Modern Control Engineering

(Chris Devlin) #1

Section 9–6 / Controllability 679


we can rewrite Equation (9–58) as


If the elements of any one row of the n*rmatrixFare all zero, then the corresponding


state variable cannot be controlled by any of the ui.Hence, the condition of complete


state controllability is that if the eigenvectors of Aare distinct, then the system is com-


pletely state controllable if and only if no row of has all zero elements. It is im-


portant to note that, to apply this condition for complete state controllability, we must


put the matrix in Equation (9–58) in diagonal form.


If the Amatrix in Equation (9–56) does not possess distinct eigenvectors, then


diagonalization is impossible. In such a case, we may transform Ainto a Jordan canonical


form. If, for example,Ahas eigenvalues and has n-3distinct


eigenvectors, then the Jordan canonical form of Ais


The square submatrices on the main diagonal are called Jordan blocks.


Suppose that we can find a transformation matrix Ssuch that


If we define a new state vector zby


(9–59)


then substitution of Equation (9–59) into Equation (9–56) yields


(9–60)


The condition for complete state controllability of the system of Equation (9–56) may


then be stated as follows: The system is completely state controllable if and only if (1)


=Jz+S-^1 Bu


z# =S-^1 ASz+S-^1 Bu


x=Sz


S-^1 AS=J


J=I


l 1


0


0


0


1


l 1


0


0


1


l 1


l 4


0


1


l 4


l 6











0


ln


Y


l 1 ,l 1 ,l 1 ,l 4 ,l 4 ,l 6 ,p, ln


P-^1 AP


P-^1 B


z#n=ln zn+fn1 u 1 +fn2 u 2 +p+fnr ur











z



2 =l 2 z 2 +f 21 u 1 +f 22 u 2 +p+f2r^ ur


z



1 =l 1 z 1 +f 11 u 1 +f 12 u 2 +p+f1r^ ur

Free download pdf