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

# Z y z 5 z 2 z 2 z 3 c omment a nd e

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: s um of t he four inputs to t hat summer. Therefore so t hat (11.41) Moreover, Y[z], t he o utput of the second summer, is e qual t o t he s um o ffour signals t o t hat summer. Therefore C onnection between t he l aplace and t he Z- Transform We now show t hat d iscrete-time systems also c an b e analyzed using t he Laplace transform. In fact, we shall see t hat t he z -transform is the Laplace t ransform in disguise a nd t hat discrete-time systems can be analyzed as if t hey were continuoustime systems. So far we have considered t he discrete-time signal as a sequence of numbers a nd n ot as a n electrical signal (voltage o r c urrent). Similarly, we have considered a discrete-time system as a mechanism t hat processes a sequence o f n umbers (input) t o yield a nother sequence o f numbers (output). T he s ystem was built by using delays (along with adders a nd multipliers) t hat delay sequences o f numbers, n ot electrical signals (voltages o r c urrents). A digital computer is a perfect example: every signal is a sequence o f numbers, a nd t he processing involves delaying sequences of numbers (along with addition a nd m ultiplication). Consider a discrete-time system with transfer function H[z] a nd a n i nput j [k], as shown in Fig. 11.8a. We &lt;:,an t hink o f (or generate, for t hat m atter) a corresponding continuous-time signal j (t) consisting of impulses spaced T seconds a part. L et t he k th impulse o f s trength b e j[k] as depicted in Fig. 11.8b. T hus 00 i (t) = (11.42) F rom Eqs. (11.41) a nd (11.42), i t follows t hat Y[z] F[z] b3Z3 + b2Z2 + bIZ + bo z3 + a2z2 + a iz + ao T his result shows t hat Fig. 11.7 is indeed a realization of H[z] in Eq. (11.40). Similarly, the cascade a nd parallel realizations of t he continuous-time case are directly applicable t o discrete-time systems, with integrators replaced by unit delays. T he s econd canonical realization developed in Appendix 6.1 also applies t o discrete-time c ase w ith l /s replaced by l /z. V' L j[k]o(t - kT) (11.43) k=O F igure 11.8 shows j[k] a nd corresponding J(t). L et us now consider a system identical in s tructure t o t he discrete-time system with transfer function H [z], e xcept t hat t he delays in H[z] a re replaced by elements t hat delay continuous-time signals (such as voltages or currents). I f a continuous-time impulse o(t) is applied t o such a delay of T seconds, t he o utput will be o(t - T). T he continuous-time transfer function of such a delay is e - sT [see Eq. (6.54)]. Hence, t he delay elements with transfer function 1 / z in t he realization of H [z] will be replaced by t he delay elements with transfer function e - sT in t he realization of t he corresponding i I (s). T his s tep is t he s ame as z being replaced by esT. Therefore, t he t ransfer function of this system is H[z] w ith z replaced by esT. T hus i I(s) = H[e sT ]. Now whatever operations are performed by t he discrete-time system H[z] o n j[k] (Fig. 11.8a) are also performed by t he corresp...
View Full Document

Ask a homework question - tutors are online