4.12 Signals and Systems As a limit . 7 a iting case, At becomes continuous variable dt and the summation jy integral. Therefore the representation of x(¢) becomes 7 x(0= { x(t) 6(¢-4t 4.36, arbitrary input signal can be represen .d shifted unit impulse functions, 1p th then the response of the S¥Sten, The importance of Eq. (4.36) is that an as a linear combination of the scaled an response of the system to the impulse is known, for any input signal x(t) can be obtained. Review Questions 1) can be represented as a Linear comp, 1. Show that‘an arbitrary signal x( ‘unit impulse functions. nation of the scaled and shifted ‘convolution Integral initially relaxed att = 0. If the input to the S¥ste Consider a LTI system which is ed at is an impulse; then the output of the system iS denoted by h(t) and is calleg n impulse response of the system. We can denote the A(t) = T(6()) (3p ML systerm x a hn Fig. 4.3 AnLTI system. We know that any arbitrary signal x(t) can be represented as x(t)= | x(1)6(¢—2)de an The system output is given by v(t) =T{x(s)] 43)Time Domain Analysis of Continuous-Time Systems 4,13 substituting Eq. (4.38) in Eq, (4.39) yields y)=T [Zo - ve (4.40) For linear system y(t) = / x(t)T(8(¢—1)]de (41) If the response of the system due to impulse 6(1) is h(t) then the response of the system due,to delayed impulse is A(t,t) = T[5(t-7)] (4.42) Substituting Eq. (4.42) in Eq, (4.41) we get w= f x(t) h(t, t)dt (4.43) For a time-invariant system the output due to delayed input by T is equal to de- layed output by 7. That is A(t,t) =A(t—1) (4.44) Substituting Eq. (4.44) and Eq. (4.43) we get w= fatane—ode (4.45) This is called convolution integral, or simply convolution. The convolution of two signal x(t) and A(t) can be represented as Y(t) =x(t) +h(¢) 48° Properties of Convolution Let us consider two signals x, (t) and x(t). The convolution of two signals x(t) and x(t) is given by the equation x(t) #x2(t) = f xi(t) xa(t—t)dt (4.46) 1, Commutative property: Convolution obeys commutative property. That is 21 (t) #x2(t) =x2(t) #x4(1) (4.47)414 gy '4 Signals and Systems Proof We have a@en()= [s(2)n0—aar Let t-t=p then —dt=dp Substituting these values in Eq. (4.48) we get 24 (t)#x2(t) = - / 22(p) x(t p)dp 7 [20)n(t- pap =nien() >. [n@n() =n()en 2. Distributive property: 21(t)* a(t) +230] =21(1) +2214) +211) +x5(4) 3. Associative property: x(t) [x2(¢) #33(¢)] = b(t) #x2(0)] #x3(1) 4, Shift property: If x(t) *x2(t) =2(¢) then X(t) *x2(t-T) =2(¢-T) Proof n(Q*n(-7) = [at)ne-r-nae : =2x(t-T)Time Domain Analysis of Continuous-Time Systems 4.15 Similarly ai(t=T) #x,(0) =2(1- 17) (4.54) and xy (t-T))an(t-) =2(t- 1 - B) (4.55) 5, Convolution with an impulse: Convolution of a signal x(t) with a unit impulse is the signal x(t) it self. That is x(t)+8(0) =x(¢) (4.56) Proof = 0 otherwise x)#8() = / (2)80-Hae 8(t-1) =1fort=t x) 6. Convolution with shifted impulse Convolution of a signal x(t) = shifted impulse 5(¢ —%) is equal to x(t—f). That is x(* a to) =x(t-t0) (457) 3(0)+8(t-6) = | x(1)5(t-1-t9) dt =x(1) [retry =2(t-f0) 7. Convolution with unit step Convolution of a signal x(t) with unit step signal u(¢) is given by t x(t)#u(t) = J x(t) dt (4.58)4.16 Signals and Systems Proof x(t) #u(1) = I x(t)ut—1)4t ! = [xa arate ifort s 8. Convolution with shifted unit step | Convolution of a signal x(¢) with shifted unit step signal is given by fo x(t) «u(t 40) = [xe 4. 9. Width property: Let the duration x; (#) and.x2 (#) are T; and Th respectively. Then the duratiog of the signal obtained by ‘convolving x1(t) and x2(0) is T+B. Review Questions 1. Define convolution integral. 2, Explain in detail about the properties of convo! 