.

− 5 − 1  2
X (t ) =   X (t ) +  U (t )
 3 − 1 5
Y (t ) = 1 2X (t )
 UNIT - V
There are two types of system representation in control systems

1. Transfer function – conventional method

2. State space – modern method

State Variables: The variables that determine the future state

(behavior) and output of the system when the present state and and
input signals are known.

Transfer function method State variable method

For LTIV systems only For nonlinear, time-varying
and Multivariable systems

Initial value is zero Initial values are considered.

Consider a linear time invariant system
U1(t) y1(t)

U2(t) y2(t)
MIMO system
Inputs (x1(t), x2(t)….xn(t)) Outputs
State Variables
Um(t) yp(t)
The state model of the system is

State equation
Output equation
State Variables can be

1. Physical variables – system variables

2. Phase variables – system output and its derivatives

3. Canonical variables – variables associated with poles of

the system
State Space Representation using Physical Variables
1. Obtain the state model using physical variables.
In mechanical
systems

Displacement
(spring) and
velocity (mass)
fv are state
variables

No. of state variables = No. of storage elements = 4

• •
No. of mass elements = 2 ∴ Two velocities y1 ( t ) are y 2 ( t )
chosen as state variables

No. of spring elements = 2 ∴ Two displacements y1(t) and y2(t) are

chosen as state variables
The four state variables are

The two differential equations of the system are

d 2 y1
m1 2
+ k1 y1 − k 2 [ y2 − y1 ] = f ( t )
dt

d 2 y2 dy2
m2 2
+ fv + k 2 [ y2 − y1 ] = 0
dt dt
State equation is

fv
KVL equations of the circuit are
State Space Representation using Phase Variables
 Method 1: the differential equation relating output and input of a system

n• ( n −1)• ( n − 2 )• •• •
y + a1 y + a2 y + ............. + a n− 2 y + a n−1 y + a n y = bu

Order of the equation = n

∴Choosing n state variables

x1 = y
•
x2 = y
••
x3 = y
.
.
( n −1)•
xn = y
•
x n + a1 x n + a 2 x n−1 + ............. + a n− 2 x 3 + a n−1 x 2 + a n x1 = bu

•
x1 = x 2
•
x2 = x3
.
.
•
x n −1 = x n
State equation

⎡ • ⎤
⎢ x1 ⎥ ⎡ 0 1 0 0 0 ⎤ ⎡ x1 ⎤ ⎡0⎤
⎢ • ⎥ ⎢ 0 0 1 0 0 ⎥⎥ ⎢⎢ x 2 ⎥⎥ ⎢⎢0⎥⎥
⎢ x2 ⎥ ⎢
⎢ . ⎥=⎢ . . . . . ⎥ ⎢ . ⎥ + ⎢ . ⎥ [u]
⎢ ⎥ ⎢ . . . .
⎥⎢ ⎥ ⎢ ⎥
. ⎥⎢ . ⎥ ⎢ . ⎥
⎢ •. ⎥ ⎢
⎢ x ⎥ ⎢⎣ − a n − a n −1 . . − a1 ⎥⎦ ⎢⎣ x n ⎥⎦ ⎢⎣1⎥⎦
⎣ n⎦
Output equation
⎡ x1 ⎤
⎢x ⎥
⎢ 2⎥
y = [1 0 . . 0]⎢ . ⎥
⎢ ⎥
⎢ . ⎥
⎢⎣ x n ⎥⎦
 Method 2 and 3

n• ( n −1)• ( n − 2 )• •• •
y + a1 y + a 2 y + ............. + a n− 2 y + a n−1 y + a n y =
m• ( m −1)• ( m − 2 )• •• •
b0 u + b1 u + b2 u + ............. + a m − 2 u + a m −1 u+ bm u

Consider n=m
 Method 2
State equation

⎡ • ⎤
⎢ x1 ⎥ ⎡ − a1 1 0 0 0⎤ ⎡ x1 ⎤ ⎡ b1 − a1b0 ⎤
⎢ • ⎥ ⎢ − a2 0 1 0 0⎥⎥ ⎢⎢ x 2 ⎥⎥ ⎢⎢ b2 − a 2 b0 ⎥⎥
⎢ x2 ⎥ ⎢
⎢ . ⎥=⎢ . . . . . ⎥⎢ . ⎥ + ⎢ . ⎥ [u]
⎢ ⎥ ⎢− a ⎥⎢ ⎥ ⎢
. . . 1⎥ ⎢ . ⎥ ⎢bn−1 − a n−1b0 ⎥
⎥
⎢ •. ⎥ ⎢ n−1
⎢ x ⎥ ⎢⎣ − a n . . 0⎥⎦ ⎢⎣ x n ⎥⎦ ⎢⎣ bn − a n b0 ⎥⎦
⎣ n⎦

