You are on page 1of 92

# EEE 4018 ADVANCED CONTROL THEORY

Review of State Models for Systems

Dr. P. Kavitha
Associate Professor
SELECT, VIT University
kavitha.p@vit.ac.in
 Katsuhiko Ogata, "Modern Control Engineering ", PHI Learning Pvt
Ltd,5th Edition, 2010

 Richard C. Dorf, Robert H. Bishop, "Modern Control Systems",
Prentice Hall, 12th Edition,

 Hassan K Khalil, "Nonlinear Control ", Pearson Prentice Hall, 1st
Edition, 2014.

2 EEE 403 Introduction to State Space Analysis 12/19/2017
Overview
 Introduction

 Modeling in state space

 Canonical forms

 State space to transfer function conversion

 Solution to State Equations

 State Transition matrix

3 EEE 403 Introduction to State Space Analysis 12/19/2017
Introduction
 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 input signals are known.

4 EEE 403 Introduction to State Space Analysis 12/19/2017
Introduction
Transfer function method

 Initial value is zero

 For LTIV systems only

State variable method

 Initial values are considered.

 For nonlinear, time-varying and Multivariable systems

5 EEE 403 Introduction to State Space Analysis 12/19/2017
State-Space Modeling
State-Space Modeling
 Alternative method of modeling a system than
 Differential / difference equations
 Transfer functions

 Uses matrices and vectors to represent the system
parameters and variables

7 EEE 403 Introduction to State Space Analysis 12/19/2017
 Easier for computers to perform matrix algebra
 e.g. MATLAB does all computations as matrix math

 Handles multiple inputs and outputs

 Provides knowledge of internal variables (states)

 Primarily used in complicated, large-scale systems

8 EEE 403 Introduction to State Space Analysis 12/19/2017
Transfer Functions vs. State-Space Models
 Transfer functions provide only input and output behavior

 No knowledge of the inner workings of the system

 System is essentially a “black box” that performs some
functions

X(s) H(s) Y(s)

 State-space models also represent the internal behavior of the
system

9 EEE 403 Introduction to State Space Analysis 12/19/2017
Definitions
u – Input vector y – Output vector
 Can be multiple inputs • A function of the input and the
 Written as a column vector present state of the internal
variables

 u1t  
 y1 t  
   y t  
 
y t    2 
u t
u t    2 
     
 
u R t   
 yM t 

10 EEE 403 Introduction to State Space Analysis 12/19/2017
Definitions

x – State vector x – “Next state” vector
 Information of the current • Derivative of the state vector
condition of the internal • Provides knowledge of where
variables the states are going
 N is the “dimension” of the state – Direction (+ or -)
model (number of internal state – How fast (magnitude)
variables)
• A function fo the input and the
present state of the internal
 x1 t  
variables
 x t    x1t  
xt    2  
  
   x
x t    2 
t
    
 xN t   
 x N t 
11 EEE 403 Introduction to State Space Analysis 12/19/2017
State-Space Equations
General form of the state-space model
x t 
Two equations –
y t 

x t   f xt , u t , t 
y t   g xt , u t , t 

12 EEE 403 Introduction to State Space Analysis 12/19/2017
Linear State-Space Equations
x t   Axt   But 
y t   Cx t   Du t 

xt , x t   N 1 vectors A  N  N system matrix
vt   R 1 vector B  N  R input matrix
y t   M 1 vector C  M  N output matrix
D  M  R matrix representing direct
coupling from system inputs
to system outputs

If A, B, C, D are constant over time, then the system is also time invariant
→ Linear Time Invariant (LTI) system
13 EEE 403 Introduction to State Space Analysis 12/19/2017
Construction of State Equations from a Differential
Equation
(Let there be no derivatives of the input)
 The dimension of the state equations (number of state variables) should
equal the order of the differential equation
 Let one state variable equal the output (y(t))
 Let one state variable equal the derivative of the output

 Let one state variable equal the (N-1)-th derivative of the output
(where N is the order of the differential equation)
 Find the derivative of each of the newly defined state equations
 In terms of the other state variables and the outputs
 Write the state equations

14 EEE 403 Introduction to State Space Analysis 12/19/2017
Some Definitions

 System variable: any variable that responds to an input or initial conditions in a
system

 State variables: the smallest set of linearly independent system variables such that
the values of the members of the set at time t0 along with known forcing functions
completely determine the value of all system variables for all t ≥ t0

 State vector: a vector whose elements are the state variables

 State space: the n-dimensional space whose axes are the state variables

 State equations: a set of first-order differential equations with ‘b’ variables, where
the ‘n’ variables to be solved are the state variables

 Output equation: the algebraic equation that expresses the output variables of a
system as linear combination of the state variables and the inputs.

