Solving the time-invariant state

equation
• Solution of Homogeneous State Equations
• Let us consider a scalar differential equation
x  ax
• We may assume a solution of the form
x(t )  b0  b1t  b2t 2   (a )
• By substituting this assumed solution
b1  2b2t  3b3t 2    a (b0  b1t  b2t 2  )
• If the assumed solution is the true solution, equating
the coefficients of the equal powers of t
1 1
b1  ab0 ; b2  ab1  a 2b0 1
2 2
 Solving the time-invariant state
equation (contd…)
1 1 3
b3  ab2  a b0
3 3 2
1 1
b4  ab3  a 4b0
4 4  3 2

1 k
bk  a b0
k!
The value of b0 is determined by substituting t=0 in eqn (a)
x(0)  b0

2
 Solving the time-invariant state
equation (contd…)
• Hence, the solution x(t) can be written as
 1 1 
x(t )  1  at  a 2t 2  a 3t 3   x(0)
 2! 3! 
 e at x(0)

• Now we shall solve matrix differential equation

X  AX
where X is a vector
A  n  n constant matrix

3
 Solving the time-invariant state
equation (contd…)
• By analogy to the scalar case, we assume that the
solution is of the form of a vector power series in t

X (t )  B0  B1t  B2t 2  

• By substituting this assumed solution

B1  2 B2t  3B3t 2    A( B0  B1t  B2t 2  )
• By equating the coefficients of like powers of t on
both sides B1  AB0 ;
1 1
B2  AB1  A2 B0
2 2
1 1
B3  A3 B0  A3 B0 4
3 2 3!
 Solving the time-invariant state
equation (contd…)
• By substituting t=0, X(0)=B(0)
• The solution X(t) can be written as
 1 1 
X (t )   I  At  A2t 2  A3t 3   X (0)
 2! 3! 
X (t )  e At X (0)

Solution in terms of matrix exponential

5
Matrix Exponential: Properties
At 1 2 2 1 33
e  I  At  A t  A t  ...
2! 3!
d At 2 A3t 2
e  A A t  
dt 2!
A2t 2
 A( I  At   )
2!
 Ae At
A2t 2
 ( I  At   ) A
2!
 e At A

6
Matrix Exponential: Properties (contd…)
A(t  s ) At As
e e e
e At e  At  e A(t t )  I

The inverse of eAt is e-At

e ( A B )t  e At e Bt if AB  BA
e ( A B )t  e At e Bt if AB  BA

For Proof refer Modern Control Engg. by Ogata

7
 Laplace Transform Approach to the
solution of Homogeneous state equations
• Scalar Case x  ax
sX ( s )  x(0)  aX ( s )
x ( 0)
X (s)   ( s  a ) 1 x(0)
sa
x(t )  e at x(0)
• Extending this to homogeneous state equation
X  AX
sX ( s )  X (0)  AX ( s )
X ( s)  ( SI  A) 1 X (0)
 
X (t )  L1 ( SI  A) 1 X (0) 8
 Laplace Transform Approach to the
solution of Homogeneous state equations
(contd…) (1 − ) = 1 + + + … .
1 I A A2
sI  A   2  3  
s s s
A2t 2 A3t 3

1
 1
L sI  A  I  At 
2!

3!
   e At

X (t )  e At X (0) (b)

f(t) F(s)
t 1/s2
tn-1/(n-1)! 1/sn

9
 State Transition Matrix
X  AX (1)
X (t )   (t ) X (0) (2)
where  (t ) is an n  n matrix and unique solution of
 (t )  A (t ); (0)  I ;

To verify
X (0)   (0) X (0)  X (0)
X (t )  
 (t ) X (0)  A (t ) X (0)  AX (t )
Thus we can say eqn (2) is a solution of eqn (1)
From eqn (2) we can also state that, solution of (1) is simply transformation of
the initial condition.
The unique matrix φ(t) is called the State Transition Matrix
10
 State Transition Matrix (contd…)
At
 (t )  e  L 1
sI  A 
1

 1 (t )  e  At
  (t )

If the eigen values λ1, λ2,… λn are distinct, then φ(t) will
contain n exponentials
e λ1t, e λ2t,… e λnt t
e 1 0 
 
If matrix A is diagonal  e 2 t 
 (t )  e At   . 
 
