Signal Processing and Linear Systems-B.P.Lathi copy

# 1 15 f ig 1 17 a c anonical r ealization o f hz r

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ontinuous-time u nit impulse 6(t). T hus, a t t he i nput of a discrete-time processor, a discrete-time signal I [k] is j ust a sequence of numbers. B ut a t t he i nput of a continuous-time system, I [k] is a sequence o f impulses. There are appropriate converters a t t he interface o f discrete-time a nd continuous-time systems t o c arry o ut signal conversion t o a ppropriate forms. To begin with, consider a basic continuous-time system (Fig. l1.lOa) with transfer function H (s). T he i nput I {t) is sampled a nd t he sampled signal I [k] is applied t o t he i nput o f H (s). A lthough y(t), t he o utput o f t his system, is continuous, we shall endeavor t o find t he values of y(t) only a t t he discrete i nstants t = kT. Such a n analysis is relatively simple using t he m ethod of z-transform. For this purpose, we consider as if t he o utput is sampled by a n h ypothetical sampler shown d otted in Fig. 11.10a. Now, we shall relate t he i nput samples I [k] a nd t he o utput samples y[k]. L et h[k] b e t he u nit impulse response relating t he o utput samples to t he i nput samples. In o ther words, y[k] = h[k] * J[k]. Recall also t hat a n u nit impulse 6[k] is 6(t) when considered in conjunction with a continuous-time system. Hence, h[k], t he u nit impulse response is t he s ampled version o f t he s ystem's u nit impulse response h(t). T hus, h[k] = h(kT) where T is t he sampling interval. For instance, if H (s) = s~A' t hen h(t) = e At a nd Therefore, t he equivalent discrete-time transfer function H [z] of this system is given by H[z] = Z {h[k]} = Z [e AkT z = z - eAT (11.49) 700 11 Discrete- Time S ystems Analysis U sing t he Z -Transform 11.6 S ampled-Data ( Hybrid) Systems 1 T[z] = - ( - Ylz] = H[z]F[z] z z -1 2 - z ) f. G[z]H[z] z -e -2T I n t his case, we use t he n otation GH[z] for T[z]. T hus, GH[z] f. G[z]H[z], b ut is t he d iscrete-time t ransfer f unction which corresponds t o G(s)H(s). For t he s ystem i n Fig. 11.1Oc, Ylz] = GH[z]F[z] Y[z] = H [z]X[z] = H[z]G[z]F[z] so t hat T[z] = G[z]H[z] For t he s ystem i n Fig. 11.1Od, Ylz] = G[z]H[z]F[z] E[z] = F [z]- W[z] Moreover, y (t) T .:&gt;' y [k] W[z] = H[z]Y[z] Y[z] = G[z]E[z] = G[z] ( F[z]- W[z]) = G [z](F[z]- H[z]Y[z]) I --i---t-. ..... .. . ~ .... G[z] Ylz] = I + G[z]H[z] F[z] Hence F ig. 1 1.10 G[z] Y[z] = 1 + G [z]H[z{[z] C omputing t he o utput i n h ybrid o r s ampled-data s ystems. T hus, H[z] is t he d iscrete-time t ransfer f unction of H (s) = s~&gt;. t hat r elates y[k] ( the o utput s amples) t o t he d iscrete-time i nput J[k].t I f we have two systems w ith t ransfer f unctions G (s) a nd H (s) in cascade (Fig. 1 l.lOb), t he equivalent transfer T[z] f. G[z]H[z], b ut is GH[z], where G[z], H[z] a nd G H [z] c orrespond t o d iscrete-time t ransfer functions of G (s ), H (s) a nd G (s ) H (s ), respectively. For instance, if 1 G(s) = s+2 a nd z - 2T z -e G[z] T[z] = 1 + G[z]H[z] • E xample 1 1.9 F ind t he o utput s amples y[k] for t he s ampled-data s ystem i llustrated i n F ig. I lolla w hen t he i nput i s a u n...
View Full Document

## This note was uploaded on 04/14/2013 for the course ENG 350 taught by Professor Bayliss during the Spring '13 term at Northwestern.

Ask a homework question - tutors are online