Output equation
⎡ x1 ⎤
⎢x ⎥
⎢ 2⎥
y = [1 0 . . 0]⎢ . ⎥ + b0 u
⎢ ⎥
⎢ . ⎥
⎢⎣ x n ⎥⎦
 Method 3
State equation

⎡ • ⎤
⎢ x1 ⎥ ⎡ 0 1 0 0 0 ⎤ ⎡ x1 ⎤ ⎡0⎤
⎢ • ⎥ ⎢ 0 0 1 0 0 ⎥⎥ ⎢⎢ x 2 ⎥⎥ ⎢⎢0⎥⎥
⎢ x2 ⎥ ⎢
⎢ . ⎥=⎢ . . . . . ⎥ ⎢ . ⎥ + ⎢ . ⎥[u]
⎢ ⎥ ⎢ 0 0 . .
⎥⎢ ⎥ ⎢ ⎥
1 ⎥ ⎢ . ⎥ ⎢ 0⎥
⎢ •. ⎥ ⎢
⎢ x ⎥ ⎢⎣ − a n − a n −1 . . − a1 ⎥⎦ ⎢⎣ x n ⎥⎦ ⎢⎣1⎥⎦
⎣ n⎦
Output equation

⎡ x1 ⎤
⎢x ⎥
⎢ 2⎥
y = [b1 − a1b0 b2 − a 2 b0 . . bn−1 − a n−1b0 bn − a n b0 ]⎢ . ⎥ + b0 u
⎢ ⎥
⎢ . ⎥
⎢⎣ x n ⎥⎦
 Method 2

n• ( n −1)• ( n − 2 )• •• •
y + a1 y + a 2 y + ............. + a n− 2 y + a n−1 y + a n y =
m• ( m −1)• ( m − 2 )• •• •
b0 u + b1 u + b2 u + ............. + a m − 2 u + a m −1 u+ bm u

Consider n=m
b0 s 3 + b1 s 2 + b2 s + b3
T ( s) = 3
s + a1 s + a 2 s + a 3
2

b1 b2 b3
b0 + + 2 + 3
T ( s) = s s s
⎛ a1 a 2 a 3 ⎞
1− ⎜− − 2 − 3 ⎟
⎝ s s s ⎠
Method 3

Y ( s ) b0 s 3 + b1 s 2 + b2 s + b3
T ( s) = = 3
U ( s) s + a1 s 2 + a 2 s + a 3

Y ( s) X1 ( s) Y ( s)
= •
U ( s) U ( s) X1 ( s)
State Space Representation using Canonical Variables
 b0 s 3 + b1 s 2 + b2 s + b3
T ( s) = 3
s + a1 s + a 2 s + a 3
2

If n=m,

c1 c2 cn
T ( s ) = b0 + + + .... +
s + λ1 s + λ2 s + λn

Where c1, c2, … Cn are residues and λ1, λ2 … λn are roots of

denominator polynomial or poles of the system
State equation

⎡ • ⎤
⎢ x1 ⎥ ⎡ − λ1 0 0 0 0 ⎤ ⎡ x1 ⎤ ⎡1⎤
⎢ • ⎥ ⎢ 0 − λ2 0 0 0 ⎥⎥ ⎢⎢ x2 ⎥⎥ ⎢⎢1⎥⎥
⎢ x2 ⎥ ⎢
⎢ . ⎥=⎢ . . . . . ⎥ ⎢ . ⎥ + ⎢ . ⎥ [u]
⎢ ⎥ ⎢ 0 0 . − λ n −1
⎥⎢ ⎥ ⎢ ⎥
0 ⎥ ⎢ . ⎥ ⎢1⎥
⎢ •. ⎥ ⎢
⎢ x ⎥ ⎢⎣ 0 0 . . − λn ⎥⎦ ⎢⎣ x n ⎥⎦ ⎢⎣1⎥⎦
⎣ n⎦

Output equation

⎡ x1 ⎤
⎢x ⎥
⎢ 2⎥
y = [c1 c2 . . c n −1 cn ]⎢ . ⎥ + b0 u
⎢ ⎥
⎢ . ⎥
⎢⎣ x n ⎥⎦
1. A feedback system has a closed-loop transfer function

2s + 6s + 5
2
T ( s) = 3
s + 4s 2 + 5s + 2
Write three different state space representation.
Method 2: using phase variables