3. Show that (a) x(t) ¥5(0) =a(4) () x(4)#8(t- = x(t-to) (©) x(s)eu(t) = J x(t)dt lution. Impulse Response of Interconnected Systems 4.6.1 Systems in parallel Consider two LTI systems with impulse responses My (t) and ha(t) connect parallel as shown in Fig. 4.4. The output of system | vit) =x) #hi(t)Time Domain Analysis of Continuous-Time Systems 4.17 yi(t) {ono | o He hw ya) ‘system 2 Fig. 4.4 Parallel connection of continuous - time systems. ‘The output of system 2 wile) = xl) hale) (461) ‘The overall output v(t) =yi(t) +yr(¢) (4.62) = x(t) *hi(e) +24) #2) = fx ahi tonart [te ing(t — t)d = f x(t) (n(t—1) + hn(t— a) dt = fo ht—t)dt =x() *h(t) where h(t) is given by A(t) = Ay(t) + Ao(t) (4.63) « That is the impulse response of two systems connected in parailel is sum of the individual impulse responses. 4.6.2 Systems in cascade Let us consider two systems with impulse responses hy (t) and z(t) connected in cascade as shown in Fig. 4.5. The output of system 1 is a(t) = x(t) + hi (2) = / x(t)iy(t— dt (4.64)4.18 Signals and Systems xit) we ‘ oh, * bt) 7 nC) System | ‘system 2 yume. Fig. 45 Cascade connection of continuous time The output of system 2 is y(m) = yr (em) hal) « [one «hn(m) = / Jaome-omser (465) Let k-t=4 won) = f x0) faster ayaa at = f x(yamn— tat : (466 = x(m)+A(rm) where aim) = faladin—aea aa = hn(m) #alm) 463} ‘Therefore, the impulse response of two systems connected in series is equal to te convolution of the individual responses. Review Questions 1. Show that the impulse of two systems connected in convolution of individual responses. 2, ‘Show that the impulse response of two syste! ‘equal to the sum of individual responses. series is equal to the mm connected in parallel isTime Domain Analysis of Continuous-Time Systems 4.19 4. .7_easuality In section (3.10) we introduced the causality condition on impulse response for discrete time systems. On the similar lines we can find that for.a causal LTT system, the impulse response h(r) must be zero fort < 0. Consider a continuous-time LTT system whose output y(?) can be obtained using convolution integral given by y(t) = / A(t)x(t—t)dt (4.69) From Eq. (4.69) we can find that the output y(r) depends on past, present and future values of input. If t > 0, the output depends on present and past values of input. If t < 0, then output depends on future values of input. But for a causal system the output depends only on present and past values of input which implies that h(t) =0 for t <0.. Therefore the limits of the integral can be changed as follows. y= [ A(t)x(t—t)dt (4.20) 0 ‘An LTI continuous system is causal if and only if its impulse. response is zero for negative values of t. Ifa causal signal is applied to a non-causal system, then y(t) = f x(t)A(t= tat 0 ; = | A(tyx(t—1)de (a7) If a causal signal is applied to a causal system, then t vt) = | x()Mt—1)de 0 ‘ = J A(a)x(t—t)dt (472) 3-4.20 - Signals and Systems Review question . =o fort < ia causal system the impulse h(t) =0 for Solved Problem 4.3 Find the convolution of xi(¢) and x(t) for the following signals : @ uQ=euy: n= e Mult) Gi) (Haul); 2) = u(t) Git) n(Q=uQ; a) =" (iv) 2x1(t) = sintu(t); x(t) =H) Solution: (i) Given x(t) = etu(t), 2 We know n()en(t) = | ay (t)xx(t— 1)4T A signal x(f) is said to be causal if x(t) = Ofort <0. Inthis case both signals x1 (¢) and x)(t) are causal. Therefore using Eq. (47, we have , si(d)asa(t) = f x(ealt~ tat 0 z fen elder 0 1 aot | eltge oO et " = —(a—b)t =(a—b)° 0 e401] boa for 1>0Time Domain Analysis of Continuous-Time Systems 4.21 + Gi) Given x (1) = u(t) and x2(¢) = u(r) Both signals are causal. Therefore the convolution of x1(¢) and-x2(¢) is given by x(x) = fi) x(t —t)dt + u(t)= fort >0 u(t—t)=1 fort >t Gi) i() seus); 22) = u(r) Both signals are causal. Therefore ' nent) = [meal or 2 z for t2>0 a t 2 [r= 7 0 2 sue) (iv) x1(t) =sintu(t); — x2(t) = u(t) f sint dt a " ai(Qen(e = —cost =(l-cost) fore >0 lo = (1—cosr)u(t)