This, Re(r) = Ralr) + Ralr)- The PSD is the sum of the Fourier transforms of Ra(r) and Ral). Hence Suu) = B sine? (22) + 5214) where Solu) is the Fourier transform of the periodic triangle function, shown in Fig. $3.8-2d. We find the ‘exponential Fourier series for this per fiodic signal to be S Damn an Ralr)= > Dae’ ae E Using Eq. (2.80). we find Dy = zysine*( 4). Hence, according to Eq (3.41) sed= = sine? (SE) Mo —man) 30i (a Cg ie Area T Ber (>) * Me % % we Me “2% 35 GH BU 3 2 TH : eb 3 ae (eo? ¢ Rte) “bin Tit te "s RTD Sy m= = 30 se Sf BF Br Bez (dy $0) -or vat ca a a Fig. $3.8-2 Therefore sats) = Bin? (2B) + YS sine? (3) ae- me) a= z 3.8-3 112) = aby and |Ht2)!? = A @) Fnwe (be) P= wo Fak [Cuenta , aif plang 0 8.8-4 The ideal differentiator transfer function is jv. Hence. the transfer function of the entire system is 2 we me (Aq) w= Ba and WIP = + 32Chapter 4 SS ‘cos 1000 4.241 (i) For mit 1 (t) e08 10,000t = cos 1000¥ cos 10, 000t 1 = }leo=90004 + cos 10001) Be Use, Posnsel!) (ii) For m1) = 2cos 10007 + cos 2000" (t) = m(t) cos 10.0004 = {2 e0s 10001 + cos 2000!| cos 10.0001 059000! + cos 11.000! + 3 feos 8000t + cos 12, 0001 = {cos 90001 + 3 cos 80001] + [eos 1.0008 + cos 12.0004) —_——_—L ——- Te rs Lait) For 1/1) = cs 10004 €0s 30001 ) = m(t}.c0s 10,0001 = Z{cos 2000¢ + cos 4000t] cos 10,000" ¥nsesclt = }[eos8000r + cos 12.0001} + 3 feos 6000 + eas 14, 0004 enor ose + fe 0 + 4.08 18 ose ‘This information is summarized in a table below. Figure S4.2-1 shows various spectra. Modulated signal stechum wal =a “6K Fig. 94.244 33Cb 2 @) a) a § Me) 2200-5 {aco «20> THieo UsB je 2B 13-5] vee =lo0 (oo ~igce0 oO a igore (ip MQ) a) . & (6) &(e> 2 wel usb LB a ue use oO = =19,£0 8 Oo 8 igeoo =iqoe< Fig. $4.2-2 Fig. $4.23 case] Baseband frequency] DSB frequency | LSB frequency | USB frequency 000 | —.000 | i 1000 000 and 11,000 | i 1000 ‘3000 and 11,0001 9000 | ‘11,000 2000" 8000 and 12.000 8000" 12,000 2000 8000 and 12.000 ‘8000 12,000 4000 GOO and 14,0007 6000 | ~14.000~—~«| 4.2-2 The relevant plots are shown in Fig. $4.2-2 4.2-3 The relevant plots are shown in Fig. $4.2-3. 4.2-4 (a) The signal at point bis : oan) dal?) = m{t) cos* wet Sed = nt) [Scosat + Fone 34425 42-6 “The term Jin(t)coswet is the desired modulated signal. whose spectrum is centered at vz. The remaining term Yiv()}eoadex! is the unwanted term, which represents the modulated signal with carrier reavency Se: Sih shectram centered at #3.c as shown in Fig, $4.2-4, The bandpass filter centered at ve allows to pass the desired term fin(t) cose, but suppresses the unwanted term t(t)cosGuct. Hence. this eystem works as desined with the output 31n(t) cos ve. (b) Figure $4.2-4 shows the spectra at points b and c. {¢) The minimum useble value of we is 2B in order to avoid spectral folding at de rc) tn(t) C08? wet = oo [1 + cos 2uet) 1 1 Jm(r) + $m(t)con duet ‘The signal st point b consists of the baseband signal Jrn(t) and a modulated signal 3rm(‘) cos 2ucs which has @ carrier frequency 2we not the desired value we. Both the components will be suppressed by the filter. whose Center center frequency is we. Hence, this system will not do the desired job. fc) The reader may verify that the identity for cos must contains a term cosuet when n is odd. This is not true shen 1 is even, Hence. the system works for a carrier cos” wet only when n is odd. ' at @ Fig. S4.2-4 We use the ring modulator shown in Fig. 4.6 with the carrier frequency fe = 100 kHz (< = 200% x 10°), and the output bandpass filter centered at fe = 300 kHz. The output 1(t) is found in Eq. (4.7b) as tthe dfmttneon se = Emineoe a.t+ b(n Boe] The outpr ondpas fer uppenes all he terme exept the one centered at SEO Ms coreaponding 19th carrier Sit). Hence. the filter output is 4 me w(t) = 'm(t) cos B2ct ‘This is the desired output kin(?) cosizet with k = -4/3n. “The resistance of each diode is » ohms while conducting. and co when off. When the cartier cos." is positive the diodes conduct (during the entire positive half cycle). and when the carrier is negative the diodes are open (during the entite negative half cycle). Thus, during the positive half cycle, the voltage 9; 0(") appeais across iach of the secistors Zt, During the negative half cycle. the output voltage is zero. Therefore, the diodes act ss f gate in the circuit that is basically a voltage divider with a gain 27/(R +r). The output is therefore an oem) colt The period of w(t) is To = 2n/we. Hence, from Eq, (2.75)we ba femur = Pom dur Leos teete] The output co(t) is cat) = eR wpm) = Etomite [E+ 2 (con eat Seon det + feo et +-~)] (a) If we pass the output co(t) through a bandband filter (centered at w.). the filter su int) eos mcf forall n # 1, Jeaving only the modulated term s1¥fry ts a modulator. (b) The same circuit can be used as a demodulator if we use & is o(t) = m(?) coset and the output is .ppresses the signal m(t) and on(t)coswet intact. Hence, the system acts basepass filter at the output. In this case, the input In the present case. [sin@ + sin(2u 2.2. c ace w@> 40 =O CD me) Boss yoo 30D We =taK -u-g “7K fo Tk ae WK ISK ) BaD or a -ak -75 5 K a, 13k “I (> Mw) Foun’? se $ ~ Coe -woo [0 400 boo Tigeo =1¥00 Geo 100|O ‘Yeo 6O# 1400 [600 or a Bana’ = Mee (6d Figures not to seale -MDH) + Ml-uh d= 5,6” ® Toot loco Fig. $4.5-3 Mo) @ be # 0 in SSB (» ae — ie Fig. 345-5 $4.5-3a shows the spectrum of m(t) and Fig. 4.5-3b shows the corresponding DSB-SC spectrum ) Figure v1 cos 18,0807 E {hy Figure S45-3¢ shows the corresponding {c) Figure $4.5-3d shows the corresponding Ave now find the inverse Fourier wansforms of the Li fuecnency shifting property os LSB spectrum obtained by suppressing the USB spectrum USB spectrum obtained by suppressing the LSB spectrum SB and USB spectra from Table 3.1 (pair 18) and the ¥usp(1) = 1000 sine (10001) cos 90007" ecau(t} = 1000sine (1000xt) cos 11.000%t _jAM(a) gn (). the transfer function of a Hilbert transformer is 4.54 Because Mn(w) = H(w) = —Jsen(~) If we apply ma(t) at the input of the Hilbert transformer. Y(w). the spectrum of the output signal 1(t) is ¥ (e) = Ma(a)H (0) = [-JA1 2) 990 ()]I-Jsen (0)] = Ce) : This shows that the Hilbert transform of mma(t) is ~m(f)- To show that the energies of u(t) and ven(t) ate equal. we have Em = £ mir(tyat fore f Mp)? zf aro) Plisen (oI de a24.5:5 The incoming SSB signsl at the receiver is given by (Eq. (4.176)] Frsn (t) = a(t) coswet + ma(t)sinwet Let the local earrier be cos [(we + Aw)t +4]. The prodeuct of the incoming signal and the local carrier is ea(t). given by alt) = egal) c08 [le + Bus}t + 4] = 2m (t) conwet + men(t) sin wet] cos [(we + Muv)t + 0] ‘The lowpass filter suppresses the sum frequency component centered at the frequency (2we + Aw). and passes ‘only the difference frequency component centered at the frequency Aw. Hence, the filter output eo(t) is given by co(t) = r(t)cos(Aw)t + 4) - rmn(t)sin(Aw)t + 4) Observe that if both Aw and 4 are zero. the output is given by eo(t) = mit) = 0, then the output is given by as expected. If only eolt) = mit) cos(Aw)t — ma(t)sin(Ao)t This is an USB signal corresponding to a carrier frequency Aw as shown in Fig. 45-5. This spectrum is the seme as the spectrum A/{w) with each frequency component shifted by a frequency Qu. This changes the sound Sf an audio signal slightly. For voice signals, the frequency shift within 20 Ha is considered tolerable, Most US systems, however.testrict the shift to 42 Ha. (b) When only Aw = 0. the lowpass filter output is colt) = mu(t) cost — ma(t)sind shifted in phase by We now show that this is a phase distortion, where each frequency component of A/(~) amount +, The Fourier transform of this equation yields = Mw) cos = Ma(wr)sind But from Eq. (4.146) -jMlw) w>0 Mie) <0 sgn (w)M (0) { and _pmwet wo ao = (hee w . F000 elt? 7 N 7 t= 9008 J t >) 4 ae, i ea +> tot a-44 “Tom to qe IOM 10-24 Fig. $5.11 S.2-1 In this case fo = 10 MHz. mp = 1 and my = 8000 For FM : Af = kynipl2 = De 108/2n = 10° Hz. Also fe = 10", Hence, (fi)mex = 107 + 10° = 10.1 MHz and Udmn = 10" = 10° = 9.9 MHz, The carrier frequency increases linearly from 9.9 MHz to 10.1 MHz over a quarter (rising) cycle of duration a seconds. For the next « seconds. when 1»({) = 1. the carrier frequency remains at 10.1 MHz. Over the next quarter (the falling) cycle of Guration a. the carrier frequency decreases linearly from 10.1 MHz to 9.9 MHz., and over the last quarter cycle, Shen m{t) = 1. the carrier frequency remains at 9.9 MHz. This cycles repeats periodically with the period 4a seconds as shown in Fig. S5.1-1a. For PM: Af = kptnh/2x = 80x x 8000/2x = 2x 10° Hz. Also fe = 10". Hence. (f.)max = 107 + 2 x 10° = 10.2 MHz and Unig = 107 2x 10° = 98 MHz. Figure $5.1-1b shows ra(t). We conclude that the frequency remains at 102 Milly over the (rising) quarter cycle, where ri(t) = 8000. For the next a seconds, a(t) = 0. and the carrier frequency remaing at 10 MHz. Over the next a seconds, where *h(t) = ~8000. the carries frequency remains at QS MH2 Over the last quarter cycle ri(t) = 0 again, and the carrier frequency remains at 10 MHz. This cycles repeats periodically with the period 4a seconds as shown in Fig. $8.1-1 8.1-2 In this case fe= 1 MHz. ry = 1 and mn}, = 2000 For FM + (Af = tyinp/2n = 20,0003/2n = 10' He. Also Je = 1 MHz. Hence, (fiJmax = 10° + 10¢ = 1.01 MHz, and Ginn 2 108 — 10" = 0.99 MHz. ‘The carrier frequency rises linearly from 0,99 MHz to 1.01 MHs over the cycle (over the interval ~292 < ¢ < 192). Then instantaneously, the carrier frequency falls to 0.99 MHz and staris Fieing linearly to 10.07 MHz over the next eyele. The cycle repeats periodically with period 10* as shown in Fig. $5.1-2a. For PM: Here, because m(t) has jump discontinuities. we shall use a direct approach. For convenience, we select the reat as ahown in Fig, Sb.1.2. Over the interval 4-2 to AP. we can express the message signal as rm(t) = 20001. Hence. rato [aries Em] ces io + E200] cos [2"(20}81 + 1000%t] = cos [2x (10% + 500) ] ‘At the discontinuity. the amount of jump is rg = 2. Hence. the phase discontinuity is kprng = x. Therefore, the cartier frequency is constant throughout at 10° + $00 Hz, But at the points of discontinuities. there is @ 435.163 B21 5.22 8.28 5.24 t> Fig. 88.1-2 phase discontinuity of x radians as shown in Fig. $5-1-2b. In this case, we must inaintain ky < = because there in demodulation. rinuity of the amount 2. For ky > , the phase discontinuity will be higher than 2 giving rse (0 (a) gpailt) = Acos [vet + Rprn({)} = 10 cos[10, 0008 + kyr) We are given that gpai(t) = 10 €0s (13.000) with fy = 1000. Clearly. m(t) = 31 over the interval tI $ 1 (b) grsilt) = Acos [ere f m(a)éa| = 10 cos [recon 4s f ra) Therefore ty ff made = 2000 ff mtorta = 001 tierce ar fi m{o)do m= 9 In this case ky = 1000n and kp = 1. For an(t) = 2 c08 1001 + 18 c082000%¢ and tht) = ~200 sin 100¢ ~ 36,000 sin 2000" “Therefore mip = 20 and my = 36.000" + 200, Also the baseband signal bandwidth B = 2000r/2x = 1 KHz For FM f= kping/2n For PM: : Af = kpin/2n kits. ult) = 10 costuet + 0.1 sin 20001), Here, the baseband signal bandwidth B = 2000"/2x = 1000 Hz. 0,000, and Bru = 2(Af + B) = 2(20.000 + 1000) = 42 kHz 18,000 + 192 Hz. and Ben = 2(AJ + B) (18,031.83 + 1000) == 38.0636 Iso, wilt) = we + 200" cos 20001 Therefore, Av = 200 and Af = 100 He and Bes: = 2(Af + B) = 2(100 + 1000) = 2.2 kHz “euult) = 8 cos(vet + 20 sin 1000x1 + 10 sin 2000rt). Hee the baneband signal bandwidth B = 2000/2 = 1000 He. Also, &,(1) = we + 20,000% cos 1000n1 + 20, 000% cos 2000"! ‘Therefore, Aw = 20,000-+20.000n = 40,000" and Af kHz. ‘The baseband signal bandwidth For FM 0 kHz and Ben: = 2(Af-+B) = 2(20,000+1000) = 42 x 1000 = 3000 Hz Af = Mt = Weel = 15.951 kilz and Bras = 2(A/ + B) = 37.831 ki. For PM: Af= S32 = 82822 = 91.891 kliz and Br = 2(A/ + B) = 66.662 kHe ‘The baseband signal bandwidth B 5 x 1000 = 5000 Hz For FM: Af = 40e = MBE = 1 kia and Bess = 2(AJ + B) = 2(2-+ 5) = 14 ke 46For PM: To find BPM. we observe from Fig. $5.1-2 that p(t) is ensentially a sequence of sinusoidal pulses of width T = 107" seconds and of frequency f, = 1 MHz. Such a pulse and its spectrum are depicted in Figs 43.22c and d. respectively. The bandwidth of the pulse, as seen from Fig. 3.224, is 4n/T rad/s or 2/T Hz. Hence, Brn = 2 kHz. 5.2-0 (2) For FM: Af = S432 = 222980521 = 100 kz and the baseband signal bandwidth B = 299" = 1 KH Therefore Bey = XA + B) = 202 kHz For PM: Af = 2322 = 105f00 = 10 kblz and Bry = (Af + B) = 2(10+1) = 22 kllz (b) ru(t) = 2 sin 2000n1. and B= 2000n/2x = 1 kHz. Also my = 2 and m;, = 400%. For FM: Af = S22 = 22002 = 200 kHz, and Bex = Of + B) = 2(200 + 1) = 402 kHz For PM: Af = 1322 = 1054000 = 20 klte and Bras = 2(0/ + B) = 2(20-+ 1) = 42 ki () mit) = sin 4000x¢, and B = 40008/2n = 2 kHz. Also my = 1 and mp = 40007. For FM: A/= 402 = S200 = 100 kita, and Bra = 2(Af + B) = 2(100 + 2) = 204 kHz For PM: Af = “208 = 10240000 = 20 kHz and Beni = 2(AS + B) = 2(20 +2) = 44 Kitz (4) Doubling the amplitude of (1) roughly doubles the bandwidth of both FM and PM. Doubling the frequency’ of m(t) jexpanding the spectrum 4V/ (a) by a factor 2] has hardly any effect on the FM bandwidth, However. it roughly doubles the bandwidth of PM. indicating that PM spectrum is sensitive to the shape of the baseband spectium. FM spectrum is relatively insensitive to the nature of the spectrum M(w)- 5.2.7 From pair 22(Table 3.1), we obtain e"!? gm VFe7%"". The spectrum M (a) = V¥e"+7/4 is a Gaussian pulse Which decays :apidiy, its 2 dB bandwidth is 1.178 rad/s=0.187 Hz. This is an extremely stall bandwidi! compared to Af ‘Also n(t) = =21e-? 