15 EEE 403 Introduction to State Space Analysis 12/19/2017
d
x(t )  Ax (t )  Bu(t ) State equation
Dynamic equation dt
y (t )  Cx(t )  Du(t ) Output equation

State variable

 x1 (t )   u1 (t )   y1 (t )   x1 (0) 
 x (t )  u (t )  y (t )   x (0)
x(t )   2  u (t )   2  y (t )    x(0)   2 
2

           
       
 x n (t ) n1 ur (t )  r1  p  p1
y (t )  n  n1
x ( 0)

State space r- input p- output

       
A   nn 
 B   nr 
 C   pn 
 D   pr 

       

16 EEE 403 Introduction to State Space Analysis 12/19/2017
u1 (t ) y1 (t )
u2 (t ) Inner state variables y2 (t )
 x1 , x2 , xn 
ur (t ) y p (t )

D

+
u (t ) + x (t ) 1 x(t ) y (t )
B C
s +
-

A

17 EEE 403 Introduction to State Space Analysis 12/19/2017
Example
y2 y1
K

M2 M1 f (t )

B3 B2 B1

M 1 y1  B1 y1  B2 ( y1  y 2 )  K ( y1  y 2 )  f (t )
M 2 y2  B3 y 2  B2 ( y 2  y1 )  K ( y 2  y1 )  0

 
 
x1  y1  y 2  x1   0
 1 1   x1   0 
 x    K ( B  B2 ) B2   x    1  f (t )
x2  y1  2  M  1
let M1 M1   2   M1 
 x3    
( B2  B3 )   x3   0 
1
x3  y 2 
K B2
 
 M 2 M2 M 2 

18 EEE 403 Introduction to State Space Analysis 12/19/2017
Example By Newton’s Law

F  My  My  by  ky  r
r  ky  by  My let x1  y, x2  y
k

b
x1  y  x2
M 
b k 1
x2  y   y  y r
y (t ), y (t ) M M M
r (t ) b
  x2 
k 1
x1  u u  r 
M M M
x1  x 2
 k b 1
x 2   x1  x2  u
M M M

 x1   0 1  x   0 
x 
  k b  1    1   u y  1 0 1   0  u
 x2   M    x2   
M M   x2 
C
A B

19 EEE 403 Introduction to State Space Analysis 12/19/2017
State Space Equation
 x1   0 1  x   0 
x  Ax  Bu For example :     k b   1    1 u
 x2   M    x2   
M M 
y  Cx  Du
x 
y  1 0 1   0  u
 x2 
Transfer Function
Y ( s) b1s  b0
G( s)  For example : G ( s ) 
U ( s) a2 s 2  a1s  a0

Example: Transfer function of the Mass-damper-spring system
d2y dy
M 2  b  ky  u (t ) x  x1 x2 
T
dt dt
Ms 2Y ( s)  bsY ( s )  kY ( s )  U ( s )
Y (s) 1
 G(s) 
U ( s) Ms 2  bs  k

20 EEE 403 Introduction to State Space Analysis 12/19/2017
Remark : the choice of states is not unique.

+ L +
R
di(t ) 1 t
ei (t ) c ec (t ) Ri(t )  L   i(t )dt  ei (t )
dt c 0
- i (t ) -

x1 (t )  i(t )  x1   R 
1  x   1 
 x    L LC      L ei (t )
1
let  2   1 0   x2   0 
x2 (t )   i(t )dt
x 
y(t )  i(t ) y (t )  1 0 1 
 x2 
 R R
xˆ1 (t )  i (t )  xˆ1   L
    ˆ
x  1
  ˆ    L  ei (t )
L
    1
1
let
xˆ2 (t )  ec (t )  xˆ2   0   x2   0 
exist a mapping  L 
y(t )  i(t )  xˆ1 
y (t )  1 0 
 xˆ2 
21 EEE 403 Introduction to State Space Analysis 12/19/2017
Modeling in state space
Example 1. Obtain the state model using physical variables.
In mechanical systems Displacement (spring) and velocity (mass) are state
variables

