Scribd
Upload
44 views4 pages

Lecture 6 (21-24)

This document discusses the design of a state observer for a system using three different methods. It begins by checking the observability of the system and substituting the observer gain matrix ke into the characteristic polynomial to determine the values of ke1, ke2, and ke3. It then presents two additional methods: 1) using Ackermann's formula to directly calculate ke and 2) equating coefficients of the characteristic equations. The effects of the observer on the closed-loop system are analyzed, showing that the pole placements and observer designs are independent. Finally, the transfer function of the observer-based controller is derived.

Uploaded by

Saif Alabdullah
Copyright
© © All Rights Reserved
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
44 views4 pages

Lecture 6 (21-24)

This document discusses the design of a state observer for a system using three different methods. It begins by checking the observability of the system and substituting the observer gain matrix ke into the characteristic polynomial to determine the values of ke1, ke2, and ke3. It then presents two additional methods: 1) using Ackermann's formula to directly calculate ke and 2) equating coefficients of the characteristic equations. The effects of the observer on the closed-loop system are analyzed, showing that the pole placements and observer designs are independent. Finally, the transfer function of the observer-based controller is derived.

Uploaded by

Saif Alabdullah
Copyright
© © All Rights Reserved
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/ 4

𝑘𝑒1

𝒌𝑒 = [𝑘𝑒2 ]
𝑘𝑒3
Check the Observability condition for the system. If the system is completely state
observable, then substitute the 𝒌𝑒 matrix into the desired characteristic polynomial:
|𝑠𝑰 − (𝑨 − 𝒌𝒆 𝒄)| = (𝑠 − 𝜇1 )(𝑠 − 𝜇2 )(𝑠 − 𝜇3 )
By equating the coefficients of the like powers of s on both sides of this last equation, we can
determine the values of 𝑘𝑒1 , 𝑘𝑒2 , and 𝑘𝑒3 . This approach is convenient if n = 1, 2, or 3, where
n is the dimension of the state vector x.
 Ackermann’s formula to obtain state observer gain matrix 𝒌𝒆 :
Check the Observability condition for the system. If the system is completely state
observable, then use the following equation to obtain state observer gain matrix 𝒌𝒆 :
𝒄 −𝟏
𝒄𝑨 0
𝒄𝑨 𝟐 0
𝒌𝒆 = 𝚫𝒄 (𝑨) : (1.22)
: 0
𝒄𝑨𝒏−𝟐 [1]
[ 𝒄𝑨𝒏 ]

Example 1.13: Consider the system


0 20.6 0
𝐴=[ ]; 𝑏 = [ ]; 𝑐 = [0 1]
1 0 1
We use the observed state feedback such that 𝑢 = −𝑘𝑥̃. Design a full-order state observer,
assuming that the desired eigenvalues of the observer matrix are:
𝑠1 = 𝜇1 = −10; 𝑠2 = 𝜇2 = −10
Solution:
The design of the state observer reduces to the determination of an appropriate observer gain
matrix 𝒌𝒆 .
Let us test the Observability of the given system
𝑐 0 1 0 1
𝑄=[ ]=[ ] → det(𝑄) = | | = −1
𝑐𝐴 1 0 1 0
The system is completely observable.
Method 1: We shall determine the observer gain matrix by use of Equation (1.21)
Since the characteristic equation of the given system is:

|𝑠𝑰 − 𝑨| = | 𝑠 −20.6| = 𝑠 2 − 20.6 = 𝑠 2 + 𝑎1 𝑠 + 𝑎2


−1 𝑠
∴ 𝑎1 = 0; 𝑎𝑛𝑑 𝑎2 = −20.6
The desired char. Eq. is:

21
(𝑠 + 10)2 = 𝑠 2 + 20𝑠 + 100 = 𝑠 2 + 𝛼1 𝑠 + 𝛼2
Hence:
𝛼1 = 20; 𝛼2 = 100
0 1 0 1
𝑊=[ ] and 𝑉 = [ ]
1 0 1 0
1 0
𝑁 = (𝑊𝑉 𝑇 )−1 = [ ]
0 1
𝛼2 − 𝑎 2 1 0 100 + 20.6 120.6
𝑘𝑒 = (𝑊𝑉 𝑇 )−1 [𝛼 − 𝑎 ] = [ ][ ]=[ ]
1 1 0 1 20 − 0 20
Method 2:
|𝑠𝐼 − 𝐴 + 𝑘𝑒 𝑐| = 0
𝑠 0 0 20.6 𝑘 𝑠 −20.6 + 𝑘𝑒1
|[ ]−[ ] + [ 𝑒1 ] [0 1]| = | | = 𝑠 2 + 𝑘𝑒2 𝑠 − 20.6 + 𝑘𝑒1 = 0
0 𝑠 1 0 𝑘𝑒2 −1 𝑠 + 𝑘𝑒2
Since the desired char. Eq. is:
𝑠 2 + 20𝑠 + 100 = 0
∴ 𝑘𝑒1 = 120.6; 𝑘𝑒2 = 20
Method 3: We shall use Ackermann's formula given by Equation (1.22):
𝒄 −𝟏 0
𝒌𝒆 = 𝚫𝒄 (𝑨) [ ] [ ]
𝒄𝑨 1
where:
∆𝑐 (𝑠) = (𝑠 − 𝜇1 )(𝑠 − 𝜇2 ) = 𝑠 2 + 20𝑠 + 100
Thus,
120.6 412
∆𝑐 (𝐴) = 𝐴2 + 20𝐴 + 100𝐼 = [ ]
20 120.6
120.6 412 0 1 0 120.6
𝑘𝑒 = [ ][ ][ ] = [ ]
20 120.6 1 0 1 20