 . 
0 e nt 
 11
 Properties of State Transition Matrix
X  AX
• For a time invariant system
for which  (t )  e At
A0
1.  (0)  e I
At
2.  (t )  e  e    (t )
 At 1 1

 1 (t )   (t )
3.  (t1  t 2 )  e A(t1 t2 )  e At1 e At2
  (t1 ) (t 2 )

12
Properties of State Transition
Matrix(contd…)

4.  (t )   (nt )
n

5.  (t 2  t1 ) (t1  t0 )   (t 2  t0 )
  (t1  t0 ) (t 2  t1 )

13
 Problem
• Obtain the STM φ(t) of the following system. Also
obtain φ-1(t)
 x1   0 1   x1 
 x    2  3  x 
 2   2 
0 1
A 
  2  3
1
 (t )  e At  L1[sI  A ]
 s 0  0 1  s 1 
sI  A       
 0 s    2  3  2 s  3
14
 Problem (contd…)
1 1  s  3 1
sI  A   
( s  1)( s  2)   2 s 
 s3 1 
 ( s  1)( s  2) ( s  1)( s  2) 
 
 2 s 
 ( s  1)( s  2) ( s  1)( s  2) 

1
 (t )  e At  L1[sI  A ]
 2e t  e  2t e t  e  2 t 
 t  2t t  2t 
  2 e  2 e  e  2e 
15
 Problem (contd…)

 1 (t )   (t )
1  At
 (t )  e
 2e t  e 2t e t  e 2t 
 t 2t t 2t 
  2e  2e  e  2e 

16
 Example
RLC Series Circuit
Ei=10V;
Initial current=1mA;
Initial voltage across capacitor=2V;
R=4.1kΩ; L=1.7H;C=0.1

17
Example

18
 Solution of Non-homogeneous State
Equations
• Scalar case
x  ax  bu
x  ax  bu
d  at
 at
e [ x  ax] 
dt

e x(t )  e  at bu 
• Integrating this equation between 0 and t gives
t
e  at x(t )  x(0)   e  a bu ( )d
0
t Response due to initial condition +
at at  a
x(t )  e x(0)  e e bu ( )d Response due to input
0 19
 Solution of Non-homogeneous State
Equations (contd…)
• Vector Case
X  AX  BU
X  AX  BU
Pr emultipying by e  At
d  At
e  At


X  AX  
dt
 
e X (t )  e  At BU
t
e  At
X (t )  X (0)   e  A BU ( ) d
0
t t
X (t )  e At X (0)  e At  e  A BU ( )d  e At X (0)   e A( t  ) BU ( )d
0 0
t
X (t )   (t ) X (0)    (t   ) BU ( )d
0

where  (t )  e At 20
 Laplace Transform Approach to the Solution of
Non-homogeneous State Equations
x (t )  Ax(t )  Bu (t )
y (t )  Cx(t )  Du (t )

sX ( s)  x(0)  AX ( s)  BU ( s)
( sI  A) X ( s )  x(0)  BU ( s)
1 1
X ( s)  ( sI  A) x(0)  ( sI  A) BU ( s )
x(t )  L1[( sI  A) 1 ]x(0)  L1[( sI  A) 1 BU ( s )]
t
x(t )  (t ) x(0)    (t   ) Bu ( )d Convolution
0
Homogeneous t 
L  f1 (t   ) f 2 ( )d   F1 ( s ) F2 ( s )
0  21
 Laplace Transform Approach to the Solution of
Non-homogeneous State Equations

t
x(t )  (t ) x(0)    (t   ) Bu ( )d
0
t
x(t )  (t  t0 ) x(t0 )   (t   ) Bu ( )d
t0
t
y (t )  C(t  t0 ) x(t0 )   C(t   ) Bu ( )d  Du (t )
t0

