Scribd
Upload
1K views6 pages

Example Pole Placement and Design of Observer.

This document provides an example MATLAB script for designing a minimum-order observer using pole placement. It begins by defining the system matrices A, B, C and determines that the system is observable. It then calculates the characteristics polynomial for the unobserved subsystem and reduced observability matrix. Finally, it places the desired observer poles and calculates the observer gain matrix Ke. The example demonstrates how to systematically design a minimum-order observer for a multi-dimensional system.

Uploaded by

Ahmed Salam
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
1K views6 pages

Example Pole Placement and Design of Observer.

This document provides an example MATLAB script for designing a minimum-order observer using pole placement. It begins by defining the system matrices A, B, C and determines that the system is observable. It then calculates the characteristics polynomial for the unobserved subsystem and reduced observability matrix. Finally, it places the desired observer poles and calculates the observer gain matrix Ke. The example demonstrates how to systematically design a minimum-order observer for a multi-dimensional system.

Uploaded by

Ahmed Salam
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

18

>> %-------- pole placement and design of observer---------


%--- First , Solving the Pole placement problem --------
% --- Entering Matrices ,A B C and D
A=[0 1;20.6 0];
B=[0;1];
An example on :
C=[1 0];
D=[0]; DESIGN OF A CONTROL SYSTEM USING
% ------- Checking controllability matrix M POLE-PLACMENT METHOD AND A
STATE OBSERVER. FIRST SOLVING THE
M=[B A*B];
POLE-PLACEMENT PROBLEM
rank(M)
ans =
Using MATLAB®

2
>>% since the rank of the matrix M is 2 ,
>> % arbitrary pole placement is possible .
>> % Entering the desired characteristics polynomial by
>> % defining the following matrix J and computing poly(J)...
>> J=[-1.8+2.4*i 0;0 -1.8-2.4*i];
>> poly(J)
ans =
1.0000 3.6000 9.0000
>> % Enter characteristics polynomial Phi
>> Phi=polyvalm(poly(J),A);
>> % State feedback gain matrix K can be given by ......
>> K=[0 1]*inv(M)*Phi
K=
29.6000 3.6000
19

% ***** The following program determines the observer matrix Ke ***


>> % Enter the observability matrix N and check its rank .....
>> N=[C' A'*C'];
rank(N)
ans =
2
>> % since the rank of the observability matrix is 2 , design of
>> % the observer is possible ....
>> % Enter the desired characteristic polynomial by defining
>> % the following matrix J0 and entering statement poly(J0)
>> J0=[-8 0;0 -8];
>> poly(J0)
ans =
1 16 64
% Enter the characteristics polynomial Ph
>> Ph=polyvalm(poly(J0),A);
>> % The observer gain matrix Ke is obtained from
>> Ke=Ph*(inv(N'))*[0;1]
Ke =
16.0000
84.6000
THE ACTUAL m-file PROGRAM IS THE FOLLOWING SCRIPT

Entering system
A=[0 1;20.6 0]; Matrices
B=[0;1];
C=[1 0]; A , B , C and D
D=[0];
M=[B A*B];
rank(M) Checking for controllability
20

J=[-1.8+2.4*i 0;0 -1.8-2.4*i]; 2.4*i];


poly(J)
Phi=polyvalm(poly(J),A); Entering char. . Poly. Phi
K=[0 1]*inv(M)*Phi Solving to find [K]
N=[C' A'*C'];
rank(N) Checking for observability
J0=[-8 0;0 -8];
poly(J0) Entering desired ch. Poly.
Ph=polyvalm(poly(J0),A); Entering ch. Poly. Phi
Ke=Ph*(inv(N'))*[0;1] Solving to find [Ke]
The state equations of the system before designing the feedback controller and the full-order
full
observer were :

Which gives an Unstable response to a step input function ,as shown below :

And this is due to the pole –zero


zero map of the system which has two poles and no zeros
21

X X

Root locus Bode plot


And after the design ,the system became of a forth order , with the characteristics equation for the
system as a whole is :

And the step response has become :

With the root locus and bode plot as shown below :


22

Root locus Bode plot

>> % ------ Enter matrices Aaa,Aab,Aba,Abb,Ba,Bb------


Aaa,Aab,Aba,Abb,Ba,Bb
>> Aaa=[0];Aab=[1 0];Aba=[0;6];Abb=[0 1;-11
1; -6];
>> Ba=[0];Bb=[0;1];
>> % ----- Not that state matrix A and control matrix B are -----
>> A=[Aaa Aab;Aba Abb],B=[Ba;Bb]

A =

0 1 0 A solved example on
0 0 1
6 -11 -6 DESIGNING A MIINMUM-ORDER
MIINMUM
OBSERVER
B =

0
Using MATLAB®
0
1
>> % -----Determine
Determine a1 and a2 of the characteristics polynomial
>> % for the unobserved portion of the system -----------
>> p=poly(Abb)

p =

1.0000 6.0000 11.0000


23

a1=p(2);a2=p(3);
>> %-----Enter the reduced observability matrix NN -----
>> NN=[Aab' Abb'*Aab'];
>> % ----Enter matrix WW -----
>> WW=[a1 1;1 0];
>> %-----Enter the desired characteristics polynomial be defining
>> % the following matrixJ and entering statement poly(j)--------
>> J=[-2+2*sqrt(3)*i 0;0 -2-2*sqrt(3)*i];
>> JJ=poly(J)

JJ =

1.0000 4.0000 16.0000


>> %----Determine aa1 and aa2 of the desired characteristics
>> % polynomial --------
>> aa1=JJ(2);aa2=JJ(3);
>> % ----- observer gain matrix Ke for the minimum-order
>> % --- observer is given by ----
>> Ke=inv(WW*NN')*[aa2-a2;aa1-a1]
Ke =

-2.0000
17.0000

References :
1. Ogata, Katsuhiko ," Linear control systems design with
Matlab® "
2. Ogata, Katsuhiko ," Modern control engineering ,second
edition, Prentice Hall, 1995.

a1 works ©2010

You might also like