Modern Control Engineering

(Chris Devlin) #1

Section 10–4 / Design of Servo Systems 745


Define


Then Equation (10–37) can be written as


(10–38)


where


(10–39)


Define a new (n+1)th-order error vector e(t)by


-vector


Then Equation (10–38) becomes


(10–40)


where


and Equation (10–39) becomes


(10–41)


where


The state error equation can be obtained by substituting Equation (10–41) into


Equation (10–40):


(10–42)


If the desired eigenvalues of matrix (that is, the desired closed-loop poles) are


specified as m 1 ,m 2 ,p,mn+1,then the state-feedback gain matrix Kand the integral


gain constant kIcan be determined by the pole-placement technique presented in Section


10–2, provided that the system defined by Equation (10–40) is completely state


controllable. Note that if the matrix


has rank n+1,then the system defined by Equation (10–40) is completely state


controllable. (See Problem A–10–12.)


B


A


- C


B


0


R


Aˆ -BˆKˆ


e# =AAˆ -BˆKˆBe


Kˆ =CK-kID


ue=-Kˆe


Aˆ = B


A


- C


0


0


R, Bˆ = B


B


0


R


e# =Aˆe+Bˆue


e(t)= B


xe(t)


je(t)


R =(n+1)


ue(t)=-Kxe(t)+kI je(t)


B


x



e(t)


j



e(t)


R = B


A


- C


0


0


RB


xe(t)


je(t)


R + B


B


0


Rue(t)


u(t)-u(q)=ue(t)


j(t)-j(q)=je(t)


x(t)-x(q)=xe(t)

Free download pdf