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

# 135b vl 12v3 10 19v3 v2 vl v3 2 1318a v

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: i jth e lement Hij(S) o f H (s) is t he t ransfer f unction t hat r elates t he o utput Yi(t) t o t he i nput f j(t). • ( 8+1):<8+2) t li = (13.38b) a nd t he z ero-state r esponse is 803 (13.41) I t is i nteresting t o o bserve t hat t he d enominator o f e very t ransfer f unction in Eq. (13.42) is (s + l )(s + 2) w ith t he e xception o f H21(S) a nd H22(S), w here t he c ancellation o fthe f actor (s + 1) occurs. T his f act is n o c oincidence. We see t hat t he d enominator o f e very e lement o f . (s) is I sI- AI b ecause . (s) = ( sl- A)-I, a nd t he i nverse o f a m atrix h as i ts d eterminant i n t he d enominator. S ince C , B , a nd D are m atrices w ith c onstant e lements, we see from E q. (13.38b) t hat t he d enominator o f . (8) will also b e t he d enominator o f H (s). H ence, t he d enominator o f e very e lement o f H (s) is l sI - AI, e xcept for t he p ossible cancellation o f t he c ommon f actors m entioned e arlier. I n o ther w ords, t he p oles o f all t ransfer f unctions of t he s ystem a re also t he z eros o f t he p olynomial l sI - AI. Therefore, t he z eros o f t he p olynomial l sI - AI a re t he c haracteristic roots o f t he system. H ence, t he c haracteristic r oots o f t he s ystem a re t he r oots o f t he e quation l sI - AI = 0 ( 13.43a) a 13 S tate-Space Analysis 804 Since I sI- AI is a n n th-order p olynomial in s w ith n zeros Al, A2, . .. , An, we c an w rite Eq. (13.43a) as l sI - AI = = sn + n a n_lS +! + . .. + a lS l sI - AI = A = [: (13.43b) = s2 2 0 [2 (13.44a) (13.44b) = ( s+l)(s+2) a nd 1] t = [0 1 2t t] t A 2=-2 1] [0 1 2 1 ]::, = [ 2 12 2 1 ]::, = [ t2 32 t2 C le Alt + C2eA2t + . .. + c ne Ant (13.45) (13.51a) (13.51b) Note t hat t he i nfinite series o n t he r ight-hand side of Eq. (13.51a) also m ay b e e xpressed as T his fact is also obvious from E q. (13.38). T he d enominator of every element of t he z ero-input response m atrix Cf1(s)x(O) is l sI - AI = (s - Al)(S - A2)··· (s An). Therefore, t he p artial-fraction expansion a nd t he s ubsequent inverse Laplace transform will yield a zero-input component of t he form i n E q. (13.45). d -e dt At [ A2t2 A 3t 3 I + A t + - - + - - + . .. + . . . A 2! 3! Hence 13.3-2 Time-Domain Solution o f S tate Equations ~eAt = A e At = e At A T he s tate e quation is (13.52) dt (13.46) X = Ax+Bf (13.50) a nd so on. We c an s how t hat t he infinite series i n E q. (13.48) is absolutely a nd u niformly convergent for all values o f t . Consequently, i t c an b e differentiated or i ntegrated t erm b y t erm. T hus, t o find ( djdt)e At , we differentiate t he series o n t he r ight-hand s ide o f E q. (13.48a) t erm b y term: Equation (13.43) is known as t he c haracteristic e quation o f t he m atrix A , a nd Al, A2, . .. , An a re t he c haracteristic roots o f A . T he t erm e igenvalue, meaning " characteristic value" i n German, is also commonly used in t he l iterature. Thus, we have shown t hat t he c haracteristic roots of a system are t he eigenvalues (chara...
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