Unit-54

State Transition Matrix

Prof HarisH Patil@CCOEW 1

 State Transition Matrix
For a linear time-invariant system, the dynamic equations are written
As:
d
State equation: x(t )  Ax(t )  Bu (t )
dt
Output equation: y (t ) Cx (t )  Du (t )

The behavior of x(t) and y(t) :

1. Homogeneous solution of x(t)
2. Non-homogeneous solution of x(t)

Prof HarisH Patil@CCOEW 2

 Homogeneous solution
x (t )  Ax(t ) 1 1
x(t ) L [( sI  A) ]x(0)
sX ( s )  x(0)  AX ( s )
e At x(0)
X ( s ) ( sI  A)  1 x(0)

At 1 1
State transition matrix  (t ) e L [( sI  A) ]

x(t0 ) e At0 x(0)

 At0
x(0) e x(t0 )
At  At0 A( t  t0 )
x(t ) e e x(t0 ) e x(t0 )  (t  t0 ) x(t0 )
Prof HarisH Patil@CCOEW 3
 State transition matrix
Δ
Properties : Φ(t )= e At = L−1 [( sI− A )−1 ]

1.  (0) I
1
2.  (t )  ( t )
3. x(0)  ( t ) x(t )
4.  (t 2  t1 )  (t1  t0 )  (t 2  t0 )
k
5.  (t )  (kt )
Prof HarisH Patil@CCOEW 4
 State transition matrix
Determine the state transition matrix for given A:

Prof HarisH Patil@CCOEW 5

State transition matrix

Prof HarisH Patil@CCOEW 6

 Non-homogeneous Solution
d
x(t )  Ax(t )  Bu (t )
dt
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)
−1 −1 −1 −1
x(t )=L [(sI − A ) ] x(0)+L [( sI− A ) BU ( s)]
t
x(t )=Φ(t ) x(0)+∫0 Φ(t−τ )Bu( τ )dτ Convolution
Prof HarisH Patil@CCOEW 7
 t
x(t )=Φ( t )x(0 )+∫0 Φ(t−τ ) Bu(τ )dτ
t
x(t )=Φ( t−t 0 )x(t 0 )+∫t Φ(t−τ )Bu (τ )dτ
0
t
y(t )=CΦ(t−t 0 ) x(t0 )+∫t CΦ (t−τ )Bu( τ )dτ +Du(t )
0

Zero-input response Zero-state response

Prof HarisH Patil@CCOEW 8

Example 1
[ ][
ẋ 1
=
0 1 x1 0
+ u(t )
ẋ 2 −2 −3 x 2 1 ][ ] [ ]
T
let x (0 )= [ 0 0]

[ ]
−t −2 t −1 −2 t
2 e −e e −e
Φ(t )=L−1 [(sI − A )−1 ]=e At =
−2 e−t +2 e−2t −e −t +2 e−2 t

t
x(t )=Φ( t )x(0 )+∫0 Φ(t−τ ) Bu(τ )dτ

[ ][ ]
1 3
x1 + 2 e−t − e−2 t
Ans: x2
= 2 2 L 1[( sI  A)  1 BU ( s )]
−2 e−t +2 e−2t

Prof HarisH Patil@CCOEW 9

How to find State transition matrix

At 1 1
 (t ) e L [( sI  A) ]

Methode 1:  (t ) L 1[( sI  A)  1 ]

Methode 2:  (t ) e At

Methode 3: Cayley-Hamilton Theorem

Prof HarisH Patil@CCOEW 10

 Controllability and Observability
In Control Problems two basic questions need to be answered
in deciding whether or not a control solution exists....
1. Can we transfer the system from any initial state to any
otherdesired state in finite time by application of
suitable control force?
2. Knowing the output vector for finite length of time, can
we determine the initial state of the system?

The answers to above basic questions were conceptualized by

Kalman into what is known as Controllability and
Observability.
Prof HarisH Patil@CCOEW 11
 Controllability and Observability
Controllability :
As a system is said to be completely state controllable if it is possible to
transfer the system state from any initial state x(t0 ) to any other desired
state x(t ) in specified time by control vector u(t).

Observability:
A system is said to be completely observable, if every state x(t0 ) can
be completely identified by measurement of the output y(t) over finite
time interval.
A system which is not completely observable, implies that some of its state
variables are shielded from observation.

Prof HarisH Patil@CCOEW 12

 Controllability
For a linear time-invariant system, the dynamic equations are
written as: ׁx = Ax + Bu
y = Cx + Du
The mathematical relations given below explain the transfer of a
system from initial state x(0) to final state x(n)
x(1) = Ax(0) + Bu(0)
x(2) = Ax(1) + Bu(1) = A {Ax(0) + Bu(0)} + Bu(1)
x(3) = Ax(2) + Bu(2) = A {AAx(0) + Bu(0) + Bu(1)} + Bu(2)
.
x(n) = A^n x(0) + A^(n−1) Bu(0) + A^(n−2) Bu(1) + · + Bu(n − 1)

Prof HarisH Patil@CCOEW 13

 Controllability
For a system to be completely controllable, it is necessary and
sufficient that the matrix U given below has a rank of n.

