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

# 1 a constant input fkj c t his is a special case of

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: nit circle (Fig. 9.8d). T he s ystem is unstable. • E xercise E9.13 Find and sketch the location in the complex plane of the characteristic roots of the system specified by the following equation: (E + 1)(E2 + 6E + 25)y[k] = 3Ef[k] f::, Is this a stable, unstable, or marginally stable system? Answer: unstable. f::, E xercise E9.14 Repeat Prob. E9.13 for (E - 1)2(E + O.5)y[k] = (E2 Answer: unstable. 9.6-1 \l (d) ( e) * f [k] 00 = L h [m]f[k - m] m =-oo a nd 00 ::; L I h[m]llf[k-mJl m =-oo + 2E + 3)f[k] \l System Response t o Bounded Inputs A s i n t he c ase o f c ontinuous-time s ystems, a symptotically s table d iscrete-time s ystems h ave t he p roperty t hat e very b ounded i nput p roduces a b ounded o utput. I f f [k] is b ounded, t hen I /[k - mJl < K 1 < 0 0, a nd 00 m =-oo C learly t he o utput is b ounded i f t he s ummation o n t he r ight-hand s ide i s b ounded; t hat is, if 608 9 Time-Domain Analysis o f Discrete-Time Systems 00 L /h[kJl < K 2 < 00 (9.75) 609 9.7 Appendix 9.1: Determining t he Impulse Response passed through a discrete-time system. T he a mount of dispersion (or spreading out) is e qual t o t he s ystem t ime c onstant (or w idth o f h[kJ). k =-oo 9 .7 For a n a symptotically stable system h[k] = 9 .6-2 For a discrete-time system specified by Eq. (9.27), we have demonstrated t hat ( cnf + C2/~ + " . + cn)'~)u[k] w ith t he c haracteristic roots ) '1> ) '2, . . . ) 'n lying within t he u nit circle. Under these conditions h[k] satisfies Eq. (9.75).t Therefore, t he response of a n a symptotically s table system to a bounded input is also bounded. Moreover, we c an show t hat for a n u nstable or marginally stable system, t he response is u nbounded for some bounded input. As seen in Sec. 2.6-1, these results lead t o t he formulation of an a lternate definition of system stability. A system is said t o b e stable in t he sense of having bounded o utput for every bounded i nput (BIBD s tability) if a nd only if its impulse response h[k] satisfies Eq. (9.75). Condition (9.75) is sufficient for a system t o produce bounded o utput for any bounded input. I t is relatively easy to show t hat i t is also a necessary condition; t hat is, for a system t hat violates (9.75), there exists a bounded input t hat p roduces unbounded o utput (see P rob. 9.6-2). Note t hat a n a symptotically stable system is always BIBD-stable.j: Appendix 9.1: Determining t he Impulse Response h[k] = Ao8[k] + Yn[k]u[k] (9.76) We now show t hat Ao = bo ao (9.77) To prove this point, we s ubstitute Eq. (9.76) in Eq. (9.28) t o o btain Q [E] (Ao8[k] + Yn[k]u[kJ) (9.78) = P [E]8[k] Because Yn[k]u[k] is a s um o f c haracteristic modes,j: [see Eq. (9.12a)] Q [E] (Yn[k]u[kJ) = 0 k~0 (9.79) E quation (9.78) now reduces t o A oQ[E]8[k] = P [E]8[k] Intuitive Insights Into S ystem Behavior or T he i ntuitive insights into t he behavior o f continuous-time systems a nd their qualitative proofs, discussed in Sec. 2.7, also apply t o d iscrete-time systems. F...
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