Beruflich Dokumente
Kultur Dokumente
If the initial voltage in capacitor 2 is zero no matter what input u is applied the mode e!t in loop II can never be exited" #he presence of e!t of loop I can never be detected from the output"
$%&
Example: The Modal Canonical Form Consider a single input single output state space system defined by
x = Ax + Bu
'
Suppose ( has distinct eigenvalues )$ * )n" #han ( can be diagonalized by the following Similarity #ransformation
x = Px
where P + V-1 ,V is an eigenvector matrix- so the system ,'- is e.uivalent to the following system in modal Canonical form
x & $ $ x2 & 2 = & & xn
A = PAP $
0 = /0
x k is
1ow if
02$
uncontrollable"
Controllability of 3inear #ime 4 Invariant Systems Consider the n!dimensional p!input state e.uation
x = Ax + Bu
'
$%$
Definition: #he state e.uation ,'- or the pair ,( 0- is said to be controllable if for any initial state x,&- + x& and any final state x$ there exists an input that transfers x& to x$ in a finite time" Otherwise ,'- or ,( 0- is said to be uncontrollable" emar!: (ccording to the definition for a controllable system the input must be able to steer the state of the system from x& at time t& to x$ at time t$" #he exact tra5ectory followed is not important" Theorem: #he n!dimensional linear 4 invariant system ,'- is controllable if and only if any of the following conditions is satisfied6 $" #he Controllability 7ramian
Wc = e A BB T e A d
& t
T
(20 *
(n!$0<
, t$ t -
& 2u $
Solution6 : + ;0 (0
$ U = & $ & & 2 $ $ $2
(20<
2 $ D 2% C
>,:- + =
system is controllable
$%=
= $ +@
#hen characteristic e.uation is )2 A @ ) A = + & ) + !$ != Bor the system to be controllable >; )I 4 (6 0< + n Bor ) + !$
! $ ! $ ;$ I A 6 0< = = =
&
So if D + ! FC Bor ) + !=
! = ! $ ;$ I A 6 0< = = =
$ $
&
D E !C
So if D + ! C the system is uncontrollable Gence for the system to be controllable $" D E !FC 2" D E !C
Controllability "ndices 3et ( and 0 be n 9 n and n 9 p constant matrices" Suppose the pair H( 0I is controllable then the controllability matrix : + ;0 (0 columns in :" It may be that we can find n such columns in the partial controllability matrix :2 + ;0 (0 (20 * (2!$0< $ J 2 J n" (20 * (n!$0< has ran2 n which mean that there are n linearly independent
$%K
Definition: #he smallest 2 say L such that :2 has ran2 n will be called the controllability index of H( 0I" Corollary: #he pair H( 0I is controllable iff the n 9 ,n ! matrix
U n p =; B (B """ ( n ! p B <
p A$-p
n p
' Un p
is nonsingular
$ & & 2
>,:@!2- + @
so H( 0I is controllable
$%?
1otice that >;0 6 (0< + @ the controllability index is 2 Mhat is the natural way to chec2 linear independenceN 3et us loo2 again on the matrix : : + ;0 (0 Or : + ;b$ b2 * bp6 (b$ (b2 * (bp6 (2b$ (2b2 * (2bp6 6 * 6 (n!$b$ (n!$b2 * (n!$bp< 1ow we search the linearly independent from left to right" 3et us rearrange the linearly independent columns as (20 * (n!$0<
[b
Ab$
"""
A $ $b$ 6 b2 """ A
p $
Ab2
"""
A 2 $b2
""" 6 b p
Ab p
bp
#he integer Li is the number of linearly independent columns associated with bi in the above set" Clearly L + maxHL$ L2 * LpI
$%O
#he set HL$ L2 * LpI will be called the controllability indices of H( 0I" Theorem: #he controllability property is invariant under any e.uivalence transformation" /roof6 Consider the pair H( 0I with the controllability matrix : + ;0 (0 (20 * (n!$0< and its E.uivalent /air , A B - with A = PAP and B = PB where / is a nonsingular matrix" #he controllability
$
B-
is
AB
n $
B<
PB
PAP $ PB
PA n $ P $ PB
+ ; PB
PAB
$ PA n B<
+ P ;B AB
$ A n B
< + /:
$%%
Theorem: #he set of the controllability indices of ,#$ %- is invariant under any e.uivalence transformation any reordering of the columns of %&
$%P