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

1 04 1 f ig p ioi 6 1 01 7 a s ignal fkj is a

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

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ~ derive t he p air o f e quations defining t he z -transform as s hows JlkJ as a s um of everlasting exponential c omponents Like t he Laplace transform, t he z -transform is a linear operator. If F[z]zk-t dz (11.2) t hen 668 (11.3) 672 11 Discrete- Time S ystems Analysis Using t he Z - Transform T he e xistence of t he z -transform is g uaranteed if IF[zJl ~ ~ 'f;~' < 673 T he Z -Transform Therefore F[zJ = ~ [_Z_ + z - z_ ] _ jP e- 2 z - e jp 00 for s ome Izl. A~y s ignal f [k] t hat grows no faster t han a n e xponential signal r k for s ome ro, satIsfies t his condition. Thus, if 0, for some ro (11.8) z(z - cos ,8) z2 - 2z cos ,8 + 1 ( d) Here f[OJ = f[IJ = f[2J = f[3J according to E q. (11.9) = f[4J 1 t hen IF[zJl~~(ro)k t:o Izl 1 = 1- M Izl> = 1 and 1 F[z] = 1 + :; + ~ 1 f[5] 1 = f[6] = . .. = O. Therefore, 1 + ;J + ? z4 + Z 3 + z 2 + z + 1 Izl > ro z4 T herefore, F[z] e xists for Izl > roo All practical signals satisfy (11.8) a nd a re t herefore z-t~ans~ormable. Some signal models (e.g. ' "(k2) which grow faster t han t he e xponentIal sIgnal r ok (for any ro) d o n ot s atisfy (11.8) a nd t herefore are n ot z transformable. Fortunately, such signals a re o f little practical o r t heoretical interest. • 11.1 We can also express this result in a closed form by summing the geometric progression on the right-hand side of the above equation, using the formula in Sec. B.7-4. Here the common ratio r = ~, M = 0, and N = 4, so t hat E xample 1 1.2 F ind the z-transforms of ( a) a[kJ ( b) u[kJ (c) cos ,8ku[kJ ( d) signal shown in Fig. 11.2. Recall that by definition F[zJ = L JlkJz-k k=O = f[OJ + f[IJ + f[2J + f[3J + ... Z ( a) For f[kJ = a[kJ, f[OJ = 1 and f[2J a[kJ = z3 z2 = (11.9) f[3J = f[4J = . .. = O. Therefore for all z 1 (11.10) ( b) For f[kJ = u[kJ, f[OJ = f[IJ = f[3J = . .. = 1. Therefore F[zJ 1 1 1 = 1 + - + - 2 + - + ... Z z z3 1 1- - z z z -1 I~I < 1 Izl > 1 Therefore u [kJ=-z(11.11) Izl > 1 z-1 ( c) Recall t hat cos,8k = (e jPk + e- jPk ) / 2. Moreover, according to Eq. (11.7), 3 o 4 456789 Fig. 11.3 Fig. 11.2 D. E xercise E ll.l ( a) Find the z-transform of a signal shown in Fig. 11.3. ( b) Using Pair 12a (Table 11.1) find the z-transform of f[k] = 20.65( y '2)k cos ( tk - 1.415) u[k]. A From E q. (11.6) it follows t hat F[zJ= - -1 2 ( nswers: ( ) F a z= ( b) z (3.2z + 17.2) z 2-2z+2 11.1-1 z5 + z4 + + z39 + z2 + z + 1 or z z ( -4 - 10) z -l z -z \l Finding the Inverse Transform As i n t he L aplace transform, we shall avoid t he i ntegration in t he c omplex plane required t o find t he inverse z-transform [Eq. (11.2)] b y using t he ( unilateral) t ransform T able. M any o f t he t ransforms F[z] o f practical i nterest a re r ational f unctions (ratiO o f polynomials in z ). Such f unctions c an b e e xpressed as a s um o f simpler functions using p artial f raction expansion. T his m ethod works because for every t ransformable f [k] defined for k 2: 0, t here is a corresponding unique F[z] defined for Izl > ro (where ro is some c onstant), a nd vice versa. 674 11 D iscrete-Time S ystems A nalysis U sing t he Z -Transfo...
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