1. Trang chủ
  2. » Kỹ Thuật - Công Nghệ

Fuzzy Systems Part 3 pdf

20 200 0
Tài liệu đã được kiểm tra trùng lặp

Đang tải... (xem toàn văn)

Tài liệu hạn chế xem trước, để xem đầy đủ mời bạn chọn Tải xuống

THÔNG TIN TÀI LIỆU

Thông tin cơ bản

Định dạng
Số trang 20
Dung lượng 442,89 KB

Các công cụ chuyển đổi và chỉnh sửa cho tài liệu này

Nội dung

Similar to Theorem 3.3, a robust stabilization is obtained from Theorem 3.2 as follows: Theorem 3.4 The system 12 with w =k 0 is robustly stabilizable via the output feedback controller

Trang 1

Definition 3.1 (i) Consider the unforced system (12) with w( =k) 0,u( =k) 0 The uncertain system (12) is said to be robustly stable if there exists a matrix X>0 such that

0 Δ

ΔXA − X<

A T

for all admissible uncertainties

(ii) The uncertain system (12) is said to be robustly stabilizable via output feedback controller if there exists an output feedback controller of the form (16) such that the resulting closed-loop system (12) with (16) is robustly stable

Definition 3.2 (i) Given a scalar γ>0, the system (12) is said to be robustly stable with H∞

disturbance attenuation γ if there exists a matrix X>0 such that

0 0

0 0

0

Δ 11 Δ 1

1 Δ

1 Δ

Δ 11 Δ 1

<

I D

C

X B

A

D B I γ

C A X

T T T T

(ii) Given a scalar γ>0, the uncertain system (12) is said to be robustly stabilizable with H∞

disturbance attenuation γ via output feedback controller if there exists an output feedback

controller of the form (16) such that the resulting closed-loop system (12) with (16) is robustly stable with H∞ disturbance attenuation γ

The robust stability and the robust stability with H∞ disturbance attenuation are converted into the stability with H∞ disturbance attenuation

Definition 3.3 Given a scalar γ>0, the system

) 1 ( +k

x = Ax(k)+Bw(k), )

(k

z = Cx(k)+Dw(k) (17)

is said to be stable with H∞ disturbance attenuation γ if it is exponentially stable and input-output stable with (14)

Now, we state our key results that show the relationship between the robust stability and the robust stability with H∞ disturbance attenuation of an uncertain system, and stability with H∞ disturbance attenuation of a nominal system

Theorem 3.1 The system (12) with w( =k) 0 is robustly stable if and only if for ε>0 the system

) 1 ( +k

x = A r x(k)+ε−1H1w(k), )

(k

z = ε A~x(k)

where w and z are of appropriate dimensions, is stable with unitary H∞ disturbance attenuation γ=1

Proof: The system (12) is robustly stable if and only if there exists a matrix X>0 such that

, 0 )

~ ) ( (

)

~ ) ( (A r+H1F1 k A T X A r+H1F1 k AX<

which can be written as

Trang 2

0 ) ( )

( 1

+HF k E E T F T k H T

where

]

0

~ [ ,

0 ,

1

H

H X A

A X Q r

T

=

It follows from Lemma 2.1 that there exists X>0 such that (18) holds if and only if there

exist a matrix X>0 and a scalar ε>0 such that

, 0

1 2

+ HH ε E E

ε

which can be written as

0 0

0 1

1

= −

I E

ε

I H ε

E ε H ε Q

T

Pre-multiplying and post-multiplying

, 0 0 0

0 0 0

0 0 0

0 0 0 1

=

I I I

I S

we have

0 0

0

0

0 0

~ 0

1 1

1 1

1

= − − −

I A

ε

X H ε A

H ε I

A ε A X

S Y S

r

T T T r

The desired result follows from Definition 3.1

Theorem 3.2 The system (12) with u( =k) 0 is robustly stable with H∞ disturbance

attenuation γ if and only if for ε>0 the system

) 1 ( +k

x = A r x(k)+[γ−1B1r ε−1H1 0]w~(k), )

(

~ k z = ~( )

0 0

~ 0 0

~ 0 )

(

~

~

11

1 11

1

1

1

k w D

εγ

B εγ

H ε D

γ k x C ε

A ε C

r

r r

⎡ +

where w ~ and z~ are of appropriate dimensions, is stable with unitary H∞ disturbance

attenuation γ=1

Trang 3

Proof: The system (12) is robustly stable with H∞ disturbance attenuation γ if and only if

there exists a matrix X>0 such that

, 0 0

~ ) (

~ ) (

0

~ ) (

~ ) (

)

~ ) ( (

)

~ ) ( (

0

)

~ ) ( (

)

