B.I.T.

SINDRI
DHANBAD, JHARKHAND

Module-II
STATE SPACE ANALYSIS

Dr. S. K. Verma
Assistant Professor
Department of ECE, BIT Sindri
 Inside
 Introduction
 Need of State Space Analysis
 Basic Definitions
 General form of State Space Model and its Dimension
 State Space Model form Differential Equation
 State Space Model form Transfer Function
 Transfer Function from State Space Model
 State Transition Matrix and it’s Properties
 Concepts of Controllability and Observability

Dr. S. K. Verma Department of Electronics & Communication Engineering

 Introduction
Plant

Mathematical Model
(Differential Equation) (Linear, Time-invariant system)

Frequency Domain Time Domain

Technique Technique

 Two approaches are followed for design and analysis of control system:
 Classical Technique or, Frequency Domain Technique.(T.F. + Graphical plots
like root locus, Nyquist, bode etc.)
 Modern Technique or, Time Domain Technique (State variable approach).

Dr. S. K. Verma Department of Electronics & Communication Engineering

 Need of Conversion of Transfer
Function Form into State Space Form
 A transfer function can be easily fitted to the determined experimental data in b
est possible manner.
 In state variable we have so many design techniques available for system. Hence in ord
er to apply these techniques T.F. must be realized into state variable model.
 For transient response simulation frequency domain design methods are not helpfu
l.
 For that It is must to invert design from s-domain to t-domain but there is not m
uch software for this.
 Hence it is better to convert transfer to state variable model and numerically integr
ating the resulting differential equations rather than attempting to compute the inver
se Laplace transform by numerical methods.
 There are certain limitations in classical design techniques which can be overcome by t
ime domain technique or state variable approach.

Dr. S. K. Verma Department of Electronics & Communication Engineering

 Basic Definitions
State
It is a group of variables, which summarizes the history of the system in order to predi
ct the future values (outputs).

State Variable
The number of the state variables required is equal to the number of the energy storag
e elements present in the system.
Examples: current flowing through inductor, voltage across capacitor

State Vector
It is a vector, which contains the state variables as elements.

Dr. S. K. Verma Department of Electronics & Communication Engineering

 General Form of State Space Model

A=System Matrix
B=Input Matrix
C=Output Matrix
D=Feed-forward Matrix

Dr. S. K. Verma Department of Electronics & Communication Engineering

 Dimensions of State Space Model

The state space model can be obtained

from any one of the two mathematical
models (Differential Equation Model
and Transfer Function Model).

Dr. S. K. Verma Department of Electronics & Communication Engineering

 State Space Model from Differential Equation
 Consider the following series of the RLC circuit.
 It is having an input voltage, vi(t) and the current flowing
through the circuit is i(t).
 It consist of two storage elements (L & C).
 So, the number of the state variables is equal to two and
these state variables are the current flowing through the
inductor, i(t) and the voltage across capacitor, vc(t).
 Since, the output voltage, v0(t) is same as vc(t) .i.e the
voltage across capacitor.
v0(t) = vc(t)
Applying KCL in the loop we get:

Dr. S. K. Verma Department of Electronics & Communication Engineering

 Expression for voltage across the capacitor is:

Differentiating both side with respect to time,

Hence, the state vector can be written as:

So, the Differential State Vector will be:

Dr. S. K. Verma Department of Electronics & Communication Engineering

 So, arrange the differential equations and output equation into the standard form:

where,

Dr. S. K. Verma Department of Electronics & Communication Engineering

 State Space Model from Transfer Function

 Consider the generalized transfer functions of a system given as:

Y (s) bn s n  bn 1s n 1  .........  b1s  b0
G (s)  
U ( s ) an s n  an 1s n 1  ..............  a1s  a0
Y ( s)  1 
  bn s  bn 1s  .........  b1s  b0 
n n 1
  n n 1
U ( s )  an s  an 1s  ..............  a1s  a0 

which can be re-written as:

Y (s)  V (s)   Y ( s) 
  
U ( s ) U ( s )   V( s ) 

V ( s) 1 Y ( s)
where,  n n 1
;&  bn s n  bn 1s n 1  .........  b1s  b0
U ( s ) an s  an 1s  ..............  a1s  a0 V ( s)

Dr. S. K. Verma Department of Electronics & Communication Engineering

 V ( s) 1
 
U ( s ) an s n  an 1s n 1  ..............  a1s  a0
So, after rearranging we get:

a s n
n
 an 1s n 1  ..............  a1s  a0  V ( s )  U ( s )
Applying inverse Laplace transform we get:
d n v(t ) d n 1v(t ) dv(t )
n
 a n 1 n 1
 ..........  a1  a0 v(t )  u (t )
dt dt dt

