You are on page 1of 19

Outline

Modern Control systems


Lecture-7 Observability

V. Sankaranarayanan

V. Sankaranarayanan Modern Control systems


Outline

Outline

1 Observability

V. Sankaranarayanan Modern Control systems


Observability

Observability

The state space equations of the system is

ẋ = Ax + Bu (1)
y = Cx + Du (2)

where
x ∈ Rn ; u ∈ Rm ; y ∈ Rp
A ∈ Rn×n B ∈ Rn×m
C ∈ Rp×n D ∈ Rp×m

Definition
A system of the form (1) is said to be observable if the initial state x(0) can
be computed from the knowledge of the input u(t) and output y(t) for a given
period of time T
Observability is concerned with what can be said about initial state when
given measurements of the plant.

V. Sankaranarayanan Modern Control systems


Observability

Observability

Definition
The state equation is said to be observable if for any unknown initiaal state x(0),
there exists a finite t1 > 0 such that the knowledge of the input u and the input y
over [0.t1 ] suffices to determine uniquely the initial state x(0). Otherwise, the
equation is said to be unobservable

Observability-Rank condition
C
 
 CA 
 CA2 
 
The system of the form (1) is observable if and only if the rank of O=
.

 
 . 
CAn−1

is n (order of the system)

V. Sankaranarayanan Modern Control systems


Observability

Observability

Rank condition
y = Cx + Du
ẏ = CAx + CBu + Du̇
ÿ = CA2 x + CABu + CB u̇ + Dü ......
y C u
     
 
 ẏ   CA  D 0 0 0 . .  u̇ 
2
 ÿ   CA   CB D 0 0 . . ü
     
= xt + 
 
 .   . CAB CB D 0 . .  .
      
 
 .   .  . . . . . .  . 
y n−1 CAn−1 un−1
y = [O]x(t) + [H]u(t)
x(t) = O−1 [y(t) − [H]u(t)]
[O]−1 should exist → Rank of O should be n

V. Sankaranarayanan Modern Control systems


Observability

Observability

Example-1
Consider the state and output equations,

ẋ1 = −x1 + x2
ẋ2 = −3x2 + u
y = x1 + x2

Let us try to check for Observability of this system by transforming these


equations into different forms.
Representing these equations in matrix form , we get

      
ẋ1 −1 1 x1 0
= + u
ẋ2 0 −3 x2 1
 
  x1
y = 1 1
x2

V. Sankaranarayanan Modern Control systems


Observability

Observability

Example-1
We know that the Observability matrix for a system having two state
variables is given by

 
C
V =
CA 2×2

 
1 1
∴V =
−1 −2

Rank of matrix V = 2
Therefore, the system is Observable.

V. Sankaranarayanan Modern Control systems


Observability

Observability

Example-1
Now, let us check the Observability of the system by transforming it into
diagonal canonical form.
The transformation matrix P is the modal matrix formed by the eigenvectors
corresponding to the eigenvalues of matrix A.
 
1 1
Matrix P =
0 −2
Transformed matrices Az ,Bz and Cz are obtained using relations
Az = P −1 AP , Bz = P −1 B, Cz = CP
   
−1 0 1/2  
Az = Bz = Cz = 1 −1
0 −3 −1/2

V. Sankaranarayanan Modern Control systems


Observability

Observability

Example-1
Representing them in matrix form , we get

      
ż1 −1 0 z1 1/2
= + u
ż2 0 −3 z2 −1/2
 
  z1
y = 1 −1
z2
 
1 −1
Observability matrix, V =
−1 3
Rank of the matrix V = 2.
Therefore, the system is Observable.

V. Sankaranarayanan Modern Control systems


Observability

Observability

Example-1
Now, let us check the observability of the system by transforming it into
Canonical Form-II using Similarity Transformation.
The transformation matrix P is obtained by Conventional method, explained
in Lecture-5.
 
1 −2
Matrix P =
−1 3
Transformed matrices Az , Bz and Cz are obtained using relations
Az = P −1 AP , Bz = P −1 B and Cz = CP
   
0 −3 2  
Az = Bz = Cz = 0 1
1 −4 1

V. Sankaranarayanan Modern Control systems


Observability

Observability

Example-1
Representing them in matrix form , we get

      
ż1 0 −3 z1 2
= + u
ż2 1 −4 z2 1
 
  z1
y = 0 1
z2
 
0 1
Controllability matrix, V =
1 −4
Rank of the matrix V = 2 .
Therefore, the system is Observable.