2, The spectrum of ri(t) is AI'(w) = jw (w) = iv %ere™~"/4. This spectrum also decars igin. and its bandwidth can also be assumed vo be negligible compared to S/ For FM: Af = S432 = SEs 2 9 kHz and Bry = 20f = 2x3 =6 kite For PM : To find mp. we set the derivative of s(t) = ~2te""/? equal to zero. Th yields sift) = -20° 9 4 ate PO aie ig and am’, = rin( Jp) = 0.858, and As = Sggh = Moe 0858 = 3.432 kHz and Bens = 2(Af) = 2(3.432) = 6.864 kHz 1 The block diagram of the design is shown in Fig. $5.3-1. & Af 21250 He > @) 10-865 MHe Fig. $5.3-1 a7fer 3M HE $= 96h Afe 4 TESOHE Bf2625H2 Afe 201 5.32 The block diagram of the design is shown in Fig. $5.3-2. B41 (a) pratt) = Acos furet + kpmn(t)] When this yeni(t) is passed through an ideal FM demodulator, the output is kpti(t) This signal. when passed through an ideal integrator. yields kyrn(t). Hence. FM demodulator followed by an ideal integrator acts as « PM demodulator, However. if m(?) has'a discontinuity, rn(t) = 00 at the point(s) of discontinuity. and the system will fail (b)— grnilt) = Acos | ath f tarde] When this signal gru(!) is passed through an ideal PM demodulator, the output is ky [‘m(o)da. When this Winal fs pateedtivough an idea! dilferentiator, the output is kymm(?). Hence. PM demedulator. followed by an sient itientiator acts as FM demodulator regardless of whether m(t) has jump discontinuities or ne. Figure $5.4-2 shows the waveforms at points b. ¢. d. and, The figure is self explanatory From Eq, (5.30). the Laplace transform of the phase error 6(t) is given by Os) = Ox(4) THAR) eee 2. and 2k Oe() = My + AKH(s)) The steady-state phase error [Eq. (8 33)] is ny We+AR) Hence. the incoming signal cannot be tracked. If ata 2k Hi)=2Z8, then Os) = SARE] and a ee im lt) = ling 800) = I TAR H @) ~ Ako ence. the incoming signa) can be tracked within a constant phase 2k/Aka radians. Now. if Hs) cimst then — @e(s) saa] and hs dim 46(1) = lim sc) = lis STAR Ga oreo) In this case. the incoming signal can he tracked with zero phase error. 48Chapter 6 SS G.A-1 The bandwidths of ox(t) and a(t) ate 100 kHz and 150 kHz, respectively. Therefore the Nyquist sampling rates for 91(t) is 200 kHz and for g2(1) is 300 kHe. Also 9:2(t) <=+ g:gulw) + gi(w), and from the width property of convolution the bandwidth ‘of i2(t) is twice the bandwidth of a1(?) and that of ¢z°(t) is three times the bandwidth of qa(t) (se also Prob, 4.3-10). Similarly the bandwidth of g1(t)ga(t) ia the sum of the bandwidth of 9:(t) and g2(t). Therefore the Nyquist rate for arth is 400 kHz. for 2°(t) is 900 kHz, for 91(t)9a(¢) is 500 kHa. 61-2 (a) sinc100nt) <=> 0.01rect (5352) ‘The bandwidth of this signal is 100 + rad/s or $0 Hz. The Nyquist rate is 100 Hz (samples/sec) (by sinc?(100nt) «== 0.01. (=) ‘The bandwidth of this signal is 200 x rad/s or 100 Hz. The Nyquist rate is 200 Hz (samples/sec) (©) sine (100n1) + sinc (5Oxt) <=» O.01rect 0.01 (sir) + 0.02rect (rE) ‘The bandwidth of the first term on the right-hand side is 50 Hz and the second term is 25 Hz, Clearly the bandwidth of the composite signal is the higher of the two, that is. 100 Hz, The Nyquist rate is 200 Hz ‘samples/sec). (4) sinc{100=1) + 3sinc®(60rt) ¢=> 0.01 reet( pie) + os O( rie) ‘The bandwidth of rect(ggfz) is 50 Hz and that of A(Z6-) is 60 Hz. The bandwidth of the sum is the higher of the two. that is. 60 Hz. The Nyquist sampling rate is 120 Hz, fe) sinc(S0nt) <= 0.02 rect( 85) sine(100rt) <=> 0.01 rect( xéf-) The two signals have bandwidths 25 Hz and 50 Hz respectively. The spectrum of the product of two signals is 1/2s times the convolution of their spectra, From width property of the convolution, the width of the convoluted signal is the sum of the widths of the signals convolved. Therefore. the bandwidth of sinc(S0x/)sinc(100st) is 25 + 50 = 75 Hz. The Nyquist rate is 150 Hz ‘The pulse train is a periodic signal with fundamental frequency 2B Hz. Hence, w, = 2n(2B) = 4B. The period is To = 1/2B. It is an even function of t, Hence, the Fourier series for the pulse train can be expressed as prilt) =Co+ D- Cneosnant Using Eqs. (272). we obtain ane Cone ae To dines > pie an On = On = af cue =dan(Z). & 1 - Hence. HN) = alter it) Lays So ain (2) atncos nent —te Fa wo Fa A sey! iE S31 002 vibe Tio 110 a o> = zs DTD w avscee s 031-002 61-4. For g(t) = sine®(Snt) (Fig. $6.1-42). the spectrum is G(u) = 0.2 A(x¥,) (Fig, $6.1-4b). The bandwidth of this signal is 5 Ha (10% rad/s). Consequently, the Nyquist rate is 10 Hz, that is, we must sample the signal at 8 rate ino less than 10 samples/s. The Nyquist interval is T = 1/2B = 0.1 second. Recall that the sampled signal spectrum consists of (1/T)G(w) = 42 (38) repeating periodically with a period equal to the sampling frequency J, Hz. We present this information in the following Table for three sampling tates: f, = 5 Hz (undersampling). 10 Hz (Nyquist rate). and 20 Hz (oversampling). sampling frequency J. | sampling interval T | 5 Hz 0.2 ‘4 (3%) | Undersampling 10 He oT 2A (aie) |_Nvauist Rate 20 He 0.05 4B (ak) | Oversampling In the frst case (undersampling). the sampling rate is 5 Hz (S samples/sec.), and the spectrum }G(~) rep every 5 Hz (10x rad/sec.). The successive spectra overlap, as shown in Fig. $6.1-4d, and the spectrum G(w) is een aie fron. Bc). thot is. 9(") cannot be reconstructed from its samples 9() in Fig. $6.1-4e. If the ‘Rinfed Signal is pased through an sea lowpass Ser of bandwidth 5 Hs, the ouput speetum ir (s?6.1-7 6.1-8 and the output signal is 10sinc(20rt). whichis not the desired signal sinc2(Sxt). In the second case. we use ane ahaat carping rate of 10 Hz (Fig, S6.1-4e). ‘The spectrum G(.) consists of back-to-back. nonossrePPg repetitions of #G() repeating every 10 Hz. Hence, G(s) can, bbe recovered from G(w) using an ideal lowpass Fee eee cintdit 9 Hz (Fig, 6.-4f). The output is 10sine (xt). Finally, in the last case of oversampling {sampling rate 20 Hz). the spectrum G(w) consists of nonoverlapping repetitions ‘of #G(~) (repeating every ine iene apty bend berween successive cycles (Fig. $6.1-4h). Hence. Glu) can Be recovered from: Gf) suing pel Topas fer or een & eeistical lowgaae filter (shown dotted in Fig, $6.1-4h). The output it sinc *(3#t) ‘Ths acheme is analyzed fully in Problem 3.4-1, where we found the bandwidths of v(t), jolt} andu(t) te bs 30 Tits hae’ as 248 Wis. respectively. Hence, the Nyquist rates for the three signals are ‘20°kH2, 10 kHz. and 30 KH. respectively. (a) When the input to this fiter i 8(¢) the output of the summer is a(¢)~ a(t —7)- This acts as the input (0 the integrator. And, (7). the output of the integrator is: nim [meee Ar = we) eT) (4) 86.1-68, ‘The impulse response (1) is shown in F (b) The transfer function of this circuit Hw) = rine(F) ure (2) ‘The amplivade response of the 6lter is shown in Fig. 861-66, Observe that the filter is lowpass fer of eae i 2n/T rad/s ene ‘He. ol q fied at the bh wowicth 24/7. rad/s Png cuca is a rectangula: pulse. When a sampled signals applied athe ier fe eae ra eps Cangula pulse at the output. proportional sarcRiscresponding sample value. Hence the sample gener cjircase approximation of the input as shown in Fig. 96.1-6¢ and lA) =T Ae) T te Figure $6.1-6 (ay Figure $6.1: shows the signal reconstruction fom ite samples using the fratarder hold Soon, Ket Seif tiers anal af a8 a arcane) ag AR oy aca oe le sequal Fa Erne a ag uignal consits of straight ine segments joining the sample Cops (G) he transfer fnetion of this eireut , , ' soe (2 ea) = FUwtt) = #{4 (ge) } = Taine (2) Because H(e) is positive forall. it alo represents the amplitude response, Fig. $6.1-7b shows the impalss response /i(t) = A(zfp). The corresponding amplitude response ‘H(w) and the idea] amplitude response (lowpass) Pore for signal Feonstruction s shown in Fig, S6.-7e {e)_A minimum of T secs delay is required to sche h(t) cousal (realizable). Such delay would cause the reconstructed signal in Fig. $6.1-7a to delayed by T secs. (d) When the input to the first filter is A(t), then as shown in Prob. 6. 1-4, its output is a rectangular pulse (2) When the iP own Fig 96.14a- This pute p() i applied co the input of the sgcond ens) Tet The output of the summer of the second filter is ott) ~ rt ‘= u(t) — 2u(t — T) + u(t - 27). which is applied The output of the sant hea the insegraror i te aren under p(t) ~ p(?~ 7). which as nye fo) ~ante = Tyne AMI ar me) — 2TH PIE = 2ryu(t 27) # shown in Fig. $6.1-7b. Aswome n signal a(t) that is simultaneously timelimited and bandlimited. Let g() = 0 for lel > 2x3. ‘Therefore aeneet(qige) = a(e) for B! > B. Therefore from the time-convolution property (3-48) a(t) = oft) = (2B'sine(2"B't)] = 2B'g(t)« sine(2nB"t) 3262-1 02-2 6.2.3 6.24 4 ne yo" Ft Ces Figure $6.1-7 0 for {tl > T. But g(t) is equal to convolution of g(¢) with sinc(2r B't) which Because o(t) is timelimited. 9(0 to obtain a time-limited signal from the convolution of a time-limited signal is not timelimited. It is impossi with a non-timelimited signal. (a) Since 128 = 27. we need 7 bits/character {(b)For 100,000 characters/second . we need 700 kbits/second. (a) 8 bits/character and 800 kbits/second. (a) The bandwidth is 15 kHz. The Nyquist rate is 30 kHz (b) 63536 = 2°°, so that 16 binary digits are needed to encode each sample, (c) 30000 x 16 = 480000 bits/s. (4) 44100 x 16 = 703600 bits/s. (a) The Nyquist rate is 2 4.5 x 10° = 9 MHz. The actual sampling rate = 1.2 x 9 = 108 MHz (b) 1024 = 2". so that 10 bits or binary pulses are nceded xo encode each sample (c) 108 x 108 x 10 = 108 x 10° or 108 Mbits/s. If'my is the peak sample amplitude. then antization error < (0-2)Gt2) = 2x quantization error < O)ete) = Fe Because the maximum quantization error is GY = Sy = 5, it follows that Meek me Le ae L=500 Because L should be a power of 2. we choose L = 512 2°. This requires 0 9-bit binary code per sample The Be Face 2 10D) = 2000 Fin. 20% above this rate i 2000 x 1.2 = 2400 Ha, Thus, cach signe bes 2100 samples/second. and each sample is ‘encoded by 9 bits. Therefore. each signal uses 9 x 2400 = 21.6 kbits/second. aa ec rais ae multiplexed. hence. we nee a total of § x 21.6 = 108 kBlits/second data bit. Fratiog and synchronization requires additic nal 0.5% bits, that is, 108,000 x 0.005 = 540 bits, yielding a total of 108540 bits/second, The minimum transmission bandwidth is 28484 = 54.27 KHz. Nyquist rate for ench signal is 200 Hz The sampling rate f, = 2x Nyquist rate 400 Hz ‘Total number of samples for 10 signals = 400 x 10 = 4000 samples/second. Quantization error < Sip# = % Moreover. quantization error = S¢ = 25% = 3% = 3m =e L = 400 Bassone Ls a power of 2. we select L = 812 2°. that is, 9 bits/sample ‘Therefore, the minimum bit rate = 9 x 4000 96 kbits/second. The minimum cable bandwidth is 36/2=18 kHz For a sinusoid, Sf! = 05. The SNR = 47 dB =S0119. From Ea. (6.16) ae So gm) 317(0.8) = 50119 = L = 1828 Because L is a power of 2. we select L = 256 = 2". The SNR for this value of L is Re seep = 3(256)?(0.5) = 98304 = 49.43 ¢B6.27 62-8 62-9 6.2-10 64-1 For this periodic s(t) each quarter cycle takes on the same set of amplitude values. Hence, euch avstie cycle Foi regan iea energy. “Consequentiy, we can compute the power for this signal by averaging is Sete saerrrntarter cycle. The equation ofthe frst quarter cycle ax shown in Fig, $6.27 is (0) & ‘4A/To. where Ais he peck amplitude and Te isthe period of m(t). The power or the mean squared value (enerEy averaged over 1 quarter eyele) i ape panty at Mea [yan viene, = ae = 3 The rest of the solution is identical to that of Prob. 6.26. From Eq, (6.16). SNR of 47 dB is @ ratio of 50119. is mages 2 - wy = 31%(1/3) = 50119 => L = 22387 Because L is a power of 2. we select L = 256 = 2°. The SNR for this value of L is Se mn = 3b72 1) = 9(256)%(1/3) = 65536 = 48.16 dB Ta Here jv = 100 and the SNR = 45 dB= 31,622.77. From Eq (6.18) So So Seg = anemaTT me r = 47988 Tai01? Because L is a power of 2. we select L = 512 = 2°. The SNR for this value of L is So _ 305127 _ . $e m= GST = 6002.84 = 45.67 2B (a) Nyquist rate = 2 108 Ha, The actual sampling rate is 1 x (2 « 10°) = 9 * 10° He, Moreover b= 256 and j= 255. From Eq, (6.18) So, _3? 