2s + 6s + 5
2
T ( s) = 3
s + 4s 2 + 5s + 2

2 6 5
+ 2+ 3
T ( s) = s s s
⎛ 4 5 2⎞
1− ⎜− − 2 − 3 ⎟
⎝ s s s ⎠
 Method 2
State equation

⎡• ⎤
⎢ x1 ⎥ ⎡ − 4 1 0⎤ ⎡ x1 ⎤ ⎡ 2⎤
⎢• ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎢ x 2 ⎥ = ⎢ − 5 0 0⎥ ⎢ x 2 ⎥ + ⎢ 6⎥ [u]
⎢ • ⎥ ⎢⎣ − 2 0 0⎥⎦ ⎢⎣ x 3 ⎥⎦ ⎢⎣ 5⎥⎦
x
⎢ 3⎥
⎣ ⎦
Output equation

⎡ x1 ⎤
⎢
y = [1 0 0]⎢ x 2 ⎥ ⎥
⎢⎣ x 3 ⎥⎦
Method 3: using phase variables

Y ( s) 2s + 6s + 5
2
= 3
U ( s) s + 4s 2 + 5s + 2

Y ( s) X1 ( s) Y ( s)
= •
U ( s) U ( s) X1 ( s)
 X 1 ( s) 1
= 3
U ( s) s + 4s 2 + 5s + 2
Taking inverse laplace transform

... .. .
x 1 + 4 x 1 + 5 x 1 + 2 x1 = u
Order = 3, ∴ taking three state variables x1, x2 and x3

.
x 1 = x2
.
x 2 = x3
•
x 3 + 4 x 3 + 5 x 2 + 2 x1 = u
 Y ( s)
= 2s 2 + 6s + 5
X1 ( s)
Taking inverse laplace transform

.. .
y = 2 x1 + 6 x 1 + 5 x1

y = 2 x 3 + 6 x 2 + 5 x1
State equation

⎡• ⎤
⎢ x1 ⎥ ⎡ 0 1 0 ⎤ ⎡ x1 ⎤ ⎡0⎤
⎢• ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
x
⎢ ⎥ ⎢
2 = 0 0 1 ⎥ ⎢ x 2 ⎥ + ⎢ 0 ⎥ [u ]
⎢ • ⎥ ⎢⎣ − 2 − 5 − 4⎥⎦ ⎢⎣ x 3 ⎥⎦ ⎢⎣1⎥⎦
x
⎢ 3⎥
⎣ ⎦
Output equation

⎡ x1 ⎤
y = [5 6 2]⎢⎢ x 2 ⎥⎥
⎢⎣ x 3 ⎥⎦
Using canonical variables

2s 2 + 6s + 5
T ( s) =
( s + 1) ( s + 2)
2

c1 c2 c3
T ( s) = + +
( s + 1) 2
s+1 s+ 2

c1 = 1, c2 = 1, c3 = 1
State equation

⎡• ⎤
⎢ x1 ⎥ ⎡ − 1 0 0 ⎤ ⎡ x1 ⎤ ⎡1⎤
⎢• ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
x
⎢ ⎥ ⎢
2 = 0 − 1 0 ⎥ ⎢ x 2 ⎥ + ⎢1⎥ [u ]
⎢ • ⎥ ⎢⎣ 0 0 − 2 ⎥
⎦ ⎢
⎣ x ⎥
⎦ ⎢
⎣1⎥
⎦
x
⎢ ⎥ 3
3
⎣ ⎦
Output equation

⎡ x1 ⎤
y = [1 1 1]⎢⎢ x 2 ⎥⎥
⎢⎣ x 3 ⎥⎦
1. A feedback system has a closed-loop transfer function

Y ( s) 5
T ( s) = = 3
X ( s) s + 4s 2 + 7 s + 2
Write three different state space representation.

Method 1: Applying inverse laplace transformation

3 2
d y d y dy
3
+ 4 2 + 7 + 2 y − 5u = 0
dt dt dt
Order of the equation = 3, therefore Choosing 3 state variables

x1 = y
•
x2 = y •
•• x3 + 4 x3 + 7 x 2 + 2 x1 = 5u
x3 = y
Differential equations are