V. Sankaranarayanan Modern Control systems


Observability

Observability

Example-2
Consider the state and output equations,

ẋ1 = 2x2
ẋ2 = −4x1 − 6x2 + u
y = x1 + x2

Let us try to check for Observability of this system by transforming these


equations into different forms.
Representing these equations in matrix form , we get

      
ẋ1 0 2 x1 0
= + u
ẋ2 −4 −6 x2 1
 
  x1
y = 1 1
x2

V. Sankaranarayanan Modern Control systems


Observability

Observability

Example-2
We know that the Observability matrix for a system having two state
variables is given by

 
C
V =
CA 2×2

 
1 1
∴V =
−4 −4

Rank of matrix V = 1, less than 2.


Therefore, the system is not Observable.

V. Sankaranarayanan Modern Control systems


Observability

Observability

Example-2
Now, let us check the Observability of the system by transforming it into
diagonal canonical form.
The transformation matrix P is the modal matrix formed by the eigenvectors
corresponding to the eigenvalues of matrix A.
 
1 1
Matrix P =
−1 −2
Transformed matrices Az ,and Cz are obtained using relations
Az = P −1 AP , Cz = CP
 
−2 0  
Az = Cz = 0 −1
0 −4

V. Sankaranarayanan Modern Control systems


Observability

Observability

Example-2
Representing them in matrix form , we get

      
ż1 −2 0 z1 1
= + u
ż2 0 −4 z2 −1
 
  z1
y = 0 −1
z2
 
0 −1
Observability matrix, V =
0 4
Rank of the matrix V = 1.
Therefore, the system is not Observable.

V. Sankaranarayanan Modern Control systems


Observability

Observability

Example-2
Now, let us check the observability of the system by transforming it into
Canonical Form-II using Similarity Transformation.
The transformation matrix P is obtained by Conventional method, explained
in Lecture-5.
 
1 0
Matrix P =
0 −4
Transformed matrices Az , Bz and Cz are obtained using relations
Az = P −1 AP , Bz = P −1 B and Cz = CP
     
0 −8 0 1
Az = Bz = Cz =
1 −6 −1/4 −4

V. Sankaranarayanan Modern Control systems


Observability

Observability

Example-2
Representing them in matrix form , we get

      
ż1 0 −8 z1 0
= + u
ż2 1 −6 z2 −1/4
 
  z1
y = 1 −4
z2
 
1 −4
Controllability matrix, V =
1 −4
Rank of the matrix V = 1 .
Therefore, the system is not Observable.

V. Sankaranarayanan Modern Control systems


Observability

Assignment Questions on Controllability & Observability

1 State space representation of the system is given by

        
x˙1 −1 −2 −2 x1 2   x1
 x˙2  =  0 −1 1  x2 + 0  u; y = 1 1 0  x2  (3)
x˙3 1 0 −1 x3 1 x3

Check whether the system is completely state controllable?


2 Check the controllability of the system (3) by transforming it into diagonal canonical form.
3 Check the controllability of the system (3) by transforming it into Canonical Form-I using
Similarity Transformation.
4 State space representation of the system is given by

        
x˙1 2 0 0 x1 0 1     x1
u1 1 0 0
 x˙2  =  0 2 0  x2 + 1 0  ; y =  x2  (4)
u2 0 1 0
x˙3 0 3 1 x3 0 1 x3

Check whether the system is completely state controllable?


5 Check the controllability of the system (4) by transforming it into diagonal canonical form.
6 Check the controllability of the system (4) by transforming it into Canonical Form-I using
Similarity Transformation.
7 State space representation of the system is given by

        
x˙1 0 1 0 x1 0   x1
 x˙2  =  0 0 1  x2 + 0  u; y = 20 9 1  x2  (5)
x˙3 −6 −11 −6 x3 1 x3

Check whether the system is completely state observable?


8 Check the observability of the system (5) by transforming it into diagonal canonical form.
V. Sankaranarayanan Modern Control systems
Observability

9 Check the observability of the system (5) by transforming it into Canonical Form-II using
Similarity Transformation.
10 State space representation of the system is given by

      
x˙1 1 1 2 x1   x1
 x˙2  =  0 2 0  x2 ; y = 1 0 0  x2  (6)
x˙3 0 0 2 x3 x3

Check whether the system is completely state observable?


11 Check the observability of the system (6) by transforming it into diagonal canonical form.
12 Check the observability of the system (6) by transforming it into Canonical Form-II using
Similarity Transformation.

V. Sankaranarayanan Modern Control systems

You might also like