3256)” gg54 = SE Tauri ~ Toaseys 7 OH = 88 SP (b) If we reduce the sampling rate ‘and increase the value of L so that the same number of bits/second is maintained. we can improve the SNR (because of increased L) with the same bandwidth. In pat (a). the sampling rate is 3 x 10° Hz and each sample is encoded by 8 bits (L = 256). Hence. the transmission rate is Bx 3 x 10° = 24 Mbits/second. If we reduce the sampling rate to 2.4 x 108 (20% above the Nyquist rate), then for the same transmission rate (24 Mbite/s). we can have (24 x 10°)/(2.4 x 10°) = 10 bits/sample. This results in L = 2° = 1024. Hence. the new SNR is Se SEF | 301020 99990 = $e = TaGeaTF 7 Ginzsey * 107800 = $9288 Cleary, the SNR is increased by more than 10 éB. Equation (6.23) shows that increasing n by one bit increases the SNR bw 6 GB. Hence. increase in the SNR Faiz a (from 30 to 42) can be accomplished by increasing n from 10 to 12, that is increasing by 20%. (a Fo Ea (630 pment wots 1 = EL a 7 = 00785 547B _ (0.0785)*(8500) _ 5 19 19-4 (e) Now 5B = OO ogy m2 x 10 (c) Here So = 4¢ = 0.5. and So os Ho Tax tot = 48x10 (2) For uniform distribution mye ond . awn Met woth $2 = 08 el ais that No ~ Tizx 10 (e) For on-off signaling with a bit rate 64 kHz, we need a bandwidth of 128 kHz. For a bipolar case, we need a bandwidth of 64 kHz = 2.94 x 10”Chapter 7 a 72-1 For full width rect pulse p(?) = wee) ls Plo) = % sine (#4) For polar signaling [see Eq. (7.12)] 2 s,(0) fet = 7 sinc? (2) For on-off case (see Eq. (7.186) oF snd (28) [ie28 o(0-22 Zh sine? (Sb) 14 E 4o-S But sine? (2%) 0 for ont for all n#0, and =1 for n=0. Hence, i Tee S,(0) = Bsine? For bipolar case (Eq, (7.206)] 2 [PM 2 (2 = POON i 3) So) 3 =, sinc? (2) sin? () 2 ‘The PSDs of the three cases are shown in Fig. $7.2 From these spectra, we find the bandwidths for all three cases to be Ry Hz. “The bandwidths for the three cases, when half-width pulses are used, are as follows: Polar and on-off: 2Ry Hz; bipolar: Ry Hz. Clearly, for polar and on-off cases the bandwidth is halved when full-width pulses are used. However, for the bipolar case, the bandwidth remains unchanged. The pulse shape bas only a minor influence in the in. Sy(q) determines its bandwidth Js bipolar ease because the term sin? (@ + Sy“) Pelar bipolar On of122 723 12d pied Tole Sy to) tt aT oO» From Fig. $7.2-2, itis clear that the bandwidth eZ rad/s of 2Ry Hz. ry For differential code (Fig. 7.17) orton es ‘To compute Ry, we observe that there are four possible 2-bit sequences 11, 0, O1, and 10, which are equally likely. The product 2,244 forthe fist rwo combinations is 1 and is-1 for the last two combinations. Hence, ‘Similarly, we can show that Ry =0 m>1 Hence, og ala Ts " ifn N A= in [ZOE | (a) Fig. $7.2-4 shows the duobinary pulse train (¢) for the sequence 1110001101001010. (b) To compute Ro, we observe that on the average, half the pulses have 2x = Oand the remaining half have ay = Jor -1. Hence, a UN? y Moyet = lim +2 Xo|-+ fo" ime alee to) 2 To determine Ry, we need to compute a,a,1- There are four possible equally likely sequences of two bits: 11, 10, 01, 00. Since bit 0 is encoded by no pulse (a, = 0), the product of a5, 41 = Ofor the last three of these sequences. This means on the average 2 combinations have a,a441 = 0and only a combinations 7have nonzero a)2441.. Because ofthe duobinary rule, the bit sequence 11 can only be encoded by two consecutive pulses of the same polarity (both positive or both negative). ‘This means ay anda,,, are | and 1 or ~1 and -1 respectively. In either case axes =1. Thus, these a combinations have apa, = 1. Therefore, ‘To compute Ry ina similar way, we need to observe the product ay aya. For this we need to observe all possible combinations of three bits in sequence. There are eight equaly Likely ‘combinations: 111, 101, eo oo O11, 010,001, and 000. The last six combinations have either the first andor the last b 0. Hence, aay. =0 forall these six combinations. The first two combinations are the onl¥ ones which yield nonzero ayay42. Using the duobinary rule, the first combination is encoded by three pulses of the same polarity (all positive or negative). Thus ay and ayy are | and ¥ or -1 and = respectively, yielding ‘2,244q =. Similarly because ofthe duobinary rule, the frst andthe third pulses in the second bit combination 101 are of opposite polarity yielding ayap4z,=—1. Thus on the average, a,4442 =1 for Sterns for terms, and 0 for terms, Henee, Rye im xoetea Molo Ina similar way we can show that R, =0 > 1, and from Eq, (7.10c), we obtain 2 > 1) =A pena) =A on) 2 orreroeoroer?e For half-width pulse P(¢) = rect(2¢/T). 5,(0)= Bsn? ( From Fig. $7.2-4 we observe thatthe bandwidth is approximately Ry /2 Hz 23-1 From Eq. (7.32) a ares (1+7)6000 1 ee 2 3 38132 13-3 134 13-5 13-6 137 m, Quantization enor SY = “P< 001m, => 12 100 @ Because Lis a power of 2, we select L = 128=27 (b) This requires 7 bit code per sample. Nyquist rate = 2 x 2000 = 4kHz for each signal. ‘The sampling rate f, = 125% 4000 = 5 KHZ. Eight signals require 8 x 5000 = 40,000 samples/sec. Bit rate = 40,000%7 = 280kbits/s. From Eq. (7.32) _lltr)Re _ 12% 280x107 rn 2 = 168 kHz. (a) By =2Ry = Ry = 1Skbitss. (b) By = Ry => Ry = 3kbitss. © Br ~Htg, Hence, s000= "25, = Ry = AB kbitvs. (@) By = Ry => Ry = Skits. (©) By = Ry > Ry = 3 kbits/s. (a) Comparison of P() with that in Fig. 7.12 shows that this P(a) does satisfy the Nyquist criterion with y= 2x 108 andr =, The excess bandwidth o, =x 108. (b) From Table 3.1, we find pit) = sinc? (xx 10%) From part (a), we have wp = 2108 and Ry = 10°, Hence, Ank)=1 n=0 =0 ne0 Hence P(t) satisfies Eq. (7.36). (©) the pulse transmission rate iste Ry = 10° bitss. i Inthis case Ab = 1M Hence, we can transmit data sta rate R, = 2 MHz. Also, By = 12 MHz. Hence, from Eq. (7.32) ler a 12x108 = "(2x10 =r = 02 fy = T00KH2. Aso, 5b = sok and f, = 700500 = 200 kHz. Ls Re -fe = Fi_ 7, = 500-200= Hencesr = ts OA and f =“ — f= 500-200 = 300 Ez ‘To obtain the inverse transform of P(a), we derive the dual of Eq, (3.35) as follows: s(t—7) © Glo"! and g(t +7) <2 Glw)e™ e+7)+ g(t-7)22G(a)cosTo ® Hence, ‘Now, P(a) in Eq. (7.34a) can be expressed as nei-gedaig) ata 4G) 9‘Using Pair 17 (Table 3.1) and Eq. (1) above, we obtain P(t) = Rgsine (2eRy)+ sine [pea (roags) a [es('-35) e [sie (2x Rt) +b sine (2m Rye) + sine (25 Rst~9) . a Som ot sin(2a Ryt+ #), 1 sin(2a Rye = 2eRy 2 2eRyrx 2 2 Ryle . sin(2eRyt) 1 sin(2x Ryt) _ 1 sin(2Ryt) Ry eR 2 aeRyte 2 WAR snarls aly ee ; 1 oR wool oan __ 2Ry coum Rytsing Ryt _ Ry Cos Ry 2m Ryi(1- 4921") 1-4Ree sine (x Ru) 2 o o )\-jwn 138 Ho)= Zo eels : dearth 1 — joel - -tm(s,) [e192 +e AR Jee ~tne(stg} tea Ry \2e Ry) Ry 2#Ry ates ens) ose eae] Sine Res sin(a Ryt- 2) aR) RA psn yt _ sine yt _ sine Ryl ARs eR- RRy(I- Rit) Hence, 73-9 The Nyquist interval is 7, “ah “The Nyquist samples are p (7) for m= 0, 1,2, 0. From Eq, (7.16), it follows that 0)= p(T) =1 and p(tnT) = 0 for all othern, Hence, from Eq. (6.10) with 7, = Tp. wndp-et, 2 2% Plt) sine m yt +sine [-(--z)] sin 2Ryt_ sin wRyt __sin Rot Ry RRi—e RRy(I- Ret)‘The Fourier transform of Eq. (1) above yields 1 @ 1 o aoe mt) aera - wel 2) [02m +erleh ers Ry \2eRy “aa @ jaol2Ry 2aRy 1.310 (0) No error because the sample values of the same polarities are separated by even mumber of zeros and the sample values of opposite polarities are separated by odd number of zeros. (b) The first sample value is 1 because there is no pulse before this digit. Hence the frst digit is 1. The detected sequence is 11000100110110100 7.3-10 The first sample value is 1, \dicating that the transmissions starts with a positive pulse, that is, first digit 1. ‘The duobinary rule is violated over the digits shown by underbracket. 12000 -200-20200-20220-2 Following are possible corect sample values in place of the 4 underbracket values: 22 0-2, 0r 2 2-2, rey On or2 000, These sample values represent the following 4 digit sequence: 1100, or 1000, or 0100, or 1010. Hence the 4 possible correct digit sequences are 11010010014)%313x.11100 where xyx,xsX is any of the four possible sequences 1100, 1000, 0100, or 1010. 74-1 $= 101010100000111 From example 7.2 T=(e?@p5eD°en%ep" ep! ep? ep" eD'e. }s R=(1@D@D*)T 7 =101110001101001 R= 101010100000111 = S 7.42 $= 101010100000111 Tx(19D?@ 4 @D'@D'eD" ep" ev" Re(1@ D?)T (see Fig. S742) T = 100010000000110 R= 101016100000111 = S Fig. S7.4-2 74-3 S=101010100000111 Ta(1@ De D* ep'ep’eptep’ ep! @p")s R=(1@D@D*)T (see Fig. 874-3) 7 =110111101001011 R= 1010101000111 = S 61151 16-1 16-2 (+) ~t + Rg Fig. S7.4-3 From Eq, (7.45), we obtain ey 1 03 -0.077 fo] [-0328 co l=] 02 1 03] [t]=| 107 c, | {0002 01 1 J Lo} [ons @ Bes (® Forpolarcase P; = 0(5)=287x107" (ii) For on-off case P, = O{5/2) = 0.00621 (ili) For bipolar case P, = 15Q{5/2) = 0.009315 In the following discussion, we assume A, = A, the pulse amplirude. (b) Energy of each pulse is E = 47, /2and there are Ry pulses/second for polar case and & pulses/second for on-off and bipolar case. Hence, the received powers are 2 2 foe Ay = OO asus 27 Pmat = hx BF coseasnio* 2 ses 10 (©) For on-off case: We require P(e) = 28710"? = QA, /20,). Hence, Ap! 20, =5 and A, = 100, = 0.003 2 2 Pott = £2008) a pasei0 For bipolar case: 4, P(e) =287 x10"? = 154, No) = 5075 Hence A= Ap =5078% 204 = 0.003045 and FRipolee = 4 wasixio* polar = For on-off case: P=l0%s dt | fe mars 6216-3 19-2 173 on = 107? = Ay 2 (4.752107) = 95x 103 For on-off case, haf the pulses are zero, and for half-width rectangular pulses, the transmitted power is: 2256x106 watts. ‘There is an attenuation of 30 dB, or equivalently, a ratio of 1000 during transmission. Therefore Sp = 10005, = 2256 107? watts For polar case: 6 A Ap 3 P= 10% 2 P| P2478 Ay =4.75x10 on on For polar case with half-width rectangular pulse: (45 10°) = 1128 10° watts Sy =(1000}(11.28% 106) = 11.28 <10™° wats For bipolar case: A, 4 P,=10% #159] 2 |= 2 = 4835 and A, = 4835x2107 = 9.67 x10 20,) 26, - For bipolar (or duobinary), half the pulses are zero and the receive power S, for half-width rectangular pulses is 2 4p 4 2 tp 3 6. Sas (06710 y= 233810" wants Sz = (1000)S, = 2338% 10? watts ‘Sampling rate = 2 x 4000 x 125 = 10,000 Hz. Quantization evor= ™E =0.00Im, => L=1000 Because L is a power of 2, we select L = 1024=2'°, Hence, n= 10 bits/sample. (a) Each 4-ary pulse conveys og, 4 = 2bits of information. Hence, we need = 4-ary pulses/sample, and a total of Sx 10,000 = 50,000 4-ary pulses/second. Therefore, the minimum transmission bandwidth is sagee = 25kH2. ® oro oe 00029 rast (a) Each 8-ary pulse carries log, 8 = 3 bits of information. Hence, the bandwidth is reduced by a factor of 3. (b) ‘The amplitudes of the 8 pulses used in this 8-ary scheme are+/2, #3A/2, £54 /2, and £74/2. Consider binary case using pulses + 4/2. Let the energy of each of these pulses (of amplitude + 4/2) be Ey. ‘The power of this binary case is Frau = Eos Because the pulse energy is proportional to the square of the amplitude, the energy of « pulse Mt is 47 Ey. Hence, the average energy of the 8 pulses in the 8-ary case above is 63TI 11-4 115 Ey) 2(41)" +23) of 2. Eq = ; = 21Ey Hence, Pony = Eay pulse rte = 218, x Bb TER. ‘Therefore, Poay = TPrinary (a) M=16. Each 16ary pulse conveys te information oflog2 16 4bits. Hence, we need 12090 3000 16-arypulsessecond. 000 Minimum transmission bandwidth = —— = 1500Hz. 2 (b) From Eq, (7.32), we have Ry = aor Hence, 3000 = 2 By = By = 18001. (a) For polar signaling, Rybits'econd requires a bandwidth of Ry H2, The half-width rectangular pulse of amplitude 4 has energy bh (34 3 4 8 @) ‘The energy ofa pulse is 42. ‘race te werage energy ofthe Mary pub eae Tie + 2(43)? +2(45)%+...02[2(M =D] 5] ua 2b z (2k+1)? wan “Sy Each M-ary pulse conveys the information oflogy M bits. Hence we require only Be My The power P is given by P= EyRy = -fhy- pulses/second. The power Py, is given by (me (M1)? ee Py aM ——=er——— Cis oe M 310g,” 2410p, M_—24l0gp M Each sample requires & bits (256= 2"). Hence: 24,0008 = 192,000 bits/sec. Br = 30kHz eer = 220,000) = 50,000 ry pulses. We nave available up 0 $0,000 Mary J pulsessecond to transmit 192,000 bitsisec. Hence, each pulse must crn ($2208) 384 bits.181 78-2 78-3 = choose 4 bits/pulse => M = 16is the smallest acceptable value (a) Baseband polar signal at a rate of IMbits/sec PSD of Psk rd using full width pulses has BWV = IMHz. PSK doubles the BW to 2MHz. (b) FSK can be viewed as a sum of 2 ASK signals. Each ASK signal BW = 2 MHz. ‘The first ASK signal cecupies a band fo 1 MHz, and the second ASK te fy signal occupies a band fy £1 MHz. Hence, the bandwidth is 2 MHz + 100 kHz =2.1 MHz. wana Ketooknad wna FigS7.6-1 (@) A baseband polar signal at a rate 1 Mbits/sec using Nyquist criterion pulses at r = 02 has @ pw =D go 2a = 60x10" He PSK doubles BH’ to 1.2 MHz. (b) Similar to Prob. 7.8-1 BWrsx = 0.6 MHz + 0.6 MHz+ 100 kHz. BW sx = 13 MHz Psd of Fst log, M=2 for =4. We need to transmit only 05x 10° 4-ary pulses/sec (a) BW is reduced by a factor of 2. BWpsx = 1 MHz (b) In FSK, there are four center (carrier) frequencies ten toy ty Set» fea» Ses» *04 fog, each separated by 100 kHz. a Ty te, + Since ASK signal occupies band f, + 05 MHz, the total seotte —f bandwidth is ee tome (05 MHz+0.5 MHz+100 kHz +100 kHz +100 kHz = 13 MHz. 5 Fig. S7.8-3 mute) fu= 2800 rotations fre mutt) myct) my tt)| 19-2 te2 7200 roladions Lee Fig. (b) Fig. $7.9-1 Either figure (a) or (b) yields the same result. ‘m(t) has 8400 samplesisec. ‘ma (t), ms(t), mq(#) each has 2800 ‘samples/sec. Hence, there are a total of 16,800 samples/sec. First, we combine ma(t), m3(t), and m4(¢) with a commutator speed of 700 rotations/sec. This combined signal is now multiplexed with m(¢) with a commutator speed of 2800 rtations/e°, ing the output of, 5600 samples/sec. wtt) cammutalin speed 2200 rotations sec19-3 Quantig € coer (a) Ahernate arrangement m4 64g kb/s a rt mye’ 2)-6 kels / cxrtput wT wT 1 i" 7 124-6 RbIls Mee" t mat? abekbls f |. ho mie) 26 pbs —~ (b) Sa er Fig7.9-3 67Chapter 8 Exercises @.1-1. Ifa plesiochronous network operates from Cesium beam clock which is accurate to +3 parts, in 10", how long will it take for a DS3 signal transmitted between two networks to become out’ 7 of sync if a 1/4 bit length time error results in desynchronization? “ Solution: A DS3 bit is transmitted in 11(44.736-10°) = 2.235336-10" sec. In the worst case, both network clocks will be out of synchronization by 6 parts in 1012. 