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

Since all t he s tates a re decoupled because o f t

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: T he u pper l imit o n t he s ummation i n E q. ( 13.93a) is n onnegative. H ence k :::: 1, a nd t he s ummation is recognized as t he c onvolution s um IAI - AI =IA i -1 A-~ 1= A2 - ~A + ~ =(A - ~) (A - ~) =° 6 6 3 2 Hence, Al = l and A2 = ~ are t he eigenvalues of A and [see Eq. (13.95)] H ence + A k -1 u [k - x [k] = A kx[O] ' --v--" 1] * Bf[k] y z ero-input ( 13.93b) , where [see Eq. (B.95b)] z ero-state a nd y[k] f3o] = [ f3I = C x+Df k -1 = C A kx[O] + L C A k -l-jBf{j] + D f [1 1 l] -1 ~ [ Wk] k = W [3 -2] -6 6 [ (3)-k] =[ ( 2)-k 3 (3)-k - 2(2)-k ] _ 6(3)-k (13.94a) j =o k1 = C Akx[O] + C A - u[k - 1] * Bf{k] + D f (13.94b) I n Sec. B.6-5, we showed t hat Ak = [ 3(3)-k - 2(2)-k] [ : = [ 3(3)-k _ 2 (2)-k A k = f30I + (31 A + (32 A 2 + . .. + (3n-1 A n -1 + 6 (2)-k a nd (13.95a) ( 3)-k _ ( 2)-k :] + [ -6(3)-k + 6 (2)-k] [:i + 6 (2)-k] _ 2(3)-k + 3 (2)-k ~] - 6(3)-k (13.97) We can now determine t he s tate vector x[k] from Eq. (13.93b). Since we are interested in t he o utput y[k], we shall use Eq. (13.94b) directly. Note t hat w here ( assuming n d istinct e igenvalues o f A ) C A k =[_15]A k =[2(3)-k_3(2)-k (30 - 4(3)-k + 9 (2)-k] (13.98) a nd t he zero-input response is C A kX[O], w ith (31 (13.95b) ( 3n-1 n An x[O] 1 a nd A1, A2, . .. , An a re t he n e igenvalues of A . We c an a lso d etermine A k f rom t he z -transform f ormula, w hich will b e d erived l ater i n E q. (13.102): (13.95c) = [ :] Hence, t he zero-input response is (13.99a) T he zero-state component is given by t he convolution sum of C A k -Iu[k - 1] a nd Bf(k]. Using t he shifting property of t he convolution sum [Eq. (9.46)], we can obtain t he zerostate component by finding t he convolution sum of C A kU[k] a nd Bf(k] and t hen replacing 13 82.8 S tate-Space A nalysis 13.6 o S tate-Space A nalysis o f D iscrete- Time S ystems 829 C omputer E xample C I3.S Using MATLAB find t he zero-state response of t he system in Example 13.12. A =[O 1 ;-1/6 5 /6]; B =[O; 1]; C =[-1 5]; D =O; [ num,den]=ss2tf{A,B,C,D); k =0:25; u =ones{ 1 : length{k) ); y =filter{ n um,den, u ); s tem(k,y) 0 J [k] 1 3.6-2 F ig. 1 3.12 System for Example 13.12. The Z - Transform Solution T he z -transform o f E q. (13.91) is given b y k w ith k - 1 in t he result. We use this procedure because the convolution sums are listed in Table 9.1 for functions of t he type f [k]u[k] r ather t han f [k]u[k - 1]. zX[z] - zx[O] = A X[z] + BF[z] ( zl - A)X[zJ) = zx[O] + BF[z] T herefore C A kU[k] * B f[k] = [ 2(3)-k _ 3 (2)-k - 4(3)-k + 9 (2)-k] * [ 0] u[k] a nd = _ 4(3)-k * u[k] + 9(2)-k * u[k] X[z] * B f[k] = - 4 [1 3 -Ck+1l] -1 _ ~ u[k] +9 [1 2- +] x[k] = Z -l[(1 - z-l A)-l]x[O] + Z -l[(zl - A)-lBF[zJJ , 1] = [12 + 6 (3)-k - 18(2)-k]u[k - 1] v " , v z ero-input c omponent Now t he desired (zero-state) response is o btained by replacing k by k - 1. Hence * B f[k - ( 13.101a) u[k] = [12 + 6 (3- Ck +1l ) - 18(2- Ck +1l)]u[k] C A kU[k] (I - H ence Ck 1l -1 _ ~ ( zl - A)-lzx[O] = U sing Table 9.1 (Pair 2a), we...
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