This preview shows page 1. Sign up to view the full content.
Unformatted text preview: +1 ( b) ~ s I F ig. 1 3.9 S ystems for E xample 13.11.
tWe can show t hat a system is completely controllable if and only if the n x n j composite matrix
[B, A B, A 2B, . .. , A nIB] has a rank n . Similarly, a system is completely observable if and
only if the n x n k composite matrix [C', A 'C', A ,2C', . .. , A mIC'] has a rank n . 8 20 X 2. 13 S tateSpace Analysis
I n b oth cases, t he s tate v ariables are identified a s t he t wo i ntegrator o utputs, x 1 a nd
T he s tate e quations for t he s ystem i n Fig. 13.9a a re 13.5 Controllability And Observability 821 T he first column o f C is zero. Hence t he m ode Z l ( corresponding t o A l = 1) is unobservable.
T he s ystem is therefore controllable b ut n ot o bservable. We come t o t he s ame conclusion
by realizing t he s ystem w ith t he s tate v ariables Z l a nd Z 2, w hose s tate e quations a re z= a nd Az y= (13.79) Cz + Bf A ccording t o Eqs. (13.80) a nd (13.81), we have H ence s 1 l sIAI
T herefore = o I = (s  I 1 a nd 1)(s + 1) y= s +1 Z2 a nd
a nd f
(13.80) (a) y W e s hall now use t he p rocedure in Sec. 13.41 t o d iagonalize t his s ystem. According t o
E q. (13.74b), we have [: :] P12 ] [ PH
P21 [ PH
P21 Pn P12] [ 1
P22 1 :] T he s olution o f t his e quation yields
P12 C hoosing PH =0 a nd  2P21 = P22
( b) = 1 a nd P21 = 1, we have
p= [: f ~2] a nd
B =PB=
A ll t he rows o f B a re nonzero. C~2] [ :] F ig. 1 3.10 C] y E quivalent o f t he s ystems i n Fig. 13.9. (13.81a)
F igure 13.10a shows a realization o f t hese e quations. I t is clear t hat e ach o f t he t wo modes
is controllable, b ut t he first m ode ( corresponding t o A = 1) is n ot o bservable a t t he o utput.
T he s tate e quations for t he s ystem i n Fig. 13.9b a re Hence t he s ystem is controllable. Also, Y =Cx
= C p 1 z (13.82) = Cz (13.81b) a nd
y a nd o
2
1 = X2 Hence
= [1 21 [ ; _ :] = [0 11 (13.81c) C =[O I I, D =O 13 S tateSpace A nalysis 8 22 s l sIAI = +1 I = (s + 1)(8  o I 1 8 1 3.6 S tateSpace A nalysis o f D iscrete Time S ystems 823 A hat=P* A * inv(P);
B hat=P*B
C hat=C*inv(P)
0 1) 1 so t hat Al =  1, A2 = 1, and 1 3.51 Inadequacy o f t he Transfer Function Description o f a S ystem (13.83)
Diagonalizing the matrix, we have [1 0]
o
1 [ PH
P21 P12] =
P22 T he solution of this equation yields PH
P21 = 1, we o btain [PH
P21 = P12] [1 0]
2 P22  P12 a nd P22 E xample 13.11 d emonstrates t he i nadequacy o f t he t ransfer f unction t o d escribe
a n L TI s ystem i n g eneral. T he s ystems i n F igs. 13.9a a nd 1 3.9b b oth h ave t he s ame
t ransfer f unction
1
H(8) = 8+1 = 1 O. Choosing P H =  1 and Y et t he t wo s ystems a re v ery different. T heir t rue n ature is revealed i n F igs. 13.lOa
a nd 1 3.lOb, respectively. B oth t he s ystems a re u nstable, b ut t heir t ransfer f unction
H(8) =
d oes n ot give a ny h int o f i t. T he s ystem i n F ig. 13.9a a ppears s table
f rom t he e xternal t erminals, b ut i t is i nternally u nstable. T he s ystem i n F ig. 13.9b,
o n t he o ther h and, will show i nstability a t t h...
View
Full
Document
 Spring '13
 Bayliss
 Signal Processing, The Land

Click to edit the document details