2.235336-10°/(6-10) = 3922.27 sec/bit or 980.57 sec/ ¥ bit 8.1-2 For the bit stream 011100101001111011001 draw an AMI waveform. Solution: mye yyy Note that typically, for illustrative purposes, the waveform is given as LAA" 8.1-3 For the following waveforms, determine if each is a valid AMI format for a DS1 signal. If not, explain why not. a AD pee Solution: No. 16 0's violation a Solution: No. bi-polar violation © 4 4 Solution: No. 1's density violation a. pe ep Solution: Yese.1-4.a) You have received the following sequence of ESF framing pattem sequence bits ~00110010110010110... Is this a legitimate framing bit sequence in order to maintain synchronization between the T! transmitter and receiver? Yes____No___ If yes, why? If no, why not? Solution: No. The bit sequence 0011 cannot be in an ESF framing bit sequence. b) The following T1 AMI signal is received: Is this an acceptable T1 signal? Yes___No. a. If yes, explain. b. If no, explain why not (what, if any, DS1 standards are violated) and draw on the figure the AMI waveform which would be transmitted by the DSU? Solution: No. 16 0's violation. The 16 0's will be replaced by a pattern of 1's by the DsU. 8.1-5 The signal 110100000000000000001 is received by the DSU in aT1 data stream which uses a B8ZS format. Draw the output of the DSU for this signal? The first 1 is already drawn. ‘Show the bit stream which is substituted by the DSU. Solution: 8.1-6 ‘T-1 synchronization at two distant locations is controlled by separate crystal controlled ‘ecillators which differ in frequency by 12S parts per million. Ifthe terminal equipment doesn't maintain sync in how many complete D4 superframes will the faster oscillator have generated (at most) one more time slot (8-bit) than the slower oscillator ? Circle the correct answer. 0a)5 b) 10 o) 15 4) 20 ¢) None of the above -if “none”, what is the number of D4 superframes before an extra time slot is generated?, Solution: e) The faster oscillator will generate 125-10°1.544-10° = 193 bits per second more than the slower oscillator. This is one frame/sec = 24.125 time slots. Hence, a time slot difference will be generated in 1/24.125 = 0.04164498 frames or 0.0034704 superframes. 8.1.7. Two plesiochronous digital networks, A and B, utilize Cesium beam clocks accurate to 3 parts in 10", The networks are operated by independent long distance companies and are synchronized to each other by means of a UTC signal. a. If a company leases aT! line with D4 framing which is terminated at one end in network A and at the other end in network B, how often must the networks be resync’d to each other to avoid a framing bit error in the customers TI signal in the worst case? {You may assume a framing bit error occurs when the two networks are out of sync by 2 1/2 of aTI "bit time".} Solution: AT! bit time is 1/(1.544:10° ) = 6.47668:10" sec/bit. In the worst case, the two clocks would be off by 2:3 = 6 parts in 10"? or 6-10"? errored bits per bit transmitted. Hence, 6.47668-10” sec/bit / 6-10" errored bits per bit = 1,07945-10° seconds per errored bit or 5.39723-10° seconds per errored half-bit. b-UTC operates via GPS satellites which are approximately 23,000 miles above the Earth. How long, in terms of TI bits, will a correction signal take to be transmitted to the network switches? . Solution: The speed of light is approximately 186000 miles/sec. 73000x2 = 46000miles up and down. 46000/186000 = 0.247 sec 0.247x1544000 = 381850 bits 10Chapter 10 i 10.1-1 101-2 10.1.3 10.14 13413 1 22 411 (@) P(black queen) = “55° = 36 2.3 icture card) = = = = (© P(picture card) = 5 = 75 (a) P(red card) = 404 @ r-5-5 25 © Ps Bas (0) $= 4 occurs as (1,1,2)(1,2,1),2,1,1). There are total of6x6%6= 216 outcomes. Hence, P(S = 4) = ae (@) $= 9 occurs a5 (12,6), (135), (14), (153), (1,62). 21,6), 22,5) 234), 243), 252). C60, (3,1,3), 2.2.4), B.3.3), 3,4,2), (35,1), (41,4), (4.2.3) 4,3,2), 44,1), (5.1.3) 5,22), G31), (6.12), 62,1) P(S=9)= & (©) $= 15oceurs a5 (36.6), (45,6), (45,5) (54,6) 5:5) (5.6.4), (6.3.6), (64,5) (655.4) (6.3) 10 P(S=10)= 356 Note: There isa typo in this problem. The probsbility that the number appears should be Ai not. ‘ ta Skin ko 2ke3ke dk o5k46k= 2k ke ia Pi oT 2123450 We can draw 2 items out of $ in 20 ways as follows: 0:02, 0:03, 0;P1, 0:Pa, 0:01, 020s, O:Pi, O,P2, 0301, 0303, 0,0, O;Ps, POs, Pie, POs, PsPs, P01, POs, POs, PaPy. All these outcomes are equally likely with probability 1/20. @ This event Ey = 0, U0, R, U0, AU0e F U0 A U0 UA UA 02UA03U F301 U F502 UFOs 1203 Hence, PEN = 3575 (ii) This event £2 = RP,URA 201 Hence, P(E2)= 35" 9 (il) This event £ = 0402 U0403U020, U0203U030) U0302 6.3 Hence, PEs) * 35" 70 (iv) This event Ey = Ey UEs and both E> & E, are disjoint. 4.2 Hence, P(Es) = P(E2)+ (Es) = 3575 n101-5 101-6 10.1-7 101-8 Let zp, be the event thatthe first chip is oscillator and xq be the event that the first chip is PLL. Also, letxo, andx represent events that the second chip drawn is an oscillator and a PLL, respectively. Then P(1 ose and 1 PLL) = P(x, 24) + ?{=q20)) = Plz )PC=nfeos}+ Aen) 21a) (a) Using the notation in the solution of Prob. 10.1-5, we find: 3 © Profan) 3 © Aeabo)=% (a) We can have(!2) ways of getting two 1's and eight 0's in 10 digits (2)- P (two 1's and eight 0's) = 45(05)'(05)' = 45(05)' om (b) Plat least four 0's)= 1-[P(exactly one 0)]+[P(exactly two 0's)]+[P(exactly three 0s)] 10 5 pone 0) = ("?\05)"° “Tare 4 Pgwo 0's) =(!$\05)" a 120 1024 oo th it) ee 512 * 1024 * 1024)” 1024 P (three 0's) =(3}03)"* = P(at least four O's) = ( (a) Total ways of drawing 6 balls out of 49 are (2)= Ak -1aoesai6 Hence, Prob(matching all 6 numbers) = = (&) Tomatch exactly 5 number means we pick 5 of the chosen 6 numbers and the last number can be picked from the remaining 43 numbers. We can choose 5 numbers of our 6 in($) = 6 ways and can choose ‘one number out of 43 in(4?) = 43 ways. Hence, we have 43 x6combinations in which exactly 5 numbers match, Thus, P (mmatching exactly 5 numbers) = mS = 1845x1075 © ‘To match exactly 4 numbers means we pick 4 out of the chosen 6 number in(§) = 15 ways and choose 2 out of the remaining 43 numbers in(4)= 903 ways. Thus there are 15 x 903 ways of picking exactly 4 numbers out of 6 and (matching exactly 4 numbers) = oe =9.686%10*@ Similarly, we can pick three numbers exactly in($)(‘}) 20x 12341 = 246820 ways. Hence, (matching exactly 3 numbers) = 576 101-9 (a) Let f represent the system failure. ‘Then P(7)=(1-0.01)'° = 090438 P(f)=1- Pf) = 00956 (b) P(7)=099and P(s) = 001 If the probability of failure of a subsystem s, is p, then P(f)=(1- 9)" 0.99 = (I= p)"* => P= 00010045 10.1-10 If f represents the system failure and f,, and f; represent the failure ofthe upper and the lower paths, respectively, inthe system, then: « Pip)= PUfafa)= AA) =[PA PU_)=1~ P(f,) =1-(1-001)"° = 0.0956 and P(f) = (0.0956)" = 0.009143 Reliability is P(7) = 1- P(/) = 0.9908 ® P(7) = 0999 P(f)=1-0.999 = 0.001 P(f,) = ¥0.001 = 00316 P(j,)=(1- P)!° = 1-00316= P = 0.003206 10,1-11 Let P be the probability of failure of a subsystem (s, oF 52) For the system in Fig. 2: ‘The system fails if the upper and lower branches fail simultaneously. The probability of any branch not failing is (1- P\(1- P)=(1- P)?. Hence, the probability of any branch failing is —(1— Py. Clearly, Py, the probability ofthe system failure is P =[|-0-PF f)-0- 7? =e”? Pecl For the system in Fig. b: ‘We may consider this system asa cascade of two subsystems x, and x2 , where x isthe parallel combination of and s,and is the parallel combination of sp and 5). Let P(xj) be the probability of failure of Then Pylai)= Pelza)= PP “The system functions if neither x, nor 23 fails. Hence, the probability of system not failing is(1- P?)(1- P?}. Therefore, the probability of system failing is Py =i-(1- P21 P?)=27?- Pt s2P? Peel Hence the system in Fig, alas twice the probability of failure of the system in Fig. b. B10.1-12 There are (5) = 2598960 ways of getting 5 cards out of 52 cards. ‘Number of ways of drawing 5 cards of the same suit (of 13 cards) is(1?)= 1287. There are 4 suits. Hence there are 4 x 1287 ways of getting a flush. Therefore, 4x 128 2598960 P(flush) = = 19808 x10"? 10,1-13 Sum of 4 can be obtained as (1,3), (2,2) and (3,1). The two dice outcomes are independent, Let x, be the ‘outcome of the regular die and x, be the outcome of| tren die. Pam (td) * Pa (Pa )=4 xt det Payy,(22)= Pe (2)Prq(2) =2x0=0 Pryzy (3A) = Pe (3)Prq(l) = det = ¢ Th Pla)= i "5 “2 Similarly, PS) = Pa (IA) + Prag (23) + a 2)+ Pra (4!) 1 1 wdrostedeteorte ted ore 36" 6 6 2 101-14 B= ABUASB P(B) = P(A)P(BLA) + P(A") (B|4°) (ola) (Ga) (2) wan (ler aT St 26 P(A 10.1-15 (a) Two 1’s and three 0°s in a sequence of 5 digits can occur in($)= 1Oways. The probability one such sequence is P=(03)?(02)? =0.00512 ‘Since the event can occur in 10 ways, its probability is 10x 0.00512 = 0.0512 (®) Three 1°s occur with probability ($)(08)°(02)? = 02048, Four 1s occur with probability (4 )(08)*(02)' = 04096 Five 1°s occur with probability ($)(08)*(02)° =03277 Hence, the probability of at least three 1°s occuring is P= 0.2048 + 0.4096 +0327 = 09421 10.1-16 Prob(no more than 3 error)= P(no error) + P(1 error) + P(2 error) + P(3 error) (1) (J) PA (1— A)” + () RA BY (PRA A)” = (1-1002,)+1002,(1-99P,) +49507,7(1-987,) + 161700P,3(1-97P,) 4101-17 Error ean occur in 10 ways. Consider case of error over the first link (comet detection over every link) =(1~ A){1~ F2)--(1~ io) Py =1-Pe=1-(1-AXI-A)-(I- Ao) (A+ Py+.not Ro) + higher order terms] BA+At++Ro A< (0) Fy (40) = 06x01+04[1- Py (i) = 096-+032 = 038 y(t) =1-Py(0) = 062 15102-2 10.23 10.2-4 10.2-5 102-6 pallid) (= © Pyl= Ray "TOR +0 AID )_Pyy( Ott) =1~ Fyy( tlt) (note that ,(1) and P, (0) are derived in Example 10. 10) @ Plu ei)= [fda tare Loris Phe” 13 ) Prob(-1 2 yo ° y= Fig. $10.2-4 Since this is a half-wave rectifier, y assumes only postive values. So Py <0) = 0. Hence, Fy(y) = 0 (fory <0) and fy <0*) =. Hence, F,(0*) = x is a gaussian rv, with mean 4 andoy =3 Hence, ( Plxz4)= 44)-010)=05 Ble o) atxz0)= o{2=*)~1-0f4)=1-o06176= 09082 ; : © P(x2-2}=o{=2=4) «1-of2)=1-002275= 09773 Fig. S10.2-5 @ From the sketch it is obvious that x is not gaussian. However, it is a unilateral (rectified) version of GQussian PDF. Hence, we can use the expression of Gaussian r.v. with a multiplier of 2. For a gaussian rv. Lit wis Py)* eee with oy =4 (b) Hence, roxzi)=2A(y21)=2¢{4) 08m oa Rix) i) muensa)-anters2)=4d{2)-df2)]-o1ss Oo 2>. (c) Ifwe take a Gaussian random variable y , Leer Fig. $10.2-6 Py) * and rectify y (ll negative of y mulkipled by -1), the resulting variable i the desired random variable x 1610.2-7 102-8 10.2.9 The volume V under Pyy(x,y) must be unity. 1 4 : Vas(ixaezeh A=2 Px(2)= J Pg) » Buty =—x+ land the limits on yare 0 to1-x. Therefore, pata) =f" saeaf fO ° cotberwise 2t-y) syst ‘Similarly, rr-| ster (sy) 2 ae raed) 226 a2 -l a Fig. $10.2-7 View Osest Similarly, py. (42) = 0 oberwise Clearly x and y are not independent, pulxs)eme tn Padahdy) ® piled lve” Pedayase de) Similarly, Bly) =e? 2p) px = =) =P Ps seta) mi nee ate 4) (b) From results in (a), it is obvious that x and y are independent. ie) apolood espe ae aI ay ty pata er aalMy ae othe Similarly we can show tha pj) = ere ‘Therefore Bayle) = Pole) “Ga eo} hm Plots) = Pal?) ge Py(y)102-10 Kf, per aay Ke? [c. oP vayfiee 1 But [° 67 Pay = dine" and, Kale Mae = wel ¥)= Hence, nif pale)e ke” lye Re y= KYre = ge “ 3-32 Because of symmetry of pyy(x,y) with respect to x and y. nol {Ee ore aad. ot) Pry (2) = and ae 2 to) 1 {er aed RE Since Pyy(x,y) # Px(#)Py(y) ¥ and y¥ are not independent. 10.211 P, = P(A) Pe(1)+ P(A0) A (0) Ifthe optimum threshold isa, then ayer nd) Pcelt), rete) rade Re =="), (y+ Ap ©) -Apta “P ce a Malton py (erred Ot 0} ° ig. $102-11 — Se P= eyed oe p (a) (0) And a= ila 20) 1034 8=2, 023. x)= oe ete aras B1032 px(x)= see Because of even symmetry of p,(x), %=0and 32 2[Ppy(sle= 202? daetas b> aff vetd=3t=6 re waa, +5? oo} 23196 Fig. 51032 103-3 pob)= $50) 207 4y). Therefore 5° [Lo = Cleves Lope G ae” 20? dy =03990 F= Lr now (raves open PP yw £ 1 “aE? oy -( = £ (03990)? = 034080” 1-3 a 103-4 on? = [x7 og(s)er #2)? oe ema, ca 8 y _N6n=32 2 Sax x 103-5 The area of the triangle must be 1. Hence K= 3 and pa(s)= G(e+1) -isxs3 pecmse Boe ae PMaee e 4 safmometiir-vord(S-] fS-)$ = nao Folppesneny{ Ei n fi -1. sl4 3 4 3 3 2agt gy all 28 pare oy? =x? -(%) Smoiao) ig- S10.3-5 103-6 so Basra) = 30) 350" 5g 39 36? LeeLee Zoe Zee ge E-7 aS aPA j= 2+ 2(0)+2.09+4(29+5(06)+ 5 (48) Fe 636 36 3636 36 Soaps Serpe Z 00m + (120 +5g(14) 5485 ”107 Wea —bee fate? 22h. Forn odd, the integrand is an odd function of, Therefore x"=0. xan = Form even, we find from tables (GS)(r-t}o} in even xt= ° n oft 10.3-8 Let x; be the outcomes of the th die, Then, ge 14243444546 7 , ees wt arrest eso 246" 91 . 