Modern Control Engineering

(Chris Devlin) #1

Section 9–2 / State-Space Representations of Transfer-Function Systems 653


(9–12)


the transformation x=Pz, where


P=


l 1 ,l 2 ,p,ln=ndistinct eigenvalues of A


will transform P–1APinto the diagonal matrix, or


If the matrix Adefined by Equation (9–12) involves multiple eigenvalues, then


diagonalization is impossible. For example, if the 3*3matrixA, where


has the eigenvalues l 1 ,l 1 ,l 3 ,then the transformation x=Sz, where


will yield


This is in the Jordan canonical form.


S-^1 AS=C


l 1


0


0


1


l 1


0


0


0


l 3


S


S= C


1


l 1


l 12


0


1


2 l 1


1


l 3


l 32


S


A=C


0


0


- a 3


1


0


- a 2


0


1


- a 1


S


P-^1 AP= F


l 1


0


l 2











0


ln


V


G


1


l 1


l 12











l 1 n-^1


1


l 2


l 22











l 2 n-^1


p


p


p


p


1


ln


ln^2











lnn-^1


W


A= G


0 0    0


- an


1 0    0


- an- 1


0 1    0


- an- 2


p


p


p


p


0 0    1


- a 1


W

Free download pdf