Zero-input response Zero-state response

22
Example 1  x1   0 1   x1  0
 x    2  3  x   1u (t )
 2   2   
x(0)  0 0
T
let
t 2 t
 2 e  e e t  e 2t 
 (t )  L1[( sI  A) 1 ]  e At   t  2t 
  2 e  2 e  e  t  2e  2 t 

t
x(t )  (t ) x(0)    (t   ) Bu ( )d
0

1 1
 x1    e t  e  2t 
Ans:
x    2 2 L1[( sI  A) 1 BU ( s )]
 2   e t  e  2t 

23
How to find State transition matrix


(t )  e At  L1[( sI  A) 1 ]

Methode 1:
(t )  L1[( sI  A) 1 ]
Methode 2:
(t )  e At
Methode 3: Cayley-Hamilton Theorem

24
Methode 2:
(t )  e At
 x1   1 0 0   x1  1 diagonal matrix
 x    0  2 0   x   1  u1 
 2    2    u 
 x3   0 0  3  x3  1  2 
 x1 
 y1 (t )  x 
 y (t )  6  6 1 2
 2   x3 

e  t 0 0 
At   2t 
(t )  e   0 e 0 
0 0 e 3t 


25
 Problem
• For a system represented by state equation
X  AX ;
• The response is

 e 2 t  1
X (t )   2t 
when X (0)   
 2e   2
 e t  1
X (t )   t  when X (0)   
 e   1
• Determine the system matrix A and the state transition
matrix

26
 Problem (contd…)
• The solution of the state equation is
X (t )  e At X (0) (a)
Pr emultiplying by e  At
 At  At At
e X (t )  e e X (0)
 X (0) (b)
• One of the response is
 e 2t  1
X (t )    2t 
; X ( 0)   
  2e   2
27
 Problem (contd…)
Substituting in (b)
 2t
 At  e  1
e   2t 
 
 2e   2
 At  e11 e12 
Let e   
e21 e22 
 2t
e e
 11 12  e   1
e    2t 
 
 21 e22   2e   2
e11e  2t  2e12e  2t  1; e21e  2t  2e22 e  2t  2 eqns (c) & (d )
28
 Problem (contd…)
• Corresponding to 2nd solution
 e t  1
X (t )   t  and X (0)   
 e   1
t
 At  e  1
e  t    
 e   1
 e11 e12   e t   1 
e   t    
 21 e22   e   1
e11e t  e12 e t  1; e21e t  e22 e t  1 eqns (e) & ( f )
29
 Problem (contd…)
• Solving c, d, e and f
• Eqn (e)×e-t - eqn©

e12  e t  e 2t
e11  2e t  e 2t
t 2t
e22  e  2e
t 2t
e21  2e  2e

30
 Controllability
• A system is said to be completely controllable at
time t0 if it is possible by means of an unconstrained
control vector to transfer the system from the initial
state x(t0) to any other state in a finite interval of
time.
• Can we design a controller?

31
 Controllability (contd…)
• Consider a continuous time system
X  AX  Bu
X  State Vector
u  Control signal
A  n  n matrix
B  n  1 matrix

• The system is completely state controllable if and only

if the n×n matrix [ B AB A2 B ... An1 B]
is of rank n
• Controllability matrix
• Output controllability CB CAB CA 2
B ... CA n 1
B 32D 
 Controllability (contd…)
• Consider the system
 x1  1 1   x1  1
 x   0  1  x   0u
 2   2   
1 1 
B AB   
0 0 

• Not state controllable

33
 Controllability (contd…)

 x1  1 1   x1  0
 x   2  1  x   1u
 2   2   
0 1 
B AB   
1  1

• Completely state controllable

• There should not be any pole zero cancellation

34
Example

35
 Observability
• A system is said to be observable at time t0 if, with
the system in state x(t0), it is possible to determine
this state from the observation of the output over
finite time interval.
• Rank of  C  must be n
 CA 
 CA 2 
 
  
CA 
n 1

C *
A*C * ( A* ) 2 C * ... ( A* ) n1 C * 
36
Observability(contd…)
 x1   0 1 0   x1  0
 x    0 0 1   x   0  u
 2   2   
 x3   6  11  6  x3  1
 x1 
y  4 5 1 x2 
 x3 

C *
A*C * ( A* ) 2 C * ... ( A* ) n 1 C * 
4  6 6 
 5  7 5 
1  1  1

37
Observability(contd…)

4 6 6
5 7 5 0
1 1 1

38
39
Controllability (contd…)

40