6 6 i21,2,..10 2 oF -(w)? =38 ofa -(R = 3 Ifx is a RV representing the sum, then Fak + Kyte Kio = wo(3)- 35 soteee Bos? = 1254167 1041 pe(2)= $a)+ pox -3) alt) een yexen pylo)= aut" al [ Sate) bale—a) een Fg S104 -t fal pee ae ie ates] pee 1a, on 4V20 Wie byt? 1 104-2 Pa(x) = 048(x) + 065(x—3) a nied PN and °°. 3 u> noleseee™ Me een Fig. S10.4-2 104-3 g(x) = 25(x-1)+(1- Q)4(x+1), a(n) = Pa(n-1)+(1- Q){n+1) py(y)=[0aty1)+(1-o)atv+ }*[ Paty +0 P)6+2)] =(P+Q-2PQ)Ay)+ POy-2)+(1- PYI-Q)ely +2)104-4 105-1 105-2 10.53 10.6-1 Prlz) = Px(2)* Py») ‘Taking Fourier transform of both sides, we have epee tei nlayen(eyrnie) wet? soley) ‘Taking inverse Fourier transform we get 1 P,l2)= Pare) Itis clear that = X+9 ando?=01+0} > For any reala, [a(x-%)-(y-¥)] 20, 0r a2o} +02 -2a0zy 20. Hence, the discriminant of this quadratic ina must be nonpositive, that is: 402, -40%03 <0, thatis, feeder or [ds eey Wheny=K\x+Kz Hence, ¥= KiX+ Kp oh = Kfod and ony = (-¥)lY~9) = (=H Kix * Ko =Kik-K) = Kyox. Hence, ory ay = sat het 1 Kyo? «1 if K; is positive, If Ky isnegative, on = Kyosis negative xPy Buto,, and ay are both positive. Hence, Pry =~! = P*cos0p(06 = = zh" c0s6d8=0 Similarly, ¥=0 Gy = 8) = sind = 1076 ht sin 26 p(6)d0 = [i sin20a0=0 4x0 0 andx, y are uncorrelated. But x? +y? =I. Hence, x and y are not independent. Hence, Oy =% In this case Ryy = Rap = Ray = mb = Po Rig = Ray = Rag = Ry = Roy = 0825 Fm Rig = Ray = Rog = 0562 Pm Ry = 0308 Pq, Substituting these values in Eq. (10.86) yields: a, = 11025, az =-02883, a5 = 00779 From Eq, (10.87), we obtain 2 =[1- (08250, + 05620, +030825)] Pq = 02753Fm Hence, the SNR improvement is ,, Ole om = dB. c ol | aoe 81Chapter 11 11.1-1 This is clearly a non-stationary process. For example, amplitudes of all sample functions are zero at same ' 1 ' instants (one is shown with a dotted line). Hence, the DAA statistics clearly depend ons. o t > be = = i t- ‘ 1 t= Fig. SU1.1-1 111-2. Ensemble statistics varies with 1. This can be seen by PP2OYRDPAD finding + XQ) = Acodar +6) = A coger +6) ple }en Uv tf U v4 AA. 00 -< ,coslar + dye>. This isa function of ¢. Heres he process non-sationy, Se i Fig. S11.1-2 111-3 This is clearly a non-stationary process since its statistics depend on ¢. For example, att = 0, the amplirudes of all sample functions is 6. This is not the case at other values of f. Ol tr Fig. S11.1-3 11-4 x{0) =acosar +8) ' Hi) = acoder +8) Bcos{ax +6) = cos(ar + 6) [4 apa(a}de er. Pac) = [rodor + 6)/24) [4 ada = 0 ———— = cA A a Aelita) =a? conan +B) cont, +0) = cosa + cola, + 6)” = cos{ar, +0)cooar, +0) [4, fan Fig. S111-4 2 = Scolar + Beaders +9) 83MAS nas Ma? a8 Hi) = aeosar +8) = fp Pcosar + 6) (a) deo “he eoker 0) d= 5 silor | = Forksin(vo0 +6)-sind] 100 0 ‘Using this result, we obtain: Raltysta) = a? eos{an + B)cos{ax, +6) = Safe $i) #26] +c0804-"2) a a . Fearn enlions +1,)+26]-sin26} + lr Taylan -4)) Hj=arsb © x ,0= 3 sj ba? gt PCO? omen Lott Hye a=0 wet aaa . Whetwleg =i atoaa=0 a a (© Reltota) <8? coder + Bhcoreeta +8) =} feosoctn—in)+ fot +12) +26)} + = feos “Alte, [pr code(s-12)+20K0 SOS OO8 eer = enteral -n) (@) The process is W.S.S. Fig. S11.1-8 a ~~ Fig. S11.1-7(€) The process is not ergodic. Time means of each sample function is different and is not equal to the ensemble mean, 9) P=RO=5 112-1 (j), (@, and (@) are valid PSDs. Others are not valid PSDs. PSD is always a real, non-negative and even function of@. Processes in (b), (c), (f), and (g) violate these conditions. 11.2-2 (a) Letx(t)=x, and x(t+1)= x2 Then, Guang) =x tag 2h 20, xg? 2 Ie But, yxy = Ry(rJand Hi? = x7? = Ry(0) Hence, Ra(0)2| Fe 0) Ryle) = Hoperreh Him R(t)» lin x(a #) Asr—+o, x(t) andx(t ++) become independent, s0 sim, Ryle) = HEF) = CD = FP 1123 R,(r)=0 forr= Fp snd its Fourier ransform S,(o) is bandlimited to B Hz. Hence, R,(r) isa waveform bandlimited to B Hz and according to Eq. 6.10b aie)= & my (a) se (2nBr -n). since (3) =0 for alln exceptn = 0. Rg(#) = R(0) sine (228) ands,(a)~ 20) ee( 22). Hence, x()is a white process bandlimited toB Hz. 1124 g(t) = Pring (te 1+ Pageg (“B=1)= Prag (le 1) = Payeg (1-1) But because of symmetry of 1 and 0, Pag I= Pug sl) and Pang (“hPa (ho) and B(#) = Pag (b= Pana tI) «#— t+} = 2F,,(0[ Phy (I~ Pega 4) re =2F, 0] Pall) =(1~ Pah) 24a 8)? Consider the case nJj <|x| <(n+1)%- In this case, there are at least n nodes and a possibility of (n +1) * nodes Protf(n+ pote] = £0 = -n h Prob(i nodes) = 1~ Probf(n+I)nodes] = (n+1)— == rs. The event(xz = tft = 1) can occur if there are N nodes and no state change at any node or state change at only 2 nodes or state change at only 4 nodes, etc. Hence, Pail) = Probl(n+!)nodes]Prob(state change at even number of nodes) + Prob(n nodes) Prob(State changes at even number of nodes) ‘The number of ways in which changes at K nodes out of N nodes occur is(f'), Hence, Pani) =[(6"?X05)°(04"*" +(3"Y06)*(04)" +.ld hy [(2yos%o4y" +(3Xo9)? (04) 2] (x 4) and Ry(r)=2P,,h9 (fl) 1 This yields a(ejet-i2id Het (o=0) =-0a4soaald %0 (Fig. $11.2-Ba) R(t) = LDAP aya 2) a” Since x, and xz can take only two values h and 0, there wil only be 4 ferms in the summation, out of which only one is nonzero (corresponding tox, =h, xz =/). Hence, R(t) = A? Pag (Oe 1) = P(A) Peatn, (AA) Since there are a pulses/second, pulses occupy ae fraction of time, Hence, Pay(h) = ae and Ry(t)= HPP aj (HH) = AP, (HA). Now, consider the range|7| <. Pye, (Af) is the nem Prob(xz) = A.given that x, = h, This means x; lies on one : of the impulses. Mark off an interval ofr from the edge of =e. . this impulse (see fig, $11.2-8b). If x, lies in the hatched L- interval, x3 falls on the same pulse. Hence, sh (HA) = Prob(xy lie in the batched region P, = w1-t C and a (s)=0H1-2) i at Since R,(r) is an even function of r, nua=a-lt) Inthe limit as ¢—+0, R,(r) becomes an impulse of strength a. Rx.) © Re(t)= a(t) [=o When > 2, x; and x, become independent. Hence, 3 es Pass (AA) = Peg (A) = oe R,(t)=athe=a? — |x{>0 Hence, Ry(t) = 08(r)+a? Fig. $11.2-811.2-9 In this case the autocorrelation function at r = Oremain same as in Prob 11.2-8. But for r > O whenever x("), wi #)are both nonzero, the product ()((-+1)is equal key tobe? and -W?. Hence, R,(2)=0, + #0 and y(t) = 25(r) 11.2-10 The process in this problem represents the ‘model for the thermal noise in conductors. Atypical sample function of this process is shown in Fig. S11.2-10. The signal x(*) changes abruptly in amplitude at random instants. “The average number of changes or shifts in amplitudes are per second, and the number of changes are Poisson-distributed. ‘The amplitude after a shift is independent of the amplitude prior to the shift. ‘The first-order probability density of the process is p(x:t). It can be shown that this process is stationary of order 2. Hence, p(x;4) can be expressed as, pz). We have Rg(#) = JL 2F%2Paye (F182 Oe = (E [osmae, (drag als = ee o To calculate py, (2/1 ‘= x), we observe that in r seconds (interval between x1 and x2), there are two ‘ther there may be no amplitude shift(x2 = x1), oF there may be an amplitude shift(x2 # x1). We can therefore express Py (2/1 = ¥1) 85 Pa (nhs = #1) = Pry (#2 = #1» ne amplitude shift) P(no amplitude shift) + Pag (a}k1 = 21» amplitude shift) P(amplitude shift) mutually exclusive possibilities t t+T ” *2 Fig. S11.2-10 ‘The number of amplitude shifts are given to have Poisson distribution. The probability of k shifts jin r seconds is given by . rity Ae where there are on the average sifs per second. The probability of no shifts is obviously Po(*) where pols) 26° ‘The probability of amplitude shift = 1— po(t)=1-e7F*. Hence Pry (22)81 = 21) =F py, (xahxy = #1» no amplitude shift) +(1-¢7*) pg, (xalei = *1- amplitude shift) @ whe there is no shift, xq = x,and the probability density fx is concentrated atthe single valve 21. ‘This is obviously an impulse located atx = x. Thus, Pry (#2/x1 = 21» no amplitude shift) = 8(2) -35) @) whenever there are one or more shifts involved, in general, x2 # xy. Moreover, we are given thatthe amplitudes before and after a shift are independent. Hence, Pr (#alei = #1» amplitude shi) = Ps, (#2) = 2) o 89M31 13-2 133 where py, (2) isthe first-order probability density of the process. This is obviously ‘p(x). Substituting Eqs. (3) and (4) in Eq. (2), we get Pr (tabs #31) Pla 21) +(1-€P* Pag (22) =e [oter-ads(e laa] ‘Substituting this equation in Eq. (1), we get ale) = oP [aura (a eCe2 21) + (eT (a) ce P(E amine (aan addin of [aval tn dr (salen | - oA [listen (eile +(e") Px (21)461 [2272 (ra) ] - oP o(ef -] where X and x? are the mean and the mean-square value of the process. For a thermal noise %=0 and Eq. (5) becomes. (6) R(t)=xteFF 0 Since autocorrelation is an even function of r, we have Ry(s) = xe and 2px? Peat S,(o) = For any rel number, (ar-y)* 20 atx? +y? -2axy20 ‘Therefore the discriminant of the quadratic in a must be non-positive. Hence, (2%)? <4 yo (i)? Ftz> fe center freq: 105 ceriley freq: 95 R Center freq: 90K Fig. $11.52 153 HG) = =a) «ef 10°? 101? } 00 Fig. S11.5-3, Sm(o) 154 ©) Hol0)= ST ye5,(0) ” 9 46 607 +60 w +10 6 " 6 ao a ® heey (©) The time constant ae Hence, a reasonable value of time-delay required to make this filter 0.949 sec. realizable is @) ‘Noise power at the output of the filter is 1 Sin(@)Sn(@) 1 6 6 -@{®_ 3 Noah fe Sa ee) tw = Lf odo = tan’ c 34 )-©S,(0)+5q(0) aeleahawe OMT ViOln VIO ‘The signal power at the output and the input are identical 1 6 =S,=—[ —;do=1 S*So" sles ge snn~ $0.08 «054 93115-5 (@) fo) heplt)= Sati) 81686 24 (© The time constant of the filter is 0.306 sec. ‘A reasonable value of time-delay required to make this filter realizable is 3 x 0306 = 0.918 sec. (@) Noise power at the output of the filter is, 1 Sn(@)Sn(@) 1 32 No 25 )@Sq(o)+Sa(a) 2 Seo ‘The signal power is sesh fi ctydont hog fea = Fig. S11.5-5Chapter 12 5, 5 8 pe 1 Saye Sh, W=2xS,()=2* ae 11d redp we? Sy(a) #210, B= = 4000H2, No y=1000= $3 5, = 008 2x10 «4000 Also, He(o)=10°3. Hence, S=—SL = 8x10 \H.()f Sr= frp [2xs000r]=8x10* = A=10 12.12 Sm? 33 so mcden le) “a a Oe Fig. $12.1-2 Sp(0) = Sq(0|Halo = 10 1 pa, of o2 +a 107 (o> a YO _ 32 97 22 f710-1| 22 iw = os Baa Norah ( a yteel =F 35 qB=3162 = 52. = Se — = 5, #33710 NM 2x107 3 3 But so(t) =m). Hence, 108 3 3 =m) ° Sy = D5 m? (0) = 337x107 = me (1) = 2157 x10 @ ) 1 h Lye, (108 s =1f salo}Heto do . zh 2696 «10° (3 “sh |" = 2696 o met fa 9 Aso, m= [des = FB = 80008 = 215.710 Hence, B= 2696x106 and Sq() = 2696106 rea{ 2 1 26% pan) 2 ax Sr -if Smlao)do = 2.6810 <, 3 [2 n6s6x10fdo = MOSES « 6865x107 S, S, 5 — 12.2.1 30 dB=1000= S2 ay = Sb =) — > 5, = 4x10" 2.2 (a) 304B = 1000" 5 = pig 28 (b) From Eq, (12.7), No = WB = 107194000) = 4 x 107 (© S,=|He(oY? Sp and 10°* Sp = 4x10 = Sp = 4xi0! 955, 5, 4 12.2-2 (@) $en100= = nao 7 = 4x10" (b) Ny = lB = 107" x8000= 4107 (©) 5, =|He(o)? Sp =108Sp = 4x10 = Sp = 4 x10" 12.2-3 Let the signals m,() and m3(1) be transmitted over the same ‘band by carriers of the same frequency (q,),but in phase quadrature. The two transmitted signals are J2[m,() cos ¢t +mp(1)sin ot] Ya ces Ot Fig. $12.2-3 ‘The bandpass noise over the channel is n¢(1)cos@et + 1n,(t)sina,t. Hence, the received signal is [vim (0) +0, (Joost + [v2ma()+ a,(i)}sinot Eliminating the high frequency terms, we get the output ofthe upper lowpass filter as m(?)+ e n,(t) Similarly, the output of the lower demodulator is m(¢)+ eu ‘These are similar to the outputs obtained for DSB-SC on page $35. Hence, we nave Se = 7 for both QAM channels. mn) 224 @ a = 72 Hence, mp = 1A s. m m 2 ) 52 = ey ee = , No Atom? omy eae (©) For tone modulation x? = mb oy and for w=, $2 acbyel Qer 3 mi 2 24m? m,z+m? Mm, (@ Ratio SE = ASP = MPT = Er stan ifx? >> + om mi? m12.2.5 12.2.6 12.27 2 (4) From Prob. 1224, 52-= Gem)’ oy 2 mi 37. For3o— loading, m,=3oqandx? “ae No xt +h ‘m S, 1 xr dwhen pel, S22——yah and when y= 9417 10 a ee ‘Whe =05, ©) When = 05 WES (ase 36 For tone modulation, let m(¢) = 4A cos@,f . For BSB-SC, nsa(t) = V2uAcosa pt Cosmet ~ Filesto. ton) +sndo.-on)] pad ae 24 A’ + ee FE ad my Bee Buk Hence, the peak power S= 1 a 2 S, S, 5, Sp =(V2ud) =24?4? and we sb ae where S, = For SSB-SC ssa!) = mi)eoser,s + m()sin art = ACOSO! CON et + HASIN yt Sin Wet = WACOM D -O_)t 2g 242 # 2 ana 5 SBA _ Sp s,=4 and m, = uA. Hence, Sp = Sonye Le ee = and mp = ence, Sp = PA? and Sm Y= ig 2B ~ 2B For AM aut) = A(l+ 120809!) cost m 242 2 2 2 im, = A(l+ aand Sp = A*(I+ 4)”. Hence, Eee ee pe ce arae No am A 4( af) AB (2407) a+ n)P ae Unde best coniton i, for a1, $= SE Ny 108 Hence, fora given peak power (given S, ) DSB-SC has 6dB superiority, and SSB-SC has S4B superiority cover AM. These results are derived for tne modulation and for = 1(the case most favorable for AM). For4c loading, mp = 4c, and the carrier amplitude A = mp = 40m (for = 1). For Gaussian m(t), im so2, (assumingii=0) Prob(E 2 A= fener tthe, = Pes! =001 on 2 2 at sence, ne Tpaasasand 5 =A = oat eh 17.2 20% oh, 2 2 2 $17 oh _11(40m) 17 Therefore, 25217. 