~ ) ( (

0

11 2 2 11 1 2

2

1

1 1

1 1 1 1

1

11 2 2 11 1

1 1 1

<

− +

+

− +

+

+ +

+ +

I D

k F H D C k

F

H

C

X B

k F H B A k

F

H

A

D k F H D B k F H B I

γ

C k F H C A k F H A X

r r

r r

T r

T r

T r

T r

which can be written as

0 ˆ ( ˆ ˆ ˆ ( ˆ

ˆ+H F k E+E T F T k H T<

where

, 0

0 0

0

ˆ

11

1

1 1

11 1

I D

C

X B

A

D B I

γ

C A X

Q

r

r

r

r

T r T r

T r T r

, 0 0 0 0 0 0 ˆ

2

1

=

H H

) ( 0 0 ) ( ) (

2

1

=

k F k F k

0 0

~

~ 0 0

~

~ ˆ

11 1

1

=

D C B A E

It can be shown from Lemma 2.1 that there exists X>0 such that (19) holds if and only if there exist X>0 and a scalar ε>0 such that

, 0 ˆ ˆ ˆ ˆ 1

+ H H ε E E ε

which can be written as

0 0

ˆ

ˆ ˆ ˆ

1 1

= −

I E

ε

I H ε

E ε H ε Q

T

Pre-multiplying and post-multiplying the above LMI by

,

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0 1

2

=

I I I

I

I I

I γ I

S

we have

Trang 4

0 0

~

~

~

~ 0

~ 0

1 2

2

I D

C

X B A

D B I

C A X

S Y S

r

T T

T T r

The result follows from Definition 3.1

3.3 Robust controller design

We are now at the position where we propose the control design of an H∞ output feedback

controller for the system (12) The controller design is based on the equivalent system (15)

The design of a robustly stabilizing output feedback controller with H∞ disturbance

attenuation for the system (15) can be converted into that of a stabilizing controller with H∞

disturbance attenuation controllers for a nominal system For the following auxiliary

systems, we can show that the following theorems hold Consider the following systems:

) 1 ( +k

x = A r x(k)+[γ−1B1r ε−1H1 0 0]w~(k)+B2r u(k),

) (

~

~

~ ) (

~ 0 0 0

~ 0 0 0

~ 0 0 0

~ 0 0 )

(

~

~

~

12 22 2 12

11

1 21

1 11

1

1 2

1

k u D ε

D ε

B ε

D k w D

εγ

D εγ

B εγ

H ε D

γ k x C ε

C ε

A ε

⎡ +

⎡ +

)

(k

y = C2r x(k)+[γ−1D21r 0 ε−1H3 0]w~(k)+D22r u(k)

(20)

and

) 1 ( +k

x = A r x(k)+[ε−1H1 0]w(k)+B2r u(k), )

(k

z = ~~ ( ) ~~ ( ),

22

2 2

k u D ε

B ε k x C ε

A ε

⎡ +

⎡ )

(k

y = C2r x(k)+[0 ε−1H3]w(k)+D22r u(k),

(21)

where 0ε> is a scaling parameter

Theorem 3.3 The system (12) is robustly stabilizable with H∞ disturbance attenuation with

γ via the output feedback controller (16) if the closed-loop system corresponding to (20)

and (16) is stable with unitary H∞ disturbance attenuation

Proof: The closed-loop system (12) with (16) is given by

) 1 ( +k

x c = (A c+H1c F1c(k)E1c)x c(k)+(B c+H1c F1c(k)E2c)w(k),

)

(k

z = (C c+H2F2(k)E3c)x(k)+(D11r+H2F2(k)D~11)w(k)

where [ T ˆT T]

c

x = x x and

, ˆ 0

0 ],

ˆ [

, ˆ ,

ˆ ˆ ˆ ˆ

ˆ

3

1 1 12

1 21

1 22

2

2

=

=

=

+

=

H B

H H C D C C D B

B B C D B A C

B

C B A

r

r c r r

r r

Trang 5

ˆ

~

~ [ ,

~

~ ,

ˆ

~

~ ˆ

~

~ ,

) ( 0 0 ) ( )

21

1 2 22 2

2 1

3

1

D

B E C D C C B A E k F k F

k

=

=

On the other hand, the closed-loop system (20) with (17) is given by

) 1 ( +k

x c = A c x c(k)+[γ−1B c ε−1H1c 0]w~(k), )

(

~ k z = ~( )

0 0

~ 0 0

0 )

(

11

1 11

1

3

D εγ

E εγ

H ε D

γ k x E ε

E ε C

r c

r c

c c c

⎡ +

The result follows from Theorem 3.2

Similar to Theorem 3.3, a robust stabilization is obtained from Theorem 3.2 as follows:

Theorem 3.4 The system (12) with w( =k) 0 is robustly stabilizable via the output feedback controller (16) if the closed-loop system corresponding to (21) and (16) is stable with unitary

H∞ disturbance attenuation

Remark 3.1 Theorem 3.3 indicates that a controller that achieves a unitary H∞ disturbance attenuation for the nominal system (20) can robustly stabilize the fuzzy system (12) with H∞

disturbance attenuation γ Similar argument can be made on robust stabilization of

Theorem 3.4 Therefore, the existing results on stability with H∞ disturbance attenuation can

be applied to solve our main problems

3.4 Numerical examples

Now, we illustrate a control design of a simple discrete-time Takagi-Sugeno fuzzy system with immeasurable premise variables We consider the following nonlinear system with uncertain parameters