U is known as controllability matrix of order n by nm

The controllability condition of a system is governed by the

coefficient matrices A and B.

However, if the controllability test matrix U is n X n i.e. square

matrix, then the condition for state controllability is |U|=0 which
implies that the matrix U is non-singular.
Prof HarisH Patil@CCOEW 14
 Controllability
For a system to be completely controllable, it is necessary and
sufficient that the matrix (n by nm) given below has a rank of n.

U is known as controllability matrix of order n by nm

The controllability condition of a system is governed by the

coefficient matrices A and B.

However, if the controllability test matrix U is n X n i.e. square

matrix, then the condition for state controllability is |U|=0 which
implies that the matrix U is non-singular.
Prof HarisH Patil@CCOEW 15
 Observability
For a linear time-invariant system, the dynamic equations are
written as: ׁx = Ax + Bu
y = Cx + Du
For the determination of observability condition, an unexcited
system is considered, because the control vector is not affecting
the said condition.ׁ
ֹx = Ax
y = Cx
x(1) = Ax(0)
y(0) = Cx(0)
Therefore, y(1) = Cx(1) = CAx(0)
y(2) = CAAx(0)
y(3) = CAAAx(0)...........
Prof HarisH Patil@CCOEW 16
Observability

Prof HarisH Patil@CCOEW 17

 Observability
For unique determination of the state x(0), the matrix



must have inverse and therefore, the matrix is to have rank n.

However if the matrix is nXn i.e. square one then for the
confirmation of observability, the same should be non-singular.

The matrix(38) is known as observability test matrix which can

be written as nXnp row matrix form and denoted by V as given
below. Prof HarisH Patil@CCOEW 18
 Observability



The above matrix is known as observability test matrix which

can be written as nXnp row matrix form and denoted by V as
given below.

If the observability condition is to be satisfied the matrix V

should have a rank of n.
The matrix pair (AC) is termed as observable pair.
Prof HarisH Patil@CCOEW 19
 Observability



The above matrix is known as observability test matrix which

can be written as nXnp row matrix form and denoted by V as
given below.

If the observability condition is to be satisfied the matrix V

should have a rank of n.
The matrix pair (AC) is termed as observable pair.
Prof HarisH Patil@CCOEW 20
 Observability



The above matrix is known as observability test matrix which

can be written as nXnp row matrix form and denoted by V as
given below.

If the observability condition is to be satisfied the matrix V

should have a rank of n.
The matrix pair (AC) is termed as observable pair.
Prof HarisH Patil@CCOEW 21
 Controllability and Observability
x (t )  Ax(t )  Bu (t )
y (t ) Cx(t )  Du (t )

Controllability matrix U =[ B AB A2 B⋯ An−1 B ]

[]
C
CA
Observability matrix V = CA2
⋮
CA n−1

Then: (1) controllable rank (U ) n

(2) observable rank (V ) n
Prof HarisH Patil@CCOEW 22
Example 1
 x 1    2 1   x1  1
 x   0  1  x   0 u (t )
 2   2   
 x1 
y 1 0  
 x2 

 1  2
U [ B AB ]   rank[U ] 1 uncontrollable
0 0 
 C   1 0
V      rank[V ] 2 observable
CA   2 1
The rank of a matrix is defined by the number of
linearly independent rows and/or the number of
linearly independent columns
Prof HarisH Patil@CCOEW 23
Example 2
 x 1    2 0   x1  3
 x   0  1  x   1 u (t )
 2   2   
 x1 
y 1 0  
 x2 

3  6 
U [ B AB ]   rank[U ] 2 controllable
1  1
 C   1 0
V     rank[V ] 1 unobservable
CA
    2 0 

Prof HarisH Patil@CCOEW 24

Example 2
 x 1    2 0   x1  3
 x   0  1  x   1 u (t )
 2   2   
 x1 
y 1 0  
 x2 

3  6 
U [ B AB ]   rank[U ] 2 controllable
1  1
 C   1 0
V     rank[V ] 1 unobservable
CA
    2 0 

Prof HarisH Patil@CCOEW 25

Example 3

Check the Controllability and Observability of the system with state

space model:

Prof HarisH Patil@CCOEW 26

Example 3

Check the Controllability and Observability of the system with state

space model:

Prof HarisH Patil@CCOEW 27

Example 3

Check the Controllability and Observability of the system with state

space model:

Prof HarisH Patil@CCOEW 28

Example 3

Check the Controllability and Observability of the system with state

space model:

Since the first three columns are linearly independent we can conclude
that rank is 3, hence the system is controllable.
Prof HarisH Patil@CCOEW 29
Example 4

Check the Controllability and Observability of the system with state

space model:

Prof HarisH Patil@CCOEW 30

Example 4

Check the Controllability and Observability of the system with state

space model:

Prof HarisH Patil@CCOEW 31

Example 4

Check the Controllability and Observability of the system with state

space model:

Prof HarisH Patil@CCOEW 32

Example 4

Check the Controllability and Observability of the system with state

space model:

Since U and V are square matrix, calculate the | U | and | V |.

| C | = 1 = not 0 = rank of U= 3 =n .......Controllable system.

| O | = 0 = rank of V < 3=n .......Unobservable system.

Prof HarisH Patil@CCOEW 33