2m 2 11/ $om | 217 (4608) = 9.7908 ae! AB 2 UB { 3 ) 9712.34 se = 28dB = 631. Hence, 5, mC) ) Soe Hele =3py ne lo mi, oh, =30)'7—B, (on) Therefore, 7 = oes. = 47325 (o) Ato, = 5525, = dB = 473.252 107! «15000 = 14197 «107 fe he aH (30m) . ©) b*F5 "te? Sonor = k som = 20,0008 5, = a2k}m?(0) = a?k}o2 = (10%) ? (20,000)? = 4x? S, -See © No = ze = 00199 123-2 and bandwidth= 2. Hence, (SolNe)py _ (28%3/To)” B® QaxV TF 3 4 GoNo)em 48/7)" 12.3.3 nt) = 08? wot and m, =1 nt) = 30, cos? @ptsinwot and #ft) = -30,[0. cos? gt cosargt ~20, 60S of si wet] & For a maximum Xt) = 0. This yields cos? wot = 2sin? argt or Las? gt =2sin? @gt = sin! = Jp, COS a@ot and im, =|+30, 005? wot sinwt| = x2 (%) - $ (Se/Nopy Gao)m ._ 90? 995 So/Mo) rw 3m {gor 30 123-4 mit) =a, cost +42 cos@2t, mp = aj +02 mkt) = -(ay0 sinayt +4202 sin. mt), mg = QO, +0202 (SolMedpy. 2x8) mB o¥(art an)” (So/No)em ore Yayo, +0,02)° 2 ofl ioSh 2 = 2 of), 211 so}aj| 14 EL ial as) 98123-5 123-7 Qax? “yea Error in this problem. There should be 47 in the denominator (see below). Sy,(@) = @7Sq(w). Hence, (edoPar [salam er= Far s?sn ene From Eq. (12.42) gu LS Salar ~ nil mT Tome lat 4x? f° mP(i)ee ‘These results are true for a waveform mt) mie) Sar —L_. 2 ano. alee Ota win! Lae =fo resi . 3) ‘The definite integrals are found from integral tables. oe Ba pat) Ask-oe, Bal = fe sin(3x/2k) seo) Berne, Baal Settdann Hence, the nomalized PSD's is OL W=3030, B= 04820 enant fa) AW) =095 = W =245e, B=039o Ae) AW) AX) 2095 W=2150, B= 03420 2) 2 = 2099 ME = 3060? > W? = PM superior 9x=095 Ya 2000? «WF 2 = x=09 Spx 540? < W? = FM superior 123-8 nko) ayconany + epcasants and Sal) =F a ar peared [as -f2)+A1F + A)] m ao Laat (ete) BB (Jo ?sairie/m)= zalte ag\.2 ide "i ast sft), a} +a} aj +a} Since B= fy,PM is superior to FM if f? > gle UVAP+Y 1 opituesty? y=2425 7 2 1(rm-20 (e) Seeger (2) my = 10,000 oF (74-20) =——— 3\- 20 sequet weeny 222 49011 123-10 From Eq, (12.40) +4 eT) -+( mi (1) Tone modulation P ales (2) Gaussian with 3o~ loading Bes a 3) Gaussian with 4o-— loading ple z os where For tone modulation, For Gaussian modulation with 30 — loading, 100wo 1 For Gaussian modulation with 40 loading, > =—"—y = 75 m, (40) 16 123-11 Lets first analyze the L*R channel. In this case, the demodulator output signal, when passed through the 2 la’ ae Eq. (12.33)}. 0-15 keiz (lowpass) filter, is given by(L-+ R)' +Mo(t), whereSy, | ‘When this signal is passed through the de-emphasis filter Hg(o) = Jorn the signal is restored to(L+R) 1 and the output noise power Nis given by 1 lm 2, Ww wen fl ales, (oko Let us now consider the (L-R) channel. Letar, = 2x x38,000 and, = 2x x2100. ‘The received signal is FM demodulated (Fig. 5.19c). The PSD of the noise atthe output of the FM demodulator is S,, (a) = Wo? / A? [see Eq. (12.33)] ‘The output of the FM demodulator is separated into (+ R)' over 0-15 kHz and (L~ R)' cosaet over the band 38 15 or 23 kHz 10 53 kHz. Let us consider the signal over this passband, where the noise can be expressed as n(t}cosa n= logy L " 0 rm Sept 7 oof So. 255 4B = 316200 No For: as distribution ale Pam amy om 1 =3" mn (a) 316200 -0"(3) Fig, $12.42 ™ =07"(2) = 2n= 1827 esr as ee choosen = 10 and L = 1024 ) 2. 32)" aa 1.048576 x 108 = 60.17 4B. Bron = 2nB = 90: a (assuming bipolar signaling) (e) To increase the SNR by 6 dB, increase m byl, that ism =11 ‘Then the new bandwidth of transmission is 22x45 = 99 MHz. S,=2BnE,, E,=2x10%, B=4000, n=8 S, =2 x 4000x8%2 x10" = 128 Si 128 y= S28. = 2561? uB” 2%625%10™7 «4000 di) = QN32 = 756910 = nB = 8x 8000 = 64 kHz (assuming bipolar line code) © ).— =} Me yeaah) * at 7 So, 52. Ne Tea wae 5) =2ies 43448. (b) If power is reduced by 10 4B, then y = 256, 32) = (1.79) = 0.0367 and Soe — (e227 = 35643, No 1+4(2'6-1}o{v32)\9 10212.44 12.45 128-1 ‘The table below gives SNR for various values of n under the reduced power. @ San ra Teas poorest rear aaa] Hence, n=3 yields the optimum SNR. The bandwidth in this case is By = 3 x8000 = 24 kHz. 1= Pg = P (correct detection overall K links) + smaller order terms (I~ PMI Pe) =[1- (KI) JN- Be] t~ Pe-(K- So Pe = Pi+(K-1)F () 7=25dB=3162, y =23dB=1995 P, = Oy 3162/8) = (6.287) = 16107" Pj = O(YI9978) = (4994) = 31077 Pg = 9916x1071 43x10"? = 316x107 & Py Dm. f Ad A Fl = [f'mpon(m)am = [mdm = Lp Aad mi = f°? pl onldm = [me ad = = 2 oh em (a) =< Se oh/me fin (+a)? m2 mp?) Qhm|/smp)+ 0/4?) SR 2 32) om? | my? = 5 BS (In2s6)" om, 4,1, FE oomata sno? mp 255mp (255)* ™ mp ‘As noted on Pg. (570), the optimum filters for DSB-SC and SSB-SC can be obtained from Eqs. (12.834) and (12.83b), provided we substitute sal 401¢)+Sq(o—0,)] for S,(0) in these equations, Let Sq(o)=3{Sn(o+02)+Sp(0-0)] Q) ‘We shall also require the power of S,,() . I “Eh Snlado We can simplify the evaluation of this integral by recognizing that te power ofthe modulated signal m(1)cosact is half the power of m(t). Hence, ° Sm "35. ado = 2 faa af ae [Sn(o\do Lf ae @ We shall use the PDE system shown in fa 12.19 103,12.52 12.53 (a) For this system 2,0 -—= [s4(0)/Sa(o) ® [rela Palette 4 Because H,(w) and S,(o) are constants, we have Vepto = srffate) , 10/ Bale) ee fe vSml@dar Tf Bale lo where Si (co) is found in Eq, (1). Also from Eq, (12.836) |He(a? oF shale wf wale ® MON Sp AfSale) 10x] Y{Su(0) (b) The output signal is Gm(r). Hence, S, =o?m() ‘We have already found the power of m(1) to be2(a/4) =2. Hence 2 @a_ (10?) (s000") oot ona 2 ro) To find the output noise power Ng, we observe that the noise signal with PSD S,(w) = 2 1079 passes through the de-emphasis filter H.4(a) in Eq. (4) above. Hence, §,(e) the noise PSD at the output of Hq(a) is iy (0)paloh = ego . ba 3 — [Also the output noise power is ng(t)/¥2 and Np = 5 [see Eq. (12.6b)), where nen = 4 FF Ss(o)do and 4. a 6 EP Sloldeo 10f" §,(@)do ‘Similar to Prob. 12.5-1 ‘The improvement ratio in FM is Bint + where dir aroin] mie af?sqlalar= (ay =200 me], 1Sulo)F =f Be VP of we Hence, the improvement ratio is B(2Bp) 4 a oe} pee.Chapter 13 tput of the Gide matched to pie) Fig. S13.1-1 For the integrate and dump filter (ID), the output i the integral of pi). Hence, tt = Ty. Po(T) = AT: If we apply 6) at the input of this filter, the output M() = 24¢)—s4¢ - 75) Hence, Ho) = fpsine (By and The Nee. m= Lie Sane? (2) M1 sine ape e 2 Fg letine (a) and 2) AE 2, mq) oR? This is exactly the value of p? for the matched filter. 13.1-2 The output po() of this R-C filter is Polt)= Afi-e*°) ostsh =A(l- eo BIRCH T)/RC i> ‘The maximum value of po(t) is Ap, which occurs at 7p: Ay = Pol) = Alle ote Wye de 82 2iele@ RC? 4RC rWIRC) 2 RC poh tee of a aon (retry IRC 105‘We now maximize p” with respect to RC. Letting x = 7/RC, we have 2 aay, (1-¢7) os Bo we z a? a per (re) & e This gives or 142x =e* and = 1.26 x=126 or RC” Th Hence, 24% Prax = (0816) (Observe that for the matched fier, 13.24 Pax = 2 LA) a(0) Pa» Fea Tho ‘The energy of p(t) is 75 times the power of p(*). Hence, Ep 2 1m apt a2 Pw), ae no thee, 2 Similarly, Ey = 4h =E Eng = (oleae = (|= 42(1-m?) 2 o -<% + Amt, togt+ Am? sin? og] at Hence, 4A T(i-m) 8; rte 4 13.2-2 LetC;be the cost of error when 1 is transmitted, and Cbe the cost of error when 0 is transmitted. Let the optimum threshold be a, in Fig. $13.2-2. Then: G.=Gup ele!) cu 02] coca p(cin=0)=Cy 0( 2") 1061323 ‘The average cost of an error is C= Paa(1) Cy + Pn(0) Co a) Le\+0)= Yfeud( “6c fe) For optimum threshold dC/day = 0. Hence, to compute dC/dag , we observe that porte) pcela) ete eh / ~ Oe) =1- ee [ew + and * Ap Ze Ap If Pa (1) = Py (0) = 05 4. Leth Fig. $13.22 a ane bd Hence, i (Ap-ae (gral 1 ao ae are C a Wc 7 \20 da, 20,Aam | 1° ae Hence, and But ae, oe and 4, = Ep. Hence, a, = in |S ee LC, We follow the procedure inthe solution of Prob. 13.2-2. The only difference is Fm (1) and Fy(0) are not 0.5, Hence, C= Pr(l) Ch + Pen(0) Co = Pall) Cio @ (Apres) Ape 28 ac = Pal) Cre 2% =. Pall) Ci we, sole ral) Cio € Hence, Pm(0) Co |_ 2404; 2, [ Pn(0) Co » [Ete] Enis altel 107ae, 2 one and A, = Ey Hence, (a Pall) Cro 13.51 Porim,) pe se ft Fig. S13.5-1 abts)= ata) } /203 * Pht 2 AE, Primo) = = oe Pem)= ea te het The thresholds are +E, /2 and ron cin) = Aeim.)ae{ 22) - |i) viele) 42) 40) 13.52 Here, p(t) and g(t) are identified with 3p{1) and p(t), respectively. Hence, H(o) =[3P(-@)- P(-0)e-/9% = 27(-a)_/0% and Ht) =2p(%-1) oa= Hap Ep] =He Ey] =46, But multiplication of H(t) by constant does not affect the performance. Hence we shall choose h() to be PAT, ~1) rather than 2p(T7, 1). This will also halve the threshold toa, = 2E». This is shown in Fig. $13.5-2. Also, Eng = i beOp na =3E, 10813.53 135-4 (EH 9E,+E, “The enerpyhitis Ey = "LP = SEp Hence, ORE, nod PE) a ForM=2, W=2x10% For 256,060 bps the baseband transmission requires a minimum bandwidth 128 kHz. But amplinude ‘modulation doubles the bandwidth. Hence By = 256 kHz 107 -f Es» Ey 227x107 S, = EgRy = 2.7 «1077 «256,000 = 0.0697 For M=16 256,000 Tog2 16 = 64 kHz 9, [24e, Vass Bre Pea = Pp l0B2 16 = 4x 10~ This yields Ey = 543x10~ S, = EyRy = 543% 10° x 256,000 = 1397 For M = 32 256,000 Tog: 32 Pag = Py logy 32 = 5% 107 = S12KH2 2) aes =) This yields E, = 1.71910 'S, = EpRy = 1.719 «107 256,000 = 447 Bre For M = 2andA= 2x10 ‘This case is identical to MASK for M =2 107 = 4 Pe) = Ey=27%1077 ay Si = EqRy = 2.7x107? «256,000 = 0.0691 y= 280 an site 109For M=16 In MPSK, the minimum Hence, For M=32 Pay = (loge 16)F, = 47, 24x107 2 4xio7 52 a => Ey = 167107 2568 S, = EyRy = 167 10 «256,000 = 042757 ‘bandwidth is equal to the number of M-ary pulses/second. 256,000 1og2 16 = 64 kHz Bre Pay = (log: 32)A, = 5% 107 lan? (SEs) 10244 5x10" g PED ensue S, = EyRy = 524 106 «256,000 = 134 256,000 log 32 Bre = 512kHz 10Chapter 14 ee 14-1 The following signals represent 2 sets of 5 mutually orthogonal signals. 141-2 FigS14.1-2 m1413 1) 041.08 Flt Psina oo In 2) @LDis FP anonrs essen] =F prmow4)] Te 1 . dy G2 -> is Fleviseow- oes] = Felis 2eteodoer 104 ] Fig. S14.1-3 42141-4 142-1 * * ¥# (3-2,3,4,9 Fig. S14.1-4 b) The energy of each signal is: 44+ 4 x,t 4 ee lels ea Fy = Se telsr is The4l % % By = SAO 39 By = StiSt4 e407, a 28 % is ©) Fy-Fa=(-6-8+6+840)=0. Hence, f(t) and fa(1) are orthogonal. Let x(t) x, x(t l= xg x(t42)=x3 We wish to determine Prax (410423) Since the process x(é) is Gaussian, x), x2, x3 are jointly Gaussian with identical variance (2, = a, = oy = R,(0)=1). ‘The covariance matrix is: ameKey = R()=t 2 Oxyxg * Fx2x Oy Fux Tus a o i Kelonn 2% Oma | ABO Ongny = Page 7 HP HEF) +2) = Rl m 2 |= aT 1 Oxy, Fayr Fay rayny © Faye, = XS = HEY AE2) = Re(2)= 41314.34 1y and (K[=(1--—> u-( 2) aie ele = eye i 1 egy = 1--k and Aya = Aa) = 43 = Aa 5-5 Ay nga ty and Ona = Ban = as = Aa 5g 1 bys Ag 0, dap =1- Ant Pe Pan V2)" C57 Fag Ug Fig. $14.3-1 rem) = Poem <) wa cma) =P 01>) 4 cin) = (Cra) == PCC) = Pron < 2) Hence Plc) = (Ca) = ig Le a 1-5) and 1? aad 4c) = {ces =o (Crs) = Jag a amt od) Hence Py =1-P(C) 21-[ton)( Chm) + ma) P( com} ora) {Cha} = 4 w-xwi de] ‘The signal energies are (zzy. (:). vn (ate) M4Hence the average pulse energy E is rn 22 92.8. wea] M4 4 4 ma 2S axes? (2 -i)e? ai Lee = 7 = (unit A ne Fe blog, M I2log, M Hence Pug eM Hicea MIE mM MV (Me=i) ‘Which agrees with the result in Eq. (13.52¢) 1432 P(cim)= Pcie) = Pcs) = AC) (Choma) = P(Chos) = PCIe) = PCIe») Heim) = om <8. >) delle (Chm) = Alul<$, ™ >3) das) reyed [aterm Amalle4ft-o( el raeet-nernsd fa) [> h)] and Fig. $14.3-2 us‘The average pulse energy E is ed - Oa) -2 et log28 2 vi (2) lB) Ey ‘This performance is considerably better than MASK in Prob. 14.3-1, which yields rau 2180 (EBSEE Ne wt 'M. Hence, the optimum receiver is the same as that 14,3-3 In this case, constants a,'s are same for k= 1, 2, .. in Fig. 14.8 with terms ay's omitted. ‘We now compare r-54, 7°82) FSM + Since r-s, = VEr cos, is the angle between r and ‘sq, itis clear that we are to pick that signal s, with which r has the smallest angle. In short, the detector is a phase comparator. It chooses that signal which is at the smallest angle with r. 143-4 Because of symmetry, (Cm) = P(Cha) where M=2% = Fm) Fig. S14.3-4 16and P(C) = P(Cm) enrol Here, M=2", Hence, each symbol carries the information logz M= Nbits. Hence Ey = Ef nbd Pm) , 4 Pm)" 2 Pi da =—in2+— Fig, S14.3-5 2@” +9 ‘ig. S14. and 143-5 wi =—h- ae” Hep) = cna) =r») 38) Achre)= Pr (sl yar-ad zea) Ac) =4 (cima) +4 2( com) +5 AC) pga) tes) 1 (a? -win2 d? + Nin2 a ae) daa) < 2 E = 05(0)+025d? +0250? = 2E 2E a ecm?) | th? ihe nord i 4 sen | 143-6 Pay =1-P(C) P(C)=3[Alcm)+ Aim] Heme Zager? Manan Also Hence ud 1 sqaten lea abl mw i " lana 72 . 