) 1 (

1 k+

x = (0.9 ) ( ) 0.2 ( ) 0.2 ( ) 2( ) 0.3 1( ),

2 1 2

x

+ )

1 (

2 k+

x = 0.2x1(k)−(0.4+β)x3(k)+0.5w1(k)+0.7u(k),

)

(k

) (

) ( 5 0 ) ( 5

1 1 2

⎡ +

k u

k x k x

)

(k

y = 0.3x1(k)−0.1x2(k)+w2(k) where α and β are uncertain scalars which satisfy α ≤0.1 and β ≤0.02,respectively Defining x(k)=[x1(k) x2(k)], w(k)=[w1(k) w2(k ] and assuming x2(k)∈[−1, 1], we have an equivalent fuzzy system description

)

1

(

1 k+

7 0

0 ) ( 0 5 0

0 3 0 ) ( ) ) ( (

)) ( ( 2

⎪⎭

⎪⎩

⎡ +

⎡ + +

=

i λ i x k A i H i F i k E i x k w k u k

)

(k

1

0 ) ( 0 0

5 0 5 1

k u k

⎡ +

⎡ )

(k

y = [0.3 −0.1]x(k)+[0 1]w(k)

where 2

1( ( )) 1x k2 x k2( ),

2( ( ))x k2 x k2( )

λ = and

Trang 6

   

2 0

0 ,

0

5 0 ,

3 0 2 0

2 0 1 1 ,

0 2 0

2 0 9 0

12 11

2

=

=

=

⎡ −

A

   

  ( ) , [0.2 0], [0 0.1] ,

)

β

β k F α

α k

which can be written as

) 1 (

1 k+

7 0

0 ) ( 0 5 0 0 3 0 ) ( )

~ ) (

~ (A2 H1F k A x k w ku k

⎡ +

⎡ + +

)

(k

1

0 ) ( 0 0

5 0 5 1

k u k

⎡ +

⎡ )

(k

y = [0.3 −0.1]x(k)+[0 1]w(k) where

=

= , ~( )

2 0 0 1

0

0 5 0 0

1

1 0 0

0 2 0 3 0 0 0 2 0

~ ], ) ( ) ( ) ( ) ( [ 1 1 1 2

⎡−

=

A k F k F k λ k

The open-loop system is originally unstable Theorem 3.3 allows us to design a robust

stabilizing controller with H disturbance attenuation γ=20:

) 1 (

ˆ +k

9861 7

8463 3 ) ( 1348 0 3181 2

0063 0 0971 0

k y k

⎡ +

− )

(k

u = [3.0029 −0.1726]x(k)

Fig 3 The state trajectories

Trang 7

Fig 4 The control trajectory

This controller is applied to the system A simulation result with the initial conditions

, ] 3

0

4

0

[

)

0

x = − x(0)=[0 0]T, the noises w1(k)=ekcos(k), w2(k)=eksin(k) and the assumption F1(k)=F2(k)=sin(k) is depicted in Figures 3 and 4, which show the trajectories of the state and control, respectively We easily see that the obtained controller stabilizes the system

4 Extension to fuzzy time-delay systems

In this section, we consider an extension to robust control problems for Takagi-Sugeno fuzzy time-delay systems Consider the Takagi-Sugeno fuzzy model, described by the following IF-THEN rule:

IF ξ is 1 M and … and i1 ξ is p M ip,

THEN x(t)=(A iA i)x(t)+(A diA di)x(th)+(B1iB1i)w(t)+(B2iB2i)u(t),

) ( ) Δ ( ) ( ) Δ ( ) ( ) Δ ( )

(t C1 C1 x t C1 C1 x t h D11 D11 w t

+(D12iD12i)u(t),

) ( ) Δ ( ) ( ) Δ ( ) ( ) Δ ( ) C2 C2 x t C2 C2 x t h D21 D21 w t

t

+(D22iD22i)u(t),                 i=1,",r

where x(t)∈ℜn is the state, w(t)∈ℜm1 is the disturbance, u(t)∈ℜm2 is the control input,

1

)

(t q

z ∈ℜ is the controlled output, y(t)∈ℜq2 is the measurement output r is the number

of IF-THEN rules M is a fuzzy set and ij ξ1, ", ξ p are premise variables We set

]

[ξ1 ξ p T

ξ= " We assume that the premise variables do not to depend on u (t) ,

i

A A di, B1i, B2i,C1i,C2i,C1di,C2di, D11i, D 12i, D21i and D22i are constant matrices of appropriate dimensions The uncertain matrices are of the form (1) with ΔA = di H1i F i(t)E di,

di i

i

di H F t E

Δ 1 = 2 and ΔC2di =H3i F i(t)E di where H1i, H2i, H3i and E are known di

constant matrices of appropriate dimensions

Trang 8

Assumption 4.1 The system (A r, A dr, B1r, B2r, C1r, C2r, C1dr, C2dr, D11r, D12r,

,

