You are on page 1of 31

Lecture 20

Controllability and Observability of Linear

Time Invariant Systems
Asst. Professor
Dept. of Aerospace Engineering
Indian Institute of Science - Bangalore

Evaluation of Matrix Exponential eAt

Asst. Professor
Dept. of Aerospace Engineering
Indian Institute of Science - Bangalore

Method 1: Power-series
2 2
3 3
A
t
A
t
At
+
+"
e = I + At +
2!
3!

This method is useful and accurate only if the

series truncates naturally. Otherwise, series
truncation introduces approximation error.

Direct computation of eAt as power series is

computationally inefficient as well.

Method 2:
Using Laplace Transform
1

e = L ( sI A )

At

This method results in closed form expressions for

eAt, can be quite useful for small matrices.

1

Can be quite cumbersome for large matrices.

Method 3:
Using Similarity Transform

2 2
3 3
A
t
A
t
At
+
+"
e = I + At +
Similarity
2!
3!
Transformation:
1
1
2
PDP
PDP
t
(
)(
)
1
1
1
= PP + PDP t +
+ " A = PDP
2!

1
D 2 t 2 D 3t 3
= P I + Dt +
+
+ " P
2!
3!

e1t 0
0

1
= P 0 % 0 P
0 0 ent

Method 4: Sylvesters Formula

Case 1: Distinct Eigenvalues

eAt satisfies the following determinant equation:

1 1
1 2
#

1 n
I A
i.e.

12 " 1n 1
22 " 2n 1
#

n2 " nn 1
A2 " An 1

e1t
e2t
#
ent
At
eN

=0

Ultimate aim

e At = 0 ( t ) I + 1 ( t ) A + 2 ( t ) A2 + " + n 1 ( t ) An 1

Method 4: Sylvesters Formula

Case 1: Distinct Eigenvalues

The coefficients 0 ( t ) , 1 ( t ) ," , n 1 ( t )

can be determined from the following set
of equations:

0 ( t ) + 1 ( t ) 1 + 2 ( t ) 12 + " + n 1 ( t ) 1n 1 = e t
0 ( t ) + 1 ( t ) 2 + 2 ( t ) 22 + " + n 1 ( t ) 2n 1 = e t
1

0 ( t ) + 1 ( t ) n + 2 ( t ) + " + n 1 ( t )
2
n

n 1
n

=e

n t

Method 4: Sylvesters Formula

Case 2: Repeated Eigenvalues

eAt satisfies the following determinant equation:

0

1 1

31

"

21 312 "

2
( n 1) 1n2

1t

"

"

1 n

n2

n3

"

nn 1

ent

A2

A3

"

An 1

e At

2
1
2
4

"

t 2 1t Eigenvalues:
e
2
1 , 1 , 1 , 4 ," , n
1t


te

3
1
3
4

1 4

i.e.

( n 1)( n 2 ) n3
n 1
1
n 1
4

e
e

4t

3 times

=0

e At = 0 ( t ) I + 1 ( t ) A + 2 ( t ) A2 + " + n 1 ( t ) An 1

Method 4: Sylvesters Formula

Case 2: Repeated Eigenvalues

The coefficients 0 ( t ) , 1 ( t ) ," , n 1 ( t )

can be determined from:
2
n 1)( n 2 )
(
t
2 ( t ) + 3 3 ( t ) 1 + " +
n 1 ( t ) 1n 3 = e t
1

2
2
1 ( t ) + 2 2 ( t ) 1 + 3 3 ( t ) 12 + " + ( n 1) n 1 ( t ) 1n 2 = te1t

0 ( t ) + 1 ( t ) 1 + 2 ( t ) 12 + " + n 1 ( t ) 1n 1 = e t
1

0 ( t ) + 1 ( t ) 4 + 2 ( t ) 42 + " + n 1 ( t ) 4n 1 = e t
4

0 ( t ) + 1 ( t ) n + 2 ( t ) n2 + " + n 1 ( t ) nn 1 = e t
1

Method 4: Sylvesters Formula

Example
0 1
A=
1,2 = 0, 2
,

0 2
To compute e At using Sylvester's formula, we have
1 1

e1t

1 2
I A

e2t = 1 2 e 2t = 0
e At
I A e At

Expanding the determinant

2e At + A + 2 I Ae 2t = 0

1
1
At
2 t

e = ( A + 2 I Ae ) =

2
0

1
2 t

e
1
(
)
2

2 t
e

10

Controllability of Linear Time Invariant Systems

Asst. Professor
Dept. of Aerospace Engineering
Indian Institute of Science - Bangalore

Controllability

if

it is possible by means of an unconstrained

control vector to transfer the system from any
initial state X 0 to any other state in a finite
interval of time

Controllability depends upon the system matrix

A and the control influence matrix B

12

Graphical Meaning

Xf

X0

13

Condition for Controllability:

(single input case)
System:

X = AX + Bu
t

Solution:

X (t ) = e At X (0) + e A( t ) Bu ( ) d
0

Assuming X (t1 ) = 0,
t1
0 = eAt1 X (0) + eA(t1 ) Bu( ) d
0

t1

X (0) = e A Bu ( ) d
0

14

Condition for Controllability:

(single input case)
n 1

e A = k ( ) Ak

(Sylvester's formula )

k =0
t1

n 1

t1

k =0

X (0) = e A Bu ( ) d = Ak B k ( ) u ( ) d
n 1

= Ak B k

t1

where

k  k ( ) u ( ) d
0

k =0

= B

AB " A B [ 0
n 1

This system should have a non-trivial solution for

1 " n 1 ]

[ 0

1 " n 1 ]

15

Controllability
If the rank of CB  B AB " An 1 B is n,
then the system is controllable.

Result:

Example:

x1 1 0 x1 2
x = 0 2 x + 1 u
2
2
2 1 0 2 2 2
CB =
=

1
0
2
1
1
2

rank ( CB ) = 2 The system is controllable.

16

Output Controllability
Result:

X = AX + BU
Y = CX + DU
X Rn , U Rm , Y R p

If the rank of CB  CB CAB " CAn 1 B D is p,

then the system is output controllable.
Note: The presence of DU term in the output equation
always helps to establish output controllability.

17

Observability of Linear Time Invariant Systems

Asst. Professor
Dept. of Aerospace Engineering
Indian Institute of Science - Bangalore

Observability

time

t0 if, with the system in state X(t0) ,it is

possible to determine this state from the
observation of the output over a finite
interval of time

Observability

matrix A and the output matrix C

19

Observability
If the rank of OB  C T AT C T

then the system is observable.

Result:

"

(A )
T

n 1

C T is n,

Example:

x 1 1 0 x1 2
x = 0 2 x + 1 u
2
2

x1
y = [1 0]
x2

1 1 0 1 1 1
OB =
=

0
0
2
0
0
0

rank ( OB ) = 1 2 The system is NOT observable.

20

Controllability and Observability

in Transfer Function Domain
z

The system is both controllable and

observable if there is no Pole-Zero
cancellation.

Note: The cancelled pole-zero pair

suppresses part of the information about
the system

21

Principle of Duality
System S1: X = AX + BU
Y1 = CX
System S2: Z = AT Z + C T V

Y2 = BT Z

CB = B
OB = C T

AB
T

AC

A2 B " An 1 B
T

T2

A C

T n1

" A

CT

CB = C T AT C T AT C T " AT C T

OB = B AB A2 B " An 1 B
2

n1

The principle of duality states that the system S1 is controllable if

and only if system S2 is observable; and vice-versa!
Hence, the problem of observer design for a system is actually a
problem of control design for its dual system.

22

z

Stabilizable system: Uncontrollable

system in which uncontrollable part is
stable

Detectable system: Unobservable

system in which the unobservable
subsystem is stable

23

Example

System Dynamics
x1 2 3 2 1 x1 1
x 2 3 0 0 x 2
2 =
2 + u
x3 2 2 4 0 x3 2

x4 2 2 2 5 x4 1
N


N N
X

Output Equation
y = [ 7 6 4 2] X


C

24

Example

Ref: B. Friedland, Control System Design, McGraw Hill, 1986

Transfer Function:

y (s)
s + 2 )( s + 3)( s + 4 )
(
1
1
= C ( sI A ) B =
=
u (s)
s + 1)( s + 2 )( s + 3)( s + 4 ) ( s + 1)
(

pole-zero cancellation

Implication: What appears to be a fourth-order system, is

actually a first-order system! Hence, there is
either loss of controllability or observability
(or both).
Question: Is this system stabilizable?

25

Example

Ref: B. Friedland, Control System Design, McGraw Hill, 1986

Define X = TX . Then
X = TX = T ( AX + Bu )
X = TAT 1 X + TB u

( )

Let
4
3
T =
2

3 2 1
1 0 0 0
1
0 2 0 0
0
3 2 1
, TB =
TAT 1 =
0 0 3 0
1
2 2 1

1 1 1
0 0 0 4
0

26

Example

Ref: B. Friedland, Control System Design, McGraw Hill, 1986

x1 x1 + u


x
x
2
2 =
,
x3 3 x3 + u


x4 4 x4

y = CX = CT 1 X = x1 + x2

Implications:
x1 : Affected by the input; visible in the output
x2 : Unaffected by the input; visible in the output
x3 : Affected by the input; Invisible in the output
x4 : Unaffected by the input; Invisible in the output

27

Block Diagram:

28

Where do uncontrollable or
unobservable systems arise?
z

29

References
z

K. Ogata: Modern Control Engineering,

3rd Ed., Prentice Hall, 1999.

B. Friedland: Control System Design,

McGraw Hill, 1986.