 No. of state variables = No. of storage elements = 4
 No. of mass elements = 2 ∴Two velocities y1(t ) and y 2 (t ) are chosen as state
variables
 No. of spring elements = 2 ∴Two displacements y1(t) and y2(t) are
chosen as state variables

23 EEE 403 Introduction to State Space Analysis 12/19/2017
 The state variables are

 Two differential equations are

24 EEE 403 Introduction to State Space Analysis 12/19/2017
 State equation is

 Algebraic output equation is

25 EEE 403 Introduction to State Space Analysis 12/19/2017
26 EEE 403 Introduction to State Space Analysis 12/19/2017

27 EEE 403 Introduction to State Space Analysis 12/19/2017

28 EEE 403 Introduction to State Space Analysis 12/19/2017
 State Variables can be

1. Controllable Canonical Form

2. Observable Canonical Form

3. Diagonal Canonical Form (or)

4. Jordan Canonical Form

29 EEE 403 Introduction to State Space Analysis 12/19/2017
Modeling –canonical forms

1. Controllable canonical form (CCF)
2 . Observable canonical form (OCF)
3 . Diagonal & Jordan canonical form (DCF or JCF)
Method 1 and Method 2

Consider n=m

31 EEE 403 Introduction to State Space Analysis 12/19/2017
Method 1 CCF
State Equation

Output Equation

32 EEE 403 Introduction to State Space Analysis 12/19/2017
Method 2 OCF
State Equation

Output Equation

33 EEE 403 Introduction to State Space Analysis 12/19/2017
METHOD 3 DCF

If n=m;

C1, C2,… Cn are residues
λ1, λ2, … λn are roots of the denominator polynomial

34 EEE 403 Introduction to State Space Analysis 12/19/2017
 State Equation

 Output Equation

35 EEE 403 Introduction to State Space Analysis 12/19/2017
Method 4 : Jordan Canonical Form

The partial fraction expansion becomes

36 EEE 403 Introduction to State Space Analysis 12/19/2017
 x1  1 1 0   x1  0
 x   0 1 1 0   x  0 
 2   2   
 x3   0 0 1   x3  1
 x   2 1   x  0 
 4     4    u
 x5   0 2   x5  1
 x6   0 3   x6  1
      
      

   
 
      
 x1 
x  Jordan block
 2
 x3 
x 
y  c1 c2 c3 c4 c5 c6  4 
 x5 
 x6 
 
 
EEE 403 Introduction to State Space Analysis

   12/19/2017
37
Example 1
A feedback system has a closed loop transfer function. Write in three different state
space representation

38 EEE 403 Introduction to State Space Analysis 12/19/2017
 Method 1 CCF
State Equation

Output Equation

39 EEE 403 Introduction to State Space Analysis 12/19/2017
 Method 2 OCF

State Equation

Output equation

40 EEE 403 Introduction to State Space Analysis 12/19/2017
Method 3

41 EEE 403 Introduction to State Space Analysis 12/19/2017
Method 3 Diagonal Canonical form
State Equation

Output Equation

42 EEE 403 Introduction to State Space Analysis 12/19/2017
Example 2
A feedback system has a closed loop transfer function.

Write three different state space representation.