1.4.2.1 Effects of the Addition of the Observer on a closed-loop System:


In the pole-placement design process, we assumed that the actual state x(t) was available for
feedback. In practice, however, the actual state 𝑥(𝑡)may not be measurable, so we will need
to design an observer and use the observed state 𝑥̃(𝑡) for feedback as shown in Figure 5. The
design process, therefore, becomes a two-stage process, the first stage being the determination
of the feedback gain matrix 𝑘 to yield the desired characteristic equation and the second stage
being the determination of the observer gain matrix 𝑘𝑒 to yield the desired observer
characteristic equation.

22
Figure 5 Observer state feedback control

Consider the completely state controllable and completely observable system defined by the
equation:
𝑥̇ = 𝐴𝑥 + 𝑏𝑢
𝑦 = 𝑐𝑥
For the state feedback control based on the observed state 𝑥̃(𝑡)
𝑢 = −𝑘𝑥̃
With this control, the state equation becomes:
̃ = (𝑨 − 𝒃𝒌)𝒙 + 𝒃𝒌(𝒙 − 𝒙
𝒙̇ = 𝑨𝒙 − 𝒃𝒌𝒙 ̃) (1.23)
The difference between the actual state 𝑥(𝑡) and the observed state 𝑥̃(𝑡)has been defined as
the error 𝑒(𝑡):
̃(𝒕)
𝒆(𝒕) = 𝒙(𝒕) − 𝒙
So that, the equation 1.23 can be rewritten as:
𝒙̇ = (𝑨 − 𝒃𝒌)𝒙 + 𝒃𝒌𝒆 (1.24)

Not that the observer error equation was given by Equation 1.20:
𝒆̇ = (𝑨 − 𝒌𝒆 𝒄)𝒆
So, we obtain that:

23
𝑥̇ 𝐴 − 𝑏𝑘 𝑏𝑘 𝑥
[ ]=[ ][ ] (1.25)
𝑒̇ 0 𝐴 − 𝑘𝑒 𝑐 𝑒
Eq. 1.25 describes the dynamics of the observed-state feedback control system.
The characteristic equation for the system is:
𝑠𝐼 − 𝐴 + 𝑏𝑘 −𝑏𝑘
| |=0
0 𝑠𝐼 − 𝐴 + 𝑘𝑒 𝑐
or
|𝑠𝐼 − 𝐴 + 𝑏𝑘||𝑠𝐼 − 𝐴 + 𝑘𝑒 𝑐| = 0 (1.26)
Notice that the closed-loop poles of the observed-state feedback control system consist of the
poles due to the pole-placement design alone and the poles due to the observer design alone.
This means that the pole-placement design and the observer design are independent of each
other. They can be designed separately and combined to form the observed-state feedback
control system. Note that, if the order of the plant is n, then the observer is also of nth order (if
the full-order state observer is used), and the resulting characteristic equation for the entire
closed-loop system becomes of order 2n.

1.4.2.2 Transfer Function of the observer based controller:


Assume that the plant is completely observable. The equations for the observer are given by:
𝑑𝑥̃
= (𝐴 − 𝑘𝑒 𝑐 − 𝑏𝑘)𝑥̃ + 𝑘𝑒 𝑦 (1.27)
𝑑𝑡
𝑢 = −𝑘𝑥̃ (1.28)
By taking the Laplace transform of Equation (1.27), assuming a zero initial condition, and
solving for 𝑋̃(𝑠):
𝑋̃(𝑠) = (𝑠𝐼 − 𝐴 + 𝑘𝑒 𝑐 + 𝑏𝑘)−1 𝑘𝑒 𝑌(𝑠)
By substituting this 𝑋̃(𝑠) into the Laplace transform of Eq. 1.28, we obtain:
𝑈(𝑠) = −𝑘(𝑠𝐼 − 𝐴 + 𝑘𝑒 𝑐 + 𝑏𝑘)−1 𝑘𝑒 𝑌(𝑠)
Then the transfer function is given by:
𝑈(𝑠)
= −𝑘(𝑠𝐼 − 𝐴 + 𝑘𝑒 𝑐 + 𝑏𝑘)−1 𝑘𝑒 (1.29)
𝑌(𝑠)
Figure 6 show the block diagram represented for the system. The transfer function (𝑘(𝑠𝐼 −
𝐴 + 𝑘𝑒 𝑐 + 𝑏𝑘)−1 ) acts as controller for the system.
The transfer function in eq. 1.29 is called the observer-controller transfer function.

24
Figure 6 the system with controller observer

You might also like