•
x1 = x 2
•
x2 = x3
•
x 3 = −4 x 3 − 7 x2 − 2 x1 + 5u
State equation

⎡• ⎤
⎢ x1 ⎥ ⎡ 0 1 0 ⎤ ⎡ x1 ⎤ ⎡0⎤
⎢• ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎢ x2 ⎥ = ⎢ 0 0 1 ⎥ ⎢ x 2 ⎥ + ⎢0⎥ [u]
•
⎢ ⎥ ⎢⎣ − 2 − 7 − 4⎥⎦ ⎢⎣ x 3 ⎥⎦ ⎢⎣5⎥⎦
x
⎢ 3⎥
⎣ ⎦
Output equation

⎡ x1 ⎤
y = [1 0 0]⎢⎢ x 2 ⎥⎥
⎢⎣ x 3 ⎥⎦
Method 2

5
T ( s) = 3
s + 4s 2 + 7s + 2

5
T ( s) = s3
⎛ 4 7 2⎞
1− ⎜− − 2 − 3 ⎟
⎝ s s s ⎠
 Method 2
State equation

⎡• ⎤
⎢ x1 ⎥ ⎡ − 4 1 0⎤ ⎡ x1 ⎤ ⎡0⎤
⎢• ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎢ x 2 ⎥ = ⎢ − 7 0 1⎥ ⎢ x 2 ⎥ + ⎢0⎥ [u]
⎢ • ⎥ ⎢⎣ − 2 0 0⎥⎦ ⎢⎣ x 3 ⎥⎦ ⎢⎣5⎥⎦
x
⎢ 3⎥
⎣ ⎦
Output equation

⎡ x1 ⎤
⎢
y = [1 0 0]⎢ x 2 ⎥ ⎥
⎢⎣ x 3 ⎥⎦
 Solution of state equations
State equation
Output equation

X (t ) = e X 0 At

where

e At
= φ (t ) Is called state transition matrix

X0 Initial state

−1 −1
e At
= φ ( t ) = L [( sI − A) ]
Properties of state transition matrrix

1. φ (0) = e A*0 = e 0 = I = unitmatrix

− At −1 −1
2. φ (t ) = e At
= (e ) = [φ ( − t )]

A( t 1 + t 2 )
3. φ ( t1 + t 2 ) = e =e At1 At 2
e = φ ( t1 ).φ ( t 2 )
1. A linear time-invariant system is characterized by homogeneous equation

⎡ . ⎤ ⎡1 0⎤ ⎡ x ⎤
⎢ x. 1 ⎥ = ⎢ ⎥ ⎢
1
⎥
⎢ x ⎥ ⎣1 1⎦ ⎣ x 2 ⎦
⎣ 2⎦
Compute the solution, assuming initial state

⎡1 ⎤
X0 = ⎢ ⎥
⎣ 0⎦
Solution of the state equation is

X (t ) = e X 0At

state transition matrix is

−1 −1
e At
= φ ( t ) = L [( sI − A) ]

⎡1 0⎤ ⎡1 0⎤ ⎡ s − 1 0 ⎤
sI − A = s ⎢ ⎥ −⎢ ⎥ =⎢ ⎥
⎣0 1⎦ ⎣1 1⎦ ⎣ − 1 s − 1⎦
 −1 1 ⎡s − 1 0 ⎤
( sI − A) = φ ( s) = 2 ⎢ ⎥
( s − 1) ⎣ 1 s − 1⎦

⎡ 1 ⎤
⎢ s −1 0 ⎥
φ ( s) = ⎢ 1 1 ⎥
⎢ ⎥
⎢⎣ ( s − 1) 2 s − 1 ⎥⎦
 ⎡ et 0⎤
e At
= φ (t ) = ⎢ t t⎥
⎣ te e ⎦
The solution of the given state equation

⎡ e t ⎤
0 ⎡1 ⎤ ⎡ e t ⎤
X (t ) = e X 0 = ⎢ t
At
t ⎥⎢ ⎥
= ⎢ t⎥
⎣ te e ⎦ ⎣0⎦ ⎣ te ⎦
 State space to transfer function

Y ( s) −1
= C ( sI − A) B + D
U ( s)
2. Compute state transition matrix At
e

⎡ 0 1 ⎤
A= ⎢ ⎥
⎣ − 2 − 3⎦

⎡1 0⎤ ⎡ 0 1 ⎤ ⎡s − 1 ⎤
sI − A = s ⎢ ⎥ −⎢ ⎥ =⎢ ⎥
⎣ 0 1 ⎦ ⎣ − 2 − 3⎦ ⎣ 2 s + 3⎦
 −1 1 ⎡ s + 3 1⎤
( sI − A) = φ ( s) = ⎢ ⎥
( s + 1)( s + 2) ⎣ − 2 s ⎦
Taking partial fraction