4 gto +a lf andes 24) af | (Chm) monde ele 1g, tan(n/8) gpd and : d= \OAE = OES tant lh Fig. S14.3-6 143-7 Note that d, 4 dy 44 e-fhn ge na shitgh d, 4 dyad gaSh-Gh weghegh 86) # $8) a [Se 4 5,8) 1 7. ,* Pt» = | ~ ° % o| T Zt Teo jo 5 fe a *s % oA Fig, $14.3-7b © P(C{ms) = Prot ite originating from s5 remains within the square of side +) sales. balexSy) 2 [doe] ade} oan flea ers =f] < = 2 We also observe that E , the average energy is E = {4}. sd? E _ 04a? _ 02d? a _ (sé a SE EE OE and yd = a a ot 2a, “Vai 2o, VON Therefore named |) ‘The decision region Ry for m, is shown in Fig, a and again in Fig. C-1. Ry can be expressed as the first quadrant (horizontally hatched area in Fig. C-1) -4,. Thus P(Chm;) = noise originating from sz lie in Ry = P(noise lie in 1st quadrant) ~ P(noise lie in Ay) and -{- if Jp rete cat tein) But —_P(noise lie in 41) = Hl P(ooise Iie within outer square) - P(noise lie within inner square)] (See Fig. C-2) = {Ao sl <8)- of al $5) “ste dete 9fad }udefe)] and nemevades,) 45) df) ae de }- ee on } 4B} nonod 2) 40 Fig. $14.3-7c Moreover, by symme see Pda) = dm)= = Plelms) = P(sima) tu atdn)] “pd 48) 143-8 ey & Ri ° . Fig. S14.3-8 nicy=dbartcin)=G[Ace) + Hcim)> Aces) + cm] ‘The decision region R, for m (see Figure) can be expressed as y= outer squae of ide 2/2 -3 (outer square ~ inner square of sided) = Louter square of side dV +3 inner square of side d 120Now P(C{m) = Prob(noise originating from m lies in) = fA lie in outer square) + 2 p(n tie in inner square) =1Wlef)-2e{ <5) fda] ld] Similarly Rp, the decision region for m, (see figure above) can be expressed as R outer square of side 2/2 - 2 (outer square ~ inner square of side d) = outer square of side df - ane square of sie d and P(Chm,) = noise ray td tie " =the) +4 “today: ‘ “el 7 wh ‘es a my ad Y a Ra ee The decision region Ry for my can be expressed as y= Ry + Rp- Re P(Cims) = Prob(noise originating from m; lie in Ry) = P(noise in R,)-+ P(noise in Rg)- P(noise in Rc) wf kce)e 2 a ake} 4 le f)-api mtd] fad Ballade) - afta eal nsfead Eade) & date)‘The decision region Rg for ms can be expressed as Ry=Ra-Re and ha) => 02> af ale) <4) {4B foe ot] For any practical scheme O() <<, and we can express {i-4atyf = 1-200) Using this approximation, we have rene to renimvad fo} renin} 4) remind) Hence 1 Pc) = 5 P(cim)+ Pima) * P(clms)+ (Chr) id ig 4)-3 af 1 fa) aya av ‘Now bynd?, Ey -2d?, Ey = Ad?, and Ey = 8d”. Therefore Bale ea oad etd) G4 , EOE And aT so that fy Ed A 4H 16 Ww ‘Therefore a3 [EEa)-Sd [8% )-2q 2 rc)=1 id sa id, 15 id 15 W Moreover TEs), of SE o> of (Ee 1S 15 A sw Hence 3 BE ncyar-4o( {5 122 And143-9 144-1 ry =t-rte)~$0( [EB Comparison ofthis result wth that in Example 14.3[Eq,(14.57)] shows that this configuration requires approximately 1.5 times the power ofthe system in Example 14.3 to achieve the same performance. If 5; is transmitted, we have byeE+arVEn, , by=-E+a-vEm bpeasVEn, =» b.ga-VEm by wa+VE ny » by za-vE my and (Chm) = Prob.(b, > 8-1, bps b-2» ~~ be» b-4) Note that by >. implies E+a+VE m>-E+a-VEn of m >-vE b> by implies E+a+VE 0 >a+VE 02 or mb.g implies E+a+ JE ny >a-VE my or m>-(vE+m) Hence by > by and b-2 implies - (nj +VE) <2 <(m+¥E) Similarly by > by and b_, implies ~ (a +E) 1, bas Beas bs, b-3s~* Ba bot) «Prob ny > VE, [pal <(ou V2} [sl «(va +42} bal <(m +E] Since ny, np~-M, are all independent gaussian random variables each with variance u/2 P(cim)=[P(01 >-VE) Peal em + ¥E)Pral Fig S14.4-1 2 tae2 Hew 44) [Ecos = exon eet 92(t) = Va0 sin wg Li Therefore si(i)= ¥20 42(1) | 1=¥ 2 a(t)=V50() | 2 =v5O s()=-5 6(0) | s=-Se reves «= 154 «Is 0,-ViB 0, + @,)- Be, : Hence the minimum = signal set is given by ali)= (2 60 = 5 431) Bye 2 sno sine nth Enon ain sone ioy a= wr Bele-6.00)-Binpeoteney BE toy ‘The optimum receiver ~ 2 » citable form — in this case would be that shown in Fig. 14.88 orb. Fig. $14.42 1443 To find he minimum energy set, wehave # = (5) +82 +83 +54) =A —#e Hence the new minimum energy set is henge hth hentia h- tah 124[Note that all the four signals form vertices ofa square because (i, 2), (2 4s) (#35 ), and (4 §) are crthogonal, The distance between these signal pars i always 22 . This set is shown in Fig, $14.4-38 “The signal set is now rotated so as to yield a new set shown in Fig. S144-3b. Q, = 2H 5SN- am %=%7 ~° “602 Fig. S144-3 Observing symmetry we obtain (c)= (cfm) = (Cr) = ACs) = ACh) = P(n >—v2 and ny >-v2) -(-4$)] = [- die] =[1-oG16)} Pay =1-P(C)=1-[1-O(316)} = 158%107 2 1 1 1 mes anghtgmtetame oo tppsmlatel Oe) wt SiR Se oe s. s sa 3 we no? t> Si ws ioe (B) The signal ext (@)The minimum energy af ce) tov! ia tov 1st () The orthogonal basis set ’ / at 38 = 88 83 2x10 5 co) ey ts q 2nert~ “4 for ens (AD The minimum energy Wave Fig. S14.4-4 125P(Chms) = P(Clm) = am pool te . coo t* ad =2 ne Aa 92 00/38 fy" 0.028% 031 Fig. S14.4-6 127Chapter 15 ae 15-11 15.1-2 151-3 151-4 151-5 A. =04, Py =03, y= 0.2 and Py = 01 Hom) =-(F lop A+ Py lop + Pslog + Fs log Ps) +=1.846 bits (source entropy) ‘There are 10* symbols/s. Henee, the rate of information generation is 1846 10° bits/s. Information/element = log, 10 = 3:32 bits. Information/picture frame = 332 x 300,000 = 9:96 «10° bit. Information/word = log) 10000 = 133 bits. Information content of 1000 words = 133 x 1000 = 13,300 bits. “The information per picture frame was found in Problem 15.1-2 to be 9.96 108 bits. Obviously, it is not possible to describe a picture completely by 1000 words, in general. Hence, # picture is worth 1000 words is very much an underrating or understating the reality. (a) Both options are equally likely. Hence, 1 = log(ds) =} bit (b) P(2 lanterns) = 0.1 1(2 lanterns) = log, 10 = 3322 bits (a) All27 symbols equiprobable and P(x;) = Ye . Hy(x) = 27(g5 logs 27) = 4.755 bits! symbol (b) Using the probability table, we compute u Hy (x) =~ PCa) log P(x) = 4.127 bits / symbol isi (©) Using Zipf's law, we compute entropy/word H (x). 327, Hy) =~ ZPee. Por) m2, = YMlog Ah) = 9:1353 bits/ word. rt Hiletter =11/82/5.5=2.14 bits/symbol. Entropy obtained by Zipf's law is much closer to the real value than Hy(x) oF a(x). 1282 83 4 1524 H(m)= ¥ Ailog A = 55 bits it Message Probability Code = an Ss 5 ™ 2 0 i 0 m= 0 i 0 7 0. 120 m; v4 10 v4 10 v4 10 4104 102 I m; vs 110 ge 110 V8 110,18 110) v4 11, m 16 110 V6 TNO 1/16 Oy VB 11, ms 132 MNO_ 321110116 1111, m 164 MM10P 132111 m 64 wii 1 1 1 1 1 1 1 Le DAL HOGA) gO Ge GOO 3 me Sp binary digits H(m) Efficiency = x100= 100% Redundancy 7 = (100-7) = 0% 4 152-2 H(m)=-¥ Plog P= 2.289 bits fet 2289 - = = 14442 3+) ieg3 2 ary units: ‘Message Probabilit Code Ss Sy ™ 13 0 7 0 0 m 3 1 V3 1 V3 1 m; VW 20 VW 20 uw 2 m VW 21. vw 21 ms wat 20 ww 22 ™ 127 21 m) 1a7 222, - Vy lay toyed 1 L= DRL =4()+2()+22)+ 52) +3551 Zab gt g@+5@)+50)+35,0) -2 3-ary digits = 1.4442 3-ary digits erin 9 = HD Ha «100 100 Redundancy y =(1-1)100= 0% 1294 152-3 H(m)=-¥ Plog F, = 169 bits —_——— Vessage Probability Code = = 7 03 70s 0, 0s 0 m 03 10, 03 10,505 1 ™ 01 110-02 11, ™ 0 1 L=ZPL, = 0541) +03(2) +01(3)-+0.1G) = 17 binary digits Efficiency = Hm 100 = 46 100 = 992% Redundancy y = (1-7)100= 03% For ternary coding, we need one dummy message of probability 0. Thus, Message Probability Code: ™ Os 005 m 03 1 03 1 m 01 20 0.2 2 m O1 21 ms 0 22, L = 05(1)+0.3(1) +01(2)+01(2) = 12. 3-ary digits Hom) =169 bits = 2 = 10663. 3-ary units og 3 Efficiency 1242100 ee x 100 = 8886% Redundancy y = (1- 7)100= 11.14% 182-4 Message _ Probability Code = = —r 12 0 Tm 0 mz 0 m Vs 1 wma 1. woo m, ves ve 20 42 m 16 we 21 ms 132 ING 22, mg 64 im 64 LaDAL a 3-ay digits 16 From Problem 15.2-1, H(m) -€ bits = 1242 3-ary units H(m) =A) «100 1282. 109 = 94.63 Efficiency 17 11008 5 % Redundancy y = (1—7)100= 537% 13015.2.5 ‘Message Probability Code s & 3 “Tn v3 T 13 T | 7S 1 18 1_ pros 0 m wa 00 18 (0018 si YL 3 ms 19 On 1 O1l V9 O11, pw2/9 010-131. my "9 0100 1 0100, = 1/9 100] J 1/9 ON ms 127 01010, 127 01010 } =» 1/9 0101, me 127 0101107} —» 227 01011 m 127 o10111, eran S = 2.4074 binary digits H(m) = 2289 bits (See Problem 15.2-2). Efticiency =H) 100 = 228 100=9508% Redundancy 7 =(1-7)100=492% 15.2-6 (a) A(m) = 3(4 10g 3) — 1585 bits (b) Ternary Code Wessage Probability Code ™m rE 0 m v3 1 m, V3 2 La tat i La ZO4 z+ za! 3-ary digits ie. 1585 . #1585 bits = 28% 21 3-ary unit ‘H(m) #1585 bits = 7=—Z=1 3-ary unit 82 Efficiency = Ao) 100 = 100% Redundancy 7 = (1-7)100 = 0% (©) Binary Code L= 40+@4@) 3 = 167 binary digits Efficiency 1-2. ts, 100 = 95.08% Redundancy y = (I~ 7)100 = 4.92% 131@® ) Second extension ~ binary code Malt 1 29 a, Le sla(3}o +2(3}o] aa 1.611 binary digits ‘H(q) = 1585 bits Efficiency = 4. ©) 100 Redundancy y = (1-7)100= Ea = gi * 1007 9839% 161% = = = 1 Ol p10 wt coor | i ooo [1 or mae ro | woo | 1 coo f 19 0) mr | wie | v8 oooL, | 19 0000” fo | in. [ie nopd 1 ooo, ie ee oop] 19 10 = = = TT 0 fix oa fis wp) as 2 no] w2® wit 2 oO 28 » Ou. 9 oO, ‘The channel matrix can be represented as shown in Fig. $15.4-1 POn) = POrlay) P(x) + POilz2)PG2) x 228 =5 “to 3 “8 PU) = I~ roped Ped 1 1 3B %2 H(x) = P(x )log: Pa? Pea dea Ty = Hog23+ 210g; 3 = 0918 bits Fig. S141 To compute H(x\y), we find _ Ponixy) PCa) _ 10 _ PoplsPen) 5 PO In) POn) B POly2) PU) n P( Pisqly)) = Pola) Pesan) = Pear = Pon) 64 1 OO TG Ai 5 Pea TS 3 13 = Blogs B+ Seg Y= 079 1 H P(x|y2)log——— (x\y2) = PC|y1) log = —— Fi Dy (x2 |¥2 log Peba) Sons 2+ Bog, St = 0624 132 2/3and H(xly) = eae P(y2)H(zly2) = Zorm-3 = (0 624) = 0.6687 ‘Thus, Uexly) = H@)~ HOaly) = 0918 - 0.6687 = 0.24893 bits/ binit HQ)= LPO Fl a Bog ti» Bosse = 08673 bits/symbol T Also, H(y\x) = H(y)~ I(xly) = 08673-02493 = 0.618 bits/ symbol 154-2. The channel matrix P(y,;) is (P) —_+—, CPD , 1 0 0 P hea @) Eee ® 0 1 * POP Also, P(n) = P, PO2)= PUs)= 2 @) *5 P 3 Now we use Plely,)= eee PEO tg obiain Fig, S15.42 LPO) 10 0 P(xily,)matrixas —-y, [0 p Imp O lp P H(x) = E P(x) log- Pa” ~Plog P-2QlogQ with (2Q=1- P) = {ome P+(l- nro 52)] =0(P)+(1-P) 1 Hosly)= rE PO Ply DB EG) = Plog + rarbee-rln}e-via t+ 9h05] a 1-p. ta f = 0+20M%p) = (I- P)ACP) Axly) = H(x)~ Healy) = A(P) + (1= P)~ (= PCP) =0(P)+(1- PYI-2(p)] Leng £=2%? or Olp)=lon 8 I (ly) = A(P)+(1- PX 10g A) Ff 4 Ziewyy=o oF Flr) +(1— PAI tog Ay} =O. Ths means SZ lpiog p+) Poett- P\t-le6)]=0 Jog P~ log(1~ P) +[1-logA]=0 133aa1+ logs Therefore log: 1-P Note: <1 logy = 10822 +1052 B= 062 4 Pak = AL ang 1- P= 2 a gen PB $0 . Bigg #2 4 2 tog B22 4 2 (1-198) = p+2 C= MAXI(xly) pap Fa pat web) eG B 18463 Consider the cascade of 2 BSCS shown in Fig. $15.4-3. In this case Pyy(it)= (I~ AXI- B+ AR = 1A 2A Py(Ol) =(1- ADR AC A)= A+ 2A (a) Cascade of The Rt» BSC , = ct) Hence, the channel matrix of the cascade is [roncine A+h-2AR J} fn" ats? ea A+P,-2AP 1-R-R-28P] LA MAILE 1-P, ‘This result will prove everything in this problem. (a) With A = Prom the above result follows thatthe channel matrix is indeed mm. © Werhave already shown thatthe channel matrix of two cascaded channels is MM2- (©) Consider a cascade of # identical channels broken UP 38 channel cascaded with the k’ channel. If M,., is the channel matrix ofthe first 1" channels in cascade then from the results derived in par (>), the channel matrix of the k cascaded channels is My = MyM. This is valid for any k. We have already proved it for k =2, that Mz = M2. Using the process of induction it is clear that M, = M* aeeaey ify these results from the development in Example 10.7. From the results in Example 10.7, we have, for a cascade of 3 channels 1- Pr =(l-B) +3P70-F) 1-3P,43P2- PO +3P -3Pe =1-3P, 4672-47 and Pp =3P,-6PP +4PP Now 134fa fe" p Te (3, -6R2+472) 3% -6R2+4P? «I= |5p,-6R244R? — 1-GR-6R,? +42?) Clearly Pe =3P,-6P2 +4P which confirms the results in Example 10.7 for k = 3. (@) From Equation 15.25 L 1 [re bape td- Pe)logs— | where Pp is the error probability of cascade of & identical channel. ‘We have shown in Example 10.7 that reat o- Ry om ee we PAQ- nw] d Wa < —(1+logp) dicep)= pand Bat ‘Substituting these quantities in Equation 15.37, we have “(1+ log p) +a, =0= p=e™™! and yn Mette 2a) <1 Hence, ted wage 2M 2M Also, 1 1 HQ)= [penton d= | spp 082 Mae = og 2M 15.52 Wehave H(x) = -[° plog par, A=|y spar, I= fp pe F(x,p) =~plog p and FF --stp) tiasp)= prand Sax #2(z,p)= pand aa ‘Substituting these quantities in Equation 15.37, we have ~(1+logp)+ayx+a2 =0 or pa et002 a (@22-Nerit Substinuting this relationship in earlier constraints, we get Hence, sett =a, p(x) =-aye™ et Ie JP pede = [Peet = 1 and A [pape ase Hence, To obtain H(x) 137H(3) = -[P pCa) log (ada = -[p° 1-8 A ine] C3 = log 4\> playa + EELS shade = log A +loge = log(ed) 155-3 Information per picture frame = 9.9610 bits. (See Problem 15.1). For 30 picture frames per second, we need a channel with capacity C given by C= 30%996 105 = 2.988 x 107 bits/sec. But for a white Gaussian noise Ss = Blog [1+— c= Blog (1+) We are given =50db =100,000 (Note: 100,000 = $0 db) Hence, B=18MHz 15.5-4 Consider a narrowband Af where Af -> 0 so that we may consider both signal noise power density to be ‘constant (bandlimited white) over the interval Af - ‘The signal and noise power Sand N ‘respectively are given by S=25,(w)4f and N=25,(o)4f ‘The maximum channel capacity over this band Af is given by S4n S,(o)+5,(0) a ap tog [StM]= ay og | 2+ Sno ow aoe W | a | x) | “The capacity of the channel over the entire band (/j, 2) is given by fa igg { Sei Salo) cuff ire [ xr ‘We now wish to maximize C where the constraint is that the signal power is constant. 