Let, dv(t ) d 2 v(t )

v(t )  x1 ;  x2  x1 ;  x3  x2 ;...............;
dt dt 2
d n 1v(t ) d n v(t )
n 1
 xn  xn 1 ; & n
 xn
dt dt

Dr. S. K. Verma Department of Electronics & Communication Engineering

  Let, u(t)=u
Then, the state equation will be:

xn   a0 x1  a1 x2  .........  an 1 xn  u
 Again consider other part of the equation
Y ( s)
 bn s n  bn 1s n 1  .........  b1s  b0
V (s)
 Rearranging it again,
Y ( s )   bn s n  bn 1s n 1  .........  b1s  b0 V ( s)

 So, after inverse Laplace transform it will give,

d n v(t ) d n 1v(t ) dv(t )
y (t )  bn  bn 1  ..........  b1  b0 v(t )
dt n dt n 1 dt

Dr. S. K. Verma Department of Electronics & Communication Engineering

 After substituting the values of state variable we get:
y (t )  bn x n  bn 1 xn  ..........  b1 x 2  b0 x1
Substitute the value of xn
y(t )  bn  a0 x1  a1 x2  .........  an1 xn  u   bn1xn  bn2 xn1  .........  b1 x2  b0 x1
or, y(t )   b0  b1a0  x1   b1  b2 a1  x2  ..........   bn1  bn an1  xn  bnu
So, the state space model will be:
 x1   0 1 0 .. .. 0 0   x1   0 
 x   0 0 1 .. .. 0 0   x2   0 
 2  
 .   . 0 0 .. .. 0 0  .   . 
      
X   .    . 0 0 .. .. 0 0   .    .  [u ]
 .   . 0 0 .. .. 1 0  .   . 
      
 xn 1   0 0 0 .. .. 0 1   xn 1   0 
 x   a a1 a2 .. .. an  2 an 1   xn  b0 
 n   0
Dr. S. K. Verma Department of Electronics & Communication Engineering
  x1   x1 
x  x 
 2   2 
 :   : 
Y  b0  bn a0 b1  bn a1 ... ... ... bn2  bn an2 bn1  bn an1    If bn=0 Y  b0 b1 ... ... ... bn  2 bn 1   
 : 
 xn1   : 
   xn 1 
 xn   
 xn 
Special Case:
If the numerator of the Transfer function having constant term only.
i.e. all bn=bn-1=bn-2=..........=b1=0

Y (s) b0  1 
G (s)       b0 
U ( s ) an s n  an 1s n 1  ..............  a1s  a0  an s n  an 1s n 1  ..............  a1s  a0 

Dr. S. K. Verma Department of Electronics & Communication Engineering

 So, the state space model will become:
 x1   0 1 0 .. .. 0 0   x1   0 
 x   0 0 1 .. .. 0 0  x  0
 2    2   
 .   . 0 0 .. .. 0 0  .   . 
      
X   .    . 0 0 .. .. 0 0   .    .  [u ]
 .   . 0 0 .. .. 1 0  .   . 
      
 xn 1   0 0 0 .. .. 0 1   xn 1   0 
 x   a a1 a2 .. .. an  2 an 1   xn  b0 
 n   0

 x1 
x 
 2 
 : 
Y  b0 0 0 ... ... 0 0  
 : 
 xn 1 
 
 xn 
Dr. S. K. Verma Department of Electronics & Communication Engineering
 Example
 Find the state space model for the system having transfer function.

Y (s) 1
 2
U ( s) s  s  1
Solution:
we can re-write the above equation as:
( s 2  s  1)Y ( s )  U ( s )
Taking Inverse-Laplace transform on both the side:
d 2 y (t ) dy (t )
  y (t )  u (t )
dt 2 dt
let, dy (t )
y (t )  x1 &  x 2  x1
dt

Dr. S. K. Verma Department of Electronics & Communication Engineering

 So, the state equations can be written as:

x2   x1  x2  u
and the output equation will be:

y (t )  y  x1
So the state space model can written as:

 x1   0 1   x1  0 
x           u
 x2   1 1  x2  1 
 x1 
Y  1 0  
 x2 
Dr. S. K. Verma Department of Electronics & Communication Engineering
 Transfer Function from State Space Model