43 EEE 403 Introduction to State Space Analysis 12/19/2017
Example 4 JCF

1 0 1 0.5 0.5
G( s)     
( s  1) ( s  1) ( s  1) s s  2
3 2

 x1   1 1 0   x1  0
 x   0  1 1 0   x  0 
 2   2   
 x3   0 0 1   x3  1
 x    0   x4   1u
 4     
 x5   0  2  x5  1
      
      
      
 x1 
x 
 2
 x3 
y   1 0  1 0.5 0.5 x4 
 
 x5 
EEE 403 Introduction to State Space Analysis
 
44   12/19/2017
 
State Space to transfer function
Dynamical equation Transfer function

x (t )  Ax (t )  Bu(t )
y (t )  Cx(t )  Du(t )
Laplace transform
sX ( s)  x(0)  AX ( s)  BU ( s )
Y ( s )  CX ( s )  DU ( s )

assume x(0)  0
X ( s )  ( sI  A) 1 BU ( s )
Y ( s )  [C ( sI  A) 1 B  D]U ( s )
matrix
Transfer function

46 EEE 403 Introduction to State Space Analysis 12/19/2017
Example1  x1   0 1 0   x1  0 0
 x    0  4 3   x   1 0   u1 
 2   2    u 
 x3   1  1  2  x3   0 1  2 
MIMO system
 x1 
 y1 (t )  1 0 0  
 y (t )  0 0 1  x2 
 2   x 
 3

( sI  A) 1 
sI  A
 s 2  6s  11 s  2 3 
1  
  3 s 2
 2 3s
s ( s  4)( s  2)  3  3s  
 s4  s  1 s  4s 
2

G( s)  [C ( sI  A) 1 B  D]
1  s2 3  Transfer function

s 3  6s 2  11s  3  ( s  1) s( s  4)

47 EEE 403 Introduction to State Space Analysis 12/19/2017
Example 2

  0 1  0
 x    x   u
  2  3 1
 y  1 3x

D  0, C  1 3
0 1 0
A  , B 
 2  3 1

48 EEE 403 Introduction to State Space Analysis 12/19/2017
H ( s)  D  C ( sI  A) 1 B
1
 1 0  0 1   0 
 0  1 3 s      1
  0 1    2  3  
1
 s  1  0 
 1 3  1
 2 s  3  
 s  3 1 0
 1 3
1
s( s  3)  2   2 s  1
1
 1 3
1
s( s  3)  2  s 
1

1
1 3 
s ( s  3)  2 s 
3s  1 3s  1
  2
s ( s  3)  2 s  3s  2
49 EEE 403 Introduction to State Space Analysis 12/19/2017
Solution to state space rep.

Solution:

50 EEE 403 Introduction to State Space Analysis 12/19/2017
Solution to State Equation
d
x(t )  Ax (t )  Bu(t )
dt
y (t )  Cx(t )  Du(t )

The behavior of x(t) et y(t) :
1. Homogeneous solution of x(t)
2. Non-homogeneous solution of x(t)

52 EEE 403 Introduction to State Space Analysis 12/19/2017
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)
State transition matrix

(t )  e At  L1[( sI  A) 1 ]
x(t0 )  e At0 x(0)
 At0
x(0)  e x(t0 )
x(t )  e At e  At0 x(t0 )  e A(t t0 ) x(t0 )   (t  t0 ) x(t0 )
53 EEE 403 Introduction to State Space Analysis 12/19/2017
Properties 
1 1
(t )  e  L [( sI  A) ]
At

1. (0)  I
1
2.  (t )  (t )
3. x(0)  (t ) x(t )
4. (t2  t1 ) (t1  t0 )  (t2  t0 )
(t )  (kt )
k
5.
54 EEE 403 Introduction to State Space Analysis 12/19/2017
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)
X ( s)  ( sI  A) 1 x(0)  ( sI  A) 1 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
55 EEE 403 Introduction to State Space Analysis 12/19/2017
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

56 EEE 403 Introduction to State Space Analysis 12/19/2017
Example 1  x1   0 1   x1  0
 x    2  3  x   1u (t )
 2   2   
x(0)  0 0
T
let

 2e t
 e 2 t
e 1  e 2t 
 (t )  L1[( sI  A) 1 ]  e At   t  2t t  2t 
  2 e  2 e  e  2e 

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

 x1    2e t  e  2t 
1 3
Ans: x    2 2 L1[( sI  A) 1 BU ( s)]
 2    2e  2e  2t 
 t

57 EEE 403 Introduction to State Space Analysis 12/19/2017
State transition matrix
State transition matrix: eAt
x  Ax
solution : x(t)  e At x(0)
1 2 2 1 33 1 n n
e  I  At  A t  A t  ...  A t  ...
At