⎡ 2 1 1 1 ⎤
−
⎢ ( s + 1) ( s + 2) − ⎥
( s + 1) ( s + 2)
φ ( s) = ⎢ ⎥
⎢ −2 + 2 −1
+
2 ⎥
⎢⎣ ( s + 1) ( s + 2) ( s + 1) ( s + 2) ⎥⎦
 ⎡ 2e − t − e −2 t e −t
−e ⎤ −2 t
e At
= φ (t ) = ⎢ −t − 2t −t − 2t ⎥
⎣ − 2e + 2e − e + 2e ⎦
 CONTROLLABILITY
A system is said to be completely controllable, if every state
variable of the system can be controlled to get the desired output in
finite time.
 Gilbert’s method of testing controllability

Case 1: When system matrix has distinct eigenvalues.

Consider the state model of a system
•
X = AX + BU
Y = CX + DU
The state model can be converted to canonical form by a
transformation, X=MZ, where M is modal matrix and Z is the
transformed state variable vector
The transformed state model of a system
• ~
Z = ΛZ + B U
~
Y = C Z + DU
where

Λ = M −1 AM The necessary and sufficient condition for ~

completely controllable is that, the matrix B
~
must have no row with all zeros, if so the
B = M −1 B
corresponding row state variable is
~
uncontrollable
C = CM
Case 2: When system matrix has repeated eigenvalues.
Consider the state model of a system
•
X = AX + BU
Y = CX + DU

The state model can be converted to canonical form by a

transformation, X=MZ, where M is modal matrix and Z is the
transformed state variable vector
The transformed state model of a system
• ~
Z = JZ + B U
~
Y = C Z + DU
where
−1 The necessary and sufficient condition for completely
J = M AM controllable is that, none of the rows of ~ be zero, if
~ so the corresponding row state B variable is
B = M −1 B uncontrollable
~
C = CM
 OBSERVABILITY
A system is said to be completely observable, if every state
variable of the system can be measured in finite time.
 OBSERVABILITY
Gilbert’s method of testing controllability

Case 1: When system matrix has distinct eigenvalues.

Consider the state model of a system
•
X = AX + BU
Y = CX + DU
The state model can be converted to canonical form by a
transformation, X=MZ, where M is modal matrix and Z is the
transformed state variable vector
The transformed state model of a system
• ~
Z = ΛZ + B U
~
Y = C Z + DU
where

Λ = M −1 AM The necessary and sufficient condition for

completely obsevable is that, the matrix ~ must
~
have no column with all zeros, C
B = M −1 B
if so the corresponding state variable is
~
unobservable
C = CM
Case 2: When system matrix has repeated eigenvalues.
Consider the state model of a system
•
X = AX + BU
Y = CX + DU

The state model can be converted to canonical form by a

transformation, X=MZ, where M is modal matrix and Z is the
transformed state variable vector
The transformed state model of a system
• ~
Z = JZ + B U
~
Y = C Z + DU
where
−1 The necessary and sufficient condition for completely
J = M AM observable is that, none of the columns of ~ be zero,
~ C
B = M −1 B if so the corresponding state variable is unobservable
~
C = CM
1.

Solution
2.

Solution
Converting the given state model to canonical form

Consider the state model of a system

•
X = AX + BU
Y = CX + DU
Where A matrix is in companion or bush form

⎡ 0 1 0 0 0 ⎤
⎢ 0 0 1 0 0 ⎥ ⎥
⎢
⎢ . . . . . ⎥
⎢ ⎥
⎢ . . . . . ⎥
⎢⎣ − a n − a n −1 . . − a1 ⎥⎦
The state model can be converted to canonical form by a
transformation, X=MZ, where M is modal matrix and Z is the
transformed state variable vector

Case 1: When system matrix has distinct eigenvalues.

The transformed state model (canonical form) of a system

• ~ where
Z = ΛZ + B U
Grammier matrix Λ = M −1 AM
~
Y = C Z + DU
~
B = M −1 B
~
The characteristic equation C = CM
Case 2: When system matrix has repeated eigenvalues.

The state model can be converted to canonical form by a

transformation, X=MZ, where M is modal matrix and Z is the
transformed state variable vector

The transformed state model (Jordan canonical form) of a system

• ~
Z = JZ + B U
~
Y = C Z + DU

where

J = M −1 AM
~
B = M −1 B
~ The characteristic equation
C = CM
3. Convert the given state model to cononical form
The transformed state model of a system
• ~
Z = ΛZ + B U where Λ = M −1 AM
~
Y = C Z + DU ~
B = M −1 B
~
C = CM