Scribd
Upload
131 views5 pages

Lab 4 Observer Design

The objective of this lab is to learn estimator design and implement full state feedback control using a regulator. The lab uses a Luenberger observer, which estimates states based on the system output and mathematical model. The observer gain L is chosen to make the estimation error asymptotically stable. For the inverted pendulum example, the system is verified to be observable and controllable. Controller and observer poles are placed appropriately, and the full compensator is implemented by combining state feedback control with state estimation. This achieves similar closed-loop performance to full state feedback control.

Uploaded by

Muhammad Asaad
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
131 views5 pages

Lab 4 Observer Design

The objective of this lab is to learn estimator design and implement full state feedback control using a regulator. The lab uses a Luenberger observer, which estimates states based on the system output and mathematical model. The observer gain L is chosen to make the estimation error asymptotically stable. For the inverted pendulum example, the system is verified to be observable and controllable. Controller and observer poles are placed appropriately, and the full compensator is implemented by combining state feedback control with state estimation. This achieves similar closed-loop performance to full state feedback control.

Uploaded by

Muhammad Asaad
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 5

Lab Session 4

Objective
The objective of this Lab is to learn the estimator design and implement it with
full state feedback control for regulator.
Apparatus
MATLAB Software
Introduction
The basic estimator design involves the creating the states using the output of the system.
The simplest form of observer is the Luenberger Observer which uses the system
mathematical model and compare the output of the system and the mathematical model and
the difference is amplified and fed back to estimate the correct value of the state.
The differential equation of the observer is
X˙ =A ^
^ X + BU + L(Y −Y^ )
Here ^X ∧Y^ represent estimated states and L is the observer gain. The gain is decided on the
basis of poles of the observer and how fast the error needs to be eliminated.
The goal of the observer is to provide an estimate x so that x → ^x ast → ∞ Remember that we
do not know x (t 0) precisely; therefore, we must provide an initial estimate ^x (t 0) to the
observer. Define the observer estimation error as
e (t )=x ( t )−^x (t)

The observer design should produce an observer with the property that e (t) → 0 as t → 0. One
of the main results of systems theory is that if the system is completely observable, we can
always find L so that the tracking error is asymptotically stable as desired. Taking the time-
derivative of the estimation error yields

ė= ẋ− x^˙

and using the system model and the observer we obtain

ė= Ax+ Bu− A x^ −Bu−L( y−C ^x )

ė (t )= ( A−LC ) e (t)

The error will approach zero (  will approach  ) if the matrix   is stable (has negative
eigenvalues)

The characteristic equation is

det ( λI −( A−LC ) )=0

the observer design process reduces to finding the matrix L such that the roots of the
characteristic equation lie in the left half-plane. This can always be accomplished if the
system is completely observable.
Example:
Consider the Inverted pendulum the system matrix is

System parameters are

• l = 0.098 m
• g=9.8 m/ s2
• m=0.825 kg
• M=8.085 kg

A=[0 1 0 0;0 0 -1 0;0 0 0 1;0 0 100 0];


B=[0;0.1237;0;-1.2621];
C=[1 0 0 0];
D=[0;0;0;0];
Pc=ctrb(A,B);
Po=obsv(A,C);
detPc=det(Pc)
detPo=det(Po)
controllability = rank(Pc) %to check how many states are conrollable
observability = rank(Po)
P=[-3+2i;-3-2i;-5;-6]
Q=[-30+20i,-30-20i,-50,-60];
K=acker(A,B,P)
L=acker(A',C',Q)

MATLAB Result
detPc =

196.5402

detPo =

P =

-3.0000 + 2.0000i
-3.0000 - 2.0000i
-5.0000 + 0.0000i
-6.0000 + 0.0000i

K =

-35.1101 -29.0784 -169.0382 -16.3196

L =

170 11000 -340000 -5000000


TASK
Using given Simulink diagram, design observer controller
Observations

Conclusion

TASK 2: Observer based control

I. Determine K such that det(AI - (A - BK)) = 0 has roots in the left half-plane and place
the poles appropriately to meet the control system design specifications
II. Determine L such that det(AI - (A - LC)) = 0 has roots in the left half-plane and place
the poles to achieve acceptable observer performance
III. Connect the observer to the full-state feedback law using
IV. U(t) = -Kx hat(t).

1. Verify that our system is observable


2. Find controller poles by eig(Ac)
3. Place the observer poles 10 times away than the slowest poles have real part .
4. Calculate L from poles
5. Now we will combine our state-feedback controller from before with our state
estimator to get the full compensator. The resulting closed-loop system is described
by the following matrix equations.

Conclusion
This response is almost identical to the response achieved when it was assumed that we had full access to the
state variables. This is because the observer poles are fast, and because the model we assumed for the
observer is identical to the model of the actual plant (including the same initial conditions). Therefore, all of the
design requirements have been met with the minimal control effort expended.

You might also like