2! 3! n!
 eAt is an nxn matrix
 eAt =ℒ-1((sI-A)-1), or ℒ (eAt)=(sI-A)-1
d At
 e = AeAt= eAtA
dt
 eAt is invertible: (eAt)-1= e(-A)t
 eA0=I
 eAt1 eAt2= eA(t1+t2)
59 EEE 403 Introduction to State Space Analysis 12/19/2017
How to find State transition matrix

1 1
(t )  e  L [( sI  A) ]
At

Method 1:
(t )  L1[( sI  A) 1 ]

Method 2: Cayley-Hamilton Theorem

60 EEE 403 Introduction to State Space Analysis 12/19/2017
1 1
Method 1:
(t )  L [( sI  A) ]
 x1   0 1 0   x1  0 0
 x    0  4 3   x   1 0   u1 
 2   2    u 
 x3   1  1  2  x3   0 1  2 
 x1 
 1  1 0 0  
y ( t )
 y (t )  0 0 1  x2 
 2   x 
 3

( sI  A) 1 
sI  A
 s 2  6s  11 s  2 3 
1  
  3 s 2
 2 3s
s ( s  4)( s  2)  3  3s  
 s4  s  1 s  4s 
2

61 EEE 403 Introduction to State Space Analysis 12/19/2017
Example 2
 0 1  s 1 
A   , sI  A   
  2  3  2 s  3
 s3 1 
 
1 1  s  3 1   ( s  1)( s  2) ( s  1)( s  2) 
( sI  A)    
s( s  3)  2   2 s   2 s 
 ( s  1)( s  2) ( s  1)( s  2) 

 2 1 1 1 
   
  s 1 s  2 s 1 s  2 
 2  2 1

2 
 
 s 1 s  2 s 1 s  2 
 2 e t
 e  2t
e t
 e  2t

e  
At
t  2t t
u (t )
 2t  s
  2e  2e  e  2e 
62 EEE 403 Introduction to State Space Analysis 12/19/2017
Method 2: Cayley-Hamilton Theorem

Theorem: Every square matrix satisfies its char. equation.

Given a square matrix A, A  R. nLet
n
f(λ) be the char. polynomial of A.

Char. Equation:
f ( )  n  an1n1    a1  a0  0
By Caley-Hamilton Theorem

f ( A)  An  an1 An1    a1 A  a0 I  0

63 EEE 403 Introduction to State Space Analysis 12/19/2017
An  an 1 An 1    a1 A  a0 I  0
An  an 1 An 1    a1 A  a0 I
An 1  an 1 An    a1 A2  a0 A
  an 1 (an 1 An 1    a1 A  a0 I )   a1 A2  a0 A

any
f ( A)  k0 I  k1 A  k2 A2    kn An  
f ( A)   0 I  1 A   2 A     n1 A2 n 1

n 1
 
k 0
k A k

64 EEE 403 Introduction to State Space Analysis 12/19/2017
Example: 1 2
A 100
? A 
0 1 
let f ( A)  A100   0 I  1 A

 1 2
 (  1)(  2)  0 , 1  1, 2  2
0  2

f (1 )   1
100
  0  11  1 100
 0  2  2100
f (2 )   2
100
  0  12  2 100 1  2100  1

1 0  1 2  1 2101
 2
f ( A)  A  (2  2 ) 
100 100
  (2  1) 
100
  
 0 1   0 1  0 1 

65 EEE 403 Introduction to State Space Analysis 12/19/2017
Example:  3  1
e ?
At
A 
2 0
 3 1
 0 , 1  1, 2  2
2 

f (1)  e  t   0  11   0  1  0  2e  t  e 2 t
f (2)  e  2t   0  12   0  1 2 1  e  2 t  e t

t 1 0  2t  2t t   3  1
e  2e  e 
At
  ( e  e )  
 0 1   2 0 
 2e  2t  e t e  2 t  e t 
  2t t  2t t 
 2e  2e  e  2e 