We know the state space model of a Linear Time-Invariant (LTI) system is written as:
x  Ax  Bu
y  Cx  Du
Apply Laplace Transform on both sides of the state equation:
sX ( s )  AX ( s )  BU ( s)
 ( sI  A) X ( s )  BU ( s )
 X ( s )  ( sI  A) 1 BU ( s )
Apply Laplace Transform on both sides of the output equation:
Y(s)  CX(s)  DU(s)
Substituting the value of X(s) we get:
Y (s) Y (s)
 C( sI  A) 1 B  D If D=0  C( sI  A) 1 B
U ( s) U ( s)
Dr. S. K. Verma Department of Electronics & Communication Engineering
 Example
Find the transfer function of the system given by the state space model as:

 x1   1 1  x1  1 
x           u
 x2   1 0   x2  0 
 x1 
Y   0 1  
 x2 
Solution:
Given,

 1 1 1 
A  ; B    ; C   0 1 ; and D   0
1 0 0 

Dr. S. K. Verma Department of Electronics & Communication Engineering

 So, the transfer function can be calculated as:
1
Y ( s) 1  s  1 1 1 
 C  sI  A  B   0 1    
U ( s)  1 s   0 
 s 1 
1 s  1 1 
Y ( s)  
   0 1   
U ( s) ( s  1) s  1(1) 0 

Y (s) 1
  2
U (s) s  s  1

Dr. S. K. Verma Department of Electronics & Communication Engineering

 State Transition Matrix
 If the system is producing output in its initial conditions.
 Since, this output is present even in the absence of input, it is called zero input response
xZIR(t).
 Mathematically, we can write it as,

 1
xZIR (t )  e At X (0)  L1  sI  A X (0) 
 From the above relation, we can write the state transition matrix (t) as


 (t )  e At  L1  sI  A
1

 So, the zero input response can be obtained by multiplying the state transition matrix (t)
with the initial conditions matrix i.e. (t).X(0).

Dr. S. K. Verma Department of Electronics & Communication Engineering

 Properties of State Transition Matrix
 If t=0, then state transition matrix will be equal to an Identity matrix.
 (0)  I
 Inverse of state transition matrix will be same as that of state transition matrix
just by replacing ‘t’ by ‘-t’.
 1 (t )   (t )
 If t=t1+t2 , then the corresponding state transition matrix is equal to the multipli
cation of the two state transition matrices at t=t1 and t=t2.

 (t1  t2 )   (t1 ). (t2 )

Dr. S. K. Verma Department of Electronics & Communication Engineering

 Controllability and Observability
Controllability:
 A system is said to be controllable if the initial states of the control system are transferred
(changed) to some other desired states by a controlled input in finite duration of time.
 We can check the controllability of a control system by using Kalman’s Test as:
Write the matrix Qc in the following form.

QC   B AB A2 B ...... An 1 B 
Here, n is order of the system.
 Find the determinant of matrix QC, if it is not equal to zero, then the control system is
controllable, otherwise the control system is said to be uncontrollable.

Dr. S. K. Verma Department of Electronics & Communication Engineering

 Observability:
 A system is said to be Observable if it is able to determine the initial states of the control
system by observing the outputs in finite duration of time.
 We can check the Observability of a control system by using Kalman’s Test as:
Write the matrix Qo in the following form:
2 n 1
QO  C T ACT T
A T
C T
...... A 
T
CT 
 
Here, n is order of the system.
 Find the determinant of matrix QO, if it is not equal to zero, then the control system is Ob
servable , otherwise the control system is said to be Unobservable.

Dr. S. K. Verma Department of Electronics & Communication Engineering

 Example
 Verify the Controllability and Observability of a control system represented in the state
space model as:

 x   1 1  x1  1 
x   1      x   0  u

x
 2  1 0  2  
x 
Y   0 1  1 
 x2 
Solution:
 1 1 1 
Given A  
1 0  ; B  0  ; C   0 1 and D  [0] and n=0
   

So, we can calculate the Controllability and Observability Matrix as:

QC   B AB  and QO  C T AT C T 

Dr. S. K. Verma Department of Electronics & Communication Engineering

  1 1 1   1  1 1   0  1 
 AB         A C  
T T
     
 1 0  0   1   1 0  1  0 

1 1 1 1 
 QC   B AB      QO  C T T T
A C    
 0 1  0 0

 QC  1  0  QO  0

The given system is Controllable but Unobservable.

Dr. S. K. Verma Department of Electronics & Communication Engineering

 References

 Linear Control System by B. S. Manke

 Control System Engineering by J. J. Nagrath & M. Gopal
 Modern Control System with Advanced topics- S. K. Bh-
aradwaj and S. K. Nagar, New Age Publication.
 Feedback Control System by R. A. Barapate
 Modern Control Engineering by K. Ogata

 [Link]
orical-review-and-types- of-control-engineering/
 [Link]
 [Link]

Dr. S. K. Verma Department of Electronics & Communication Engineering

Email: [Link]@[Link]; skv.1986@[Link];
 Block Diagram

Dr. S. K. Verma Department of Electronics & Communication Engineering