2 Sw) df =S (aconstant) ‘Using Equation 15.37, we obtain 2 tog [5st 50 |, pS = =| Ss }eat ° 55+Sn= + (aconstant) @ ‘Thus, 5,(0)+S,(0)=—* a ‘This shows that to attain the maximum channel capacity, the signal power density + noise power density must be a constant (white). Under this condition, C= [Plog [seepesied foes 2] a =| S,(0) aul be =ci- sien (-2}- foe (storey 138= Blog [5,(o)+Sq(@)]~ [7 log [Sy] 15.55 _ In this problem, we use the results of Problem 15.5-4, Under the best possible conditions, C= Blog [S,(o)+5,(0)}- 7? log [Sn(or] of consant We shall now show that the integral if Jog [Sp()] df is maximum when 5,(«) = constant if the noise is constrained to have a given mean square value (power). Thus, we wish to maximize [Plog [Son]ar under the constraint 2{ft og [Sy(o)] af = N(@-constant) Using Equation 15.37, we have leona aR 0 or dtaz0 Sy Sy(0) = 4 (a constant) ‘Thus, we have shown that for a noise with a given power, the integral Jog [Sao] is maximized when the noise is white. This shows that white Guassian noise is the worst possible kind of noise. 139Chapter 16 a 161-1 16.1.2 1641-3 16.2-1 162-2 3 We ZF)=(F) PPP) i 2048 21423423 1142377 = 2048 @) There are (1) ways in which positions can be chosen from n. But for a temery code, a digit can ‘be mistaken for two other digits. Hence the number of possible errors in j places is m3) i" ¥¥ mot yk $ (Aoi (9fo-a! oe rast £ yp! oars Ep (b) (11,6) code for t= 2 35 2()+({Ha+(p)2? =1422+220=245 This is satisfied exactly. For (18,7) code to correct up to 3 errors 22 38) o 2 2(8}+(1)+(8)-() eo wie 38, 18 1! TH" 2146” 3 15! 14184153 +816 = 988 2s 3 (3) Clearly, there exists a possiblity of 3 error correcting (18,7) code. Since the Hamming bound is oversatisfied, this code could correct some 4 error patterns in addition to all patterns with up to 3 errors. GH’ =[, alr] =POP =0 c= dG where d is a single digit (0 or 1). Ford =0 e=0([111]=[000) Ford =1 e=iftit}=[111) 140162-3 16.2-4 162-5 c= dG where dis a single (0 or 1). Ford =0 e=0[11111]=[00000) Ford =1 e=tfiiiitj=(1111a) ence in tis cade a digit repeats 5 tines Such a code can correct up to two errors using majority rule for 0 is transmitted by [0 0 0] and 1 is transmitted by [1 1 1] (a) This is clearly a systematic code with G@=[11 i] (a) P= Note that m= 1 mrocnHod moneron Hoon Hol (©) This i a parity check code. If single error occurs anywhere in the code ‘word, the parity is violated. ‘Therefore this code can detect a single error. (@) Equation (16.9a) inthe text shows that ef” = 0. Now ree@e and rHT =(c@e)H =cH” @eH™ =eH™ If there is no error ¢ = 0 and rHT = eH =0 Also 7_[? - HT =) 7 |. Butsince m= tm =[H] and 141ifthe isa single eror in the received word r, eas a single I element with al ther elements being 0. Hence HT =eH™ =1 (for single error) 162-6 ta word ovonnera eosee-™ eoneene From this code we see thatthe distance between any two code words is at least 3. Hence yin = 3 2 ouonneoel Observe that digin =3 mune ne ocd wn oconne ee | 1 1 0 0 1 1 0 weronen 162-8 HT isa 15x4 matrix with all distinct rows, One possible HT is: Liat 1110 1101 1100 10o1n 1010 1001 HT =|0011 i] Tn. Olli 0110 o10l 1000 0100 0010 0001 142100000000001111 010000000001110 001000000001101 000100000001100 000010000001011 G=[I; P]=|000001000001010 000000100001001 000000010000011 000000001000111 000000000100110 000000000010101 Ford =10111010101 c=dG=(10111010101]G=1011101 01011110 16.2-9 (a) 1a Toor) toe Gz=loro110] & AT=1, 0, O}L0 seri o10 001 ) ‘Data word Code word 00/0 00000 oo1r/o ort oO? o 1 0/0 1 011 0 o1afoi toad 10 0/1 0 O11 79 10 1/1 0101 0 11 0j/1 100028 11ijttrr1 00 (©) The minimum distance between any two code words is 3. Hence, this isa single error correcting code, ‘Since there are 6 single errors and 7 syndromes, we can correct all single errors and one double error. @ s=eH™ | HeccceH eococHa ecconcd mooncca eorecece eoo-ees| merco+ wn eonoH 143© 7 5 € c a To1100 110 010000111100 Tit 000110 110 010000 010110 010 101010 000 © 000000 101010101 16.2-10 (a) done in Prob. 16.2-7 Ol 101 ay © BT Als 49 010 oot 010000 | 101 sixsingle errors 001000 | 110 000100 100 o0o10 | 010 000001 001 1 double error 100100 i 16.2-11 1000101 010011! G=[P]*loo10011 0001110 end 144162-12 16.2-13 i ¢ 0000] {0000000 0001] |ooo1110 0010) |oo10011 oo11| |oorit01 0100) |o100111 101 0101] |o101001 wt 0110) |ors0100 on 0111} [0111010 AT =|110 1000| {1000101 100 1001] |1001011 010 1010] |1010110 001 1011] |1011000 1100! [1100010 1101] [1101100 1iio0] [1110001 anaa) [pagaaiy s=eH™ ‘ f 0000001 oor 0000010 010 0000100 100 0001000 110 0010000 oll 0100000) |111 1000000 101 s= rH where r= received code cur@e €= corrected code ‘We observe that the syndrome for all the three 2-error patterns 100010, 010100, or 001001 have the same syndrome namely 111, Since the decoding table specifies s = 111 for ¢= 100010 whenever = 100010 Ceeurs, it will be corrected. The other two pattems will not be corrected. If for example ¢ = 010100 eceurs, s= 111 and we shall read from the decoding table e = 100010 and the error is not corrected. if we wish to correct the 2-error patter 010100 (along with six single error patterns), the new decoding table is identical to that in Table 16.3 except for the last entry which should be € 8 010100 From Eq. on P.737, for a simple error correcting code 2H ened ot 282 nt] > n-82 logy(n+1) ‘This is satisfied for n212. Choose m= 12. This gives a (12, 8) code. HT is chosen to have 12 distinct rows of four elements with the last 4 rows forming an identity matrix. Hence, 45G=[1,P] oir 0101 100000000011 0110 010000000101 oi | |, 001000000110 1001 G= 000100000111 nT = | 1010 = [000010001001 1011 000001001010 1100 000000101011 1000 000000011100 0100 oo10 | ff 0001 ‘The number of non-zero syndromes = 16-115. There are 12 single error pattems. Hence we may be able to correct 3 double-error patterns. £ ¢ 0000 000000000000 0011 100000000000 0101 010000000000 0110 001000000000 0111 000100000000 1001 000010000000 1010 000001000000 1011 000000100000 1100 000000010000 1000 000000001000 0100 000000000100 0010 000000000010 0001 00000000001 1111 100000010000 1110 001000001000 1101 000000010001 16.2-14 00 11011 101110 110101 ‘The minimum distance between any two code words is dyin =4. Therefore, it can correct all I-error partems. Since the code oversatisfies Hamming bound it can also correct some 2-error and possibly some 3eerror patterns. 146G=[1,P] oir 0101 100000000011 0110 010000000101 oi | |, 001000000110 1001 G= 000100000111 nT = | 1010 = [000010001001 1011 000001001010 1100 000000101011 1000 000000011100 0100 oo10 | ff 0001 ‘The number of non-zero syndromes = 16-115. There are 12 single error pattems. Hence we may be able to correct 3 double-error patterns. £ ¢ 0000 000000000000 0011 100000000000 0101 010000000000 0110 001000000000 0111 000100000000 1001 000010000000 1010 000001000000 1011 000000100000 1100 000000010000 1000 000000001000 0100 000000000100 0010 000000000010 0001 00000000001 1111 100000010000 1110 001000001000 1101 000000010001 16.2-14 00 11011 101110 110101 ‘The minimum distance between any two code words is dyin =4. Therefore, it can correct all I-error partems. Since the code oversatisfies Hamming bound it can also correct some 2-error and possibly some 3eerror patterns. 146®) 1110 1011 1000 0100 0010 0001 z T00000| 1110 010000 | 1011 ' 001000 | 1000 6 single error pattems) 999100 | 0100 000010 | 0010 000001 | 0001 110000 | 0101 101000 | 0110 100100 | 1010 7 double-error patterns. 100010 | 1100 100001 | 1111 011000 | oo11 010010 | 1001 ooornt | O11 2triple-error pattems{ 954401 | 1101 163-1 Systematic (7, 4) cyclic code g{x)exextl Fordata 1111 d(x) etexel BP atersiestedarter Sextet axellxoe ext ex? x8 xte o saved tx+l axel da) (2224 eeei) aber “The code word is 11111111 Sextet ex? 147 tre]Fordata 1110 (x)= 22 +x7 +x tx Daxel)xoex ox! x8 tetas? ‘The code word is 1110100 A similar procedure is used to find the remaining codes (see Table 1). (b) From Table 1 itcan be seen thatthe minimum distance between any two codes is 3. Hence this is single error correcting code. € Tid) @iitiit 1110] 1110100 1101 | 1101001 1100 | 1100010 1011 | 1011000 1010) 1010011 1001 | 1001110 1000| 1000101 Table! 0111 | 0111010 0110] 0110001 0101 | 0101100 0100 | 0100111 Oo11 | OO11101 0010 | 0010110 0001 | 0001011 0000 | 0000000 (c) There are seven possible non-zero syndromes. for e=1000000 Saxet)x' se101 48163-2 ‘The remaining syndromes are shown in Table 2. @ = To00000 | TOT 0100000| 111 0010000) 110 0001000| O11 0000100] 100 0000010| 010 e000001| 001 Table2 (@) The received data is 1101100 Ha) x8 oF as? Pareles thes? States? ext ts? Bt es? xtes? xt tates axtex sa)=2 41 xt! 2101 4 From Table 2 ¢=1000000 e=r@e=11011001000000~ 0101100 Hence d=0101 a(x)ex! +P ox7oxSexS ert] e(z) = d(2)a(x) d;=000011110000, ay(x)= x? 4x8 4x5 4x4 01010101010, dy(z)axex2 4x7 4x5 +27 4x eq(x) = y(a)a(2) =a! ox! ax! 4x2 all aa? aah 43” + and ¢,=00001100011101110010000 a(x) = dy(x)gla) = 22 4x! 42! 4a 4s oat eS eat eal ex? x and €,=10001101010000100111110 149x41 z 163-3 xelxtax? tet xr? x41 xl 0 Hence Paxtanate(e sia? 1) =(e+ ier ier) =(e4? xfexel 163-4 x4ljxiext ex? +1 Hence Sash esta le (xix! +z4)) ghee’ Now try dividing x4 +241 by x+1, we getaremainder 1, Hence (x+1) is nota factor of (x4 +x+1). ‘The 2™-order prime factors not divisible by x +1 are xtandx? 4x41. Since (xt +x+1) is not divisible by x2, we try dividing by (x? +2+ 1). This also yields a remainder | Hence x4 +x+1 does not have cither a first or a second order factor. This means it cannot have a third order factor either. Hence Daoxt esate (es ifat axel) 163-4 ‘Try dividing x7 +1 by x+1 xSeeerteDextertl xel)x741 xlsx x84! 1503 Therefore (x7 +1) = (e+ fab ox8ex4 27437441) Now try dividing (oS +25 424 4x3 4274241) by (x41). It does not divide. So try dividing by (+1). It does not divide. Try dividing by (2 +241) It does not divide. Next try dividing by (29 +1). 1ndoes not divide either. Now ty dividing by (x3 x41). It divides. We find (s8+aSsate ext eret)a(Peztie ee? a)) Since (2 +2? +1) is not divisible by x or x+1 (the only two first-order prime factors), it must be a third-order prime factor. Hence x7 a e(x4li(x? tee ifs? +27 +1) For a single error correcting (7, 4) cyclic code with a generator polynomial g(z)ar axel hed n=? xt g(x)] xt? gfx) xP g(x)] [xo +25 4x7 Hence 1101000 0110100 0011010 0001101 Each code word is found by matrix multiplication c= dG" 1101000 o 0 c=[0000] sorsero = 0000000 ooo1101 1101000 0110100 0011010 0001101 ¢=(0001) = 0001101 151‘The remaining codes are found in a similar manner. See table below. TO00000 ooo1lol 0011010 oo101n1 0110100 0111001 0101110 0100001 1101000 1100101 1110010 wiatiit 1011100 1010001 1000110 1001011 0001 0010 oo1n 0100 0101 0110 o1nt 1000 1001 1010 1011 1100 1101 1110 anit 163-7 g(x)ex? +x? 41 The desired form is 1000+ = = Oy hay tar * Am 0100+ -- Oh2 haz haa * - * bmg 0010 Oy Fag fag > > ° my Ge rer 9000-3} hy hae hae * * him he — ¥ (kxk) (kxm) ‘The code is found by using e=dG Proceeding with matrix multiplication, and noting that 0400, OFtN1+0=1, 14120 and Ox0=0, OxIeix0=0, IxI=1 we get 1000110 0100011 es=[1114] |oosoiat =(nnna11a) 0001101 cys =[1110] @=[1110010) and so on. 15216.3-8 aint 1110 |1110010 1101 | 1101000 1100 | 1100101 1011 | 1011100 1010 | 1010001 1001 {1001011 1000 | 1000110 0111 }0111001 0110 |0110100 0101 |0101110 0100 [0100011 0011 | 0011010 0010 | 0010111 0001 |0001101 0000 | 0000000 ‘These results agree with those of Table 16.5 @) 1011000 0101100 “loo10110 ooo10rt (b) The code is found by matrix multiplication. = dG" In general (x)= ant +. a + Broke Forthiscase gj =, g2=1 83>, g4 =! Since iy = £2. Pax = 83. Pax = £4, the fourth row is immediately found. Thus, so far we have G= 0001101 Next, to get row 3, use row 4 with one left shift. 0011010 0001101 ‘The i is eliminated by adding row 4 to row 3. oorolll 0001101 ‘Next for row 2, use row 3 with 1 left shift. 153,163-9 o101110 010111 0001101 ‘The | is eliminated by adding row 4 to row 2. 0100011 oo1oind 0001101, Next for row 1, use row 2 with I left shift. 1000110 0100011 oorolnt 0001101 ‘This is the desired form. TT00000 ooo1oil 0010 [0010110 oo11101 0100 | 0101100 0100111 0110 | 0111010 0110001 1000 | 1011000 1010011 1001110 1000101 1110100 wiaadia 1100010 1101001 (©) All code words are at a minimum distance of 3 units. Hence this is a single error correcting code. e(s) =x? +241, Hence row 4 is 0001011. 1011000 ,,{0101100) 0010110 ooo10rl Row 4 is ok. Row 3 is left shift of row 4. For row 2, left shift row 3. ‘And add row I to obtain row 2 For row 1, left shift row 2. ‘And add row 1 to obtain row 1. 0001011 0010110 0101100 0100111 1001110 1000101 154 + rows + row3 + row2 + row!1000101 o1001i1 S"loo10110 0001011 164-1 The burst (of length 5) detection ability is obvious. The single error correcting ability can be demonstrated as follows, If in any segment of & digits a single error occur, it will violate the parity in that segment ‘Hence we locate the segment where the error exists. This error will also cause parity violation in the augmented segment. By checking which bit in the augmented segment violates the party, we can locate the wrong bit position exactly. 165-1 The code can correct any 3 bursts of length 10 or less. It can also correct any 3 random errors in each code word. 167-1 Poy = HO{ PE y/o) = 120{ 12912) = 9825 x 10-° Pee -e[d 2) (oes «920-00 To achieve a value 9872 107 for Pry, we need new value Ey/W say Eb/\. Then 9872x107 = i FE). na FE) Hence rf P| -onn a0 ws B 603 => => = 1818 atte ‘This means E,/ must be increased from 9.12 to 18.18 (nearly doubled). 155