66 EEE 403 Introduction to State Space Analysis 12/19/2017
Example:
2 0  1 1
x  Ax  bu  4 1  4 x  1u
2 0  1 0
y  1 0 1x

1  1 0
  0,1,1 v1  4, v 2  0, 1
2 1 0

67 EEE 403 Introduction to State Space Analysis 12/19/2017
Example:

 0 1 0  0 
x  Ax  bu  0 0 1  x  0u
2 1  2 1
y  1 0 0x

68 EEE 403 Introduction to State Space Analysis 12/19/2017
Controllability and Observability

69 EEE 403 Introduction to State Space Analysis 12/19/2017
Controllability and Observability

 Controllability and observability represent two major concepts of modern control
system theory. These concepts were introduced by R. Kalman in 1960.

 Controllability: In order to be able to do whatever we want with the given
dynamic system under control input, the system must be controllable.
 Observability: In order to see what is going on inside the system under observation,
the system must be observable

 In this lecture we show that the concepts of controllability and observability are
related to linear systems of algebraic equations. It is well known that a solvable
system of linear algebraic equations has a solution if and only if the rank of the
system matrix is full
 Observability and controllability tests will be connected to the rank tests of certain
matrices: the controllability and observability matrices.
70 EEE 403 Introduction to State Space Analysis 12/19/2017
Controllability and Observability

 Controllability deals with whether or not the state of a state-space
equation can be controlled from the input.
 Controllability studies the possibility of steering the state from the
input while observability studies the possibility of estimating the
state from the output. The two concepts are indeed dual as will be
described later.
 Observability deals with whether or not the initial state can be
observed from the output.
 Observability is defined under the assumption that the state equation
or {A,B,C,D} is known. It is different from the problem of
realization or identification, which is to determine or estimate
{A,B,C,D} from the information of inputs and outputs.

71 EEE 403 Introduction to State Space Analysis 12/19/2017
Controllability

Plant: x  Ax  Bu, x  R n
y  Cx  Du
Definition of Controllability

A system is said to be (state) controllable at time t0 , if
there exists a finite t1  t0 such for any x(t0 ) and any x1 ,
there exist an input u[t0 ,t1 ] that will transfer the state x (t0 )
to the state x1 at time t1 , otherwise the system is said to
be uncontrollable at time t0 .

72 EEE 403 Introduction to State Space Analysis 12/19/2017
Controllability:
 x  Ax  Bu

 y  Cx  Du
is completely controllab le if  any x(0),
 control u (t ) which can bring x(t ) to
x  0 in finite time.

Thm : c.c. iff rank([ B | AB | A2 B |  An 1B])  n
n 1
or det[ B | AB  A B]  0 (if B is n 1)
73
or rank( I - A B)  n 
EEE 403 Introduction to State Space Analysis 12/19/2017
Controllability Matrix
 A, B  Controllab le  rank(U )  n,
U B  AB A2 B  An 1 B 
 det(U )  0 if u  R
Controllab ility Matrix U  B  AB A2 B  An1B 
Example: An Uncontrollable System

 1 0  1
x    x    u
 0  3 0
y  1 2x

74 EEE 403 Introduction to State Space Analysis 12/19/2017
Example:1
0 1 0 
A  , B 
 2  3 1
n2
0 |  0 1  0  
[ B AB ]    2  3 1 
1 |    
0 | 1  0 1 
  rank    2
1 |  3  1  3

linearly ind.  rank  2
or det( B AB )  1  0
75 EEE 403 Introduction to State Space Analysis 12/19/2017
In Matlab:
S  [ B | AB | A B ]
2

>> S=ctrb(A,B)
>> r=rank(S)
1 1
e.g. rank    1
 2 2
If S is square (when B is nx1)
>> det(S)

76 EEE 403 Introduction to State Space Analysis 12/19/2017
Observability
A system is said to be (completely state) observable at
time t0 , if there exists a finite t1  t0 such that for any x (t 0 )
at time t0 , the knowledge of the input u[t t ] and the
0, 1

output y[t t ] over the time interval [t0 , t1 ] suffices to
0, 1

determine the state x0 , otherwise the system is said to be
unobservable at t0 .

77 EEE 403 Introduction to State Space Analysis 12/19/2017
Observability Matrix
A, C  Observable  rank(V )  n  det(V )  0 if y  R

 C 
 CA 
 
Observability Matrix V   CA 
2

 
  
CAn 1 

Example: An Unobservable System

0 1  0
x    x   u
0  2 1
y  0 4x

78 EEE 403 Introduction to State Space Analysis 12/19/2017
Observability
 x  Ax  Bu

 y  Cx  Du
is completely obserrable if the knowledge of u (t ), y (t )
over a finite time can enable us to determine x(0).
Without loss of generality , can set u  0
 C   C 
 CA   
   CA 
Thm : c.o. iff rank  n, or det   0 (if C is 1 n)
   
 n 1   
 n 1 
CA   CA 
 I - A 
or rank(   )  n, 
 C 
79 EEE 403 Introduction to State Space Analysis 12/19/2017
Example:2

0 1
A  , C  1 0
 2  3
n2
 1 0 
C   

CA 1 0 0 1 
    2  3 
   
1 0
 
 0 1 
C 
det( )  1  0
 CA 
 c.o.
80 EEE 403 Introduction to State Space Analysis 12/19/2017
In Matlab:
 C 
>> V=obsv(C,A)  CA 
V   2
>> r=rank(V) CA 
 
rank must = n   
Look for controllability
Look for observability
Or if single output (ie V is square), can use
>> det(V)
det must be nonzero

81 EEE 403 Introduction to State Space Analysis 12/19/2017
Example

0 1  0 
x  Ax  Bu, x  R n A  , B   , C  0 1
Plant: 1 0 1
y  Cx  Du

0 1
Controllab ility Matr ix V  B AB    
1 0
 C  0 1 
Obervabili ty Matrix N      
  
CA 1 0
rank(V )  rank( N )  2

Hence the system is both controllable and observable.

82 EEE 403 Introduction to State Space Analysis 12/19/2017
Theorem I

xc (t )  Ac xc (t )  B c u (t )
Controllable canonical form Controllable

Theorem II
xo (t )  Ao xo (t )  B o u (t )
y (t )  Co xo (t )

Observable canonical form Observable

83 EEE 403 Introduction to State Space Analysis 12/19/2017
s2
example T ( s) 
(s  1)( s  2)
0 1 0
xc    xc   u
Controllable canonical form  2  3 1
y  2 1xc
0 1 
U  B 
AB  
1  3
 rank[U ]  2  n
C   2
V  
1

rank[V ]  1  n
  
CA  2  1
0  2  2
xo    xo   u
Observable canonical form 1  3 1
y  0 1xo
 2  2
U  B AB     rank[U ]  1  n
1  1 
 C  0 1 
V   rank[V ]  2  n
 
  
CA 1 3
EEE 403 Introduction to State Space Analysis 12/19/2017
84
Principle of Duality (Due to Kalman)
 To clarify the apparent analogies between controllability and
observability.
 Consider the system described by ‘S1’
𝑋 = 𝐴𝑋 + 𝐵𝑈
Y=CX
 And the dual system is ‘S2’ given by
𝑍 = 𝐴𝑇𝑍 + 𝐶𝑇𝑉
n=BTZ
 The principle of duality states that the system S1 is completely state
controllable if and only if system S2 is completely observable and vice
versa.

85 EEE 403 Introduction to State Space Analysis 12/19/2017
Controllability & Observability

86 EEE 403 Introduction to State Space Analysis 12/19/2017
Not Controllable

87 EEE 403 Introduction to State Space Analysis 12/19/2017
Not Controllable

88 EEE 403 Introduction to State Space Analysis 12/19/2017
Not Observable

89 EEE 403 Introduction to State Space Analysis 12/19/2017
Not Observable

90 EEE 403 Introduction to State Space Analysis 12/19/2017
Summary
 State Space Modeling

 Different forms of representation

 State Space to transfer function

 Solution to State equation

 State transition matrix

 Controllability & Observability

 Stabilizability & Detectability

91 EEE 403 Introduction to State Space Analysis 12/19/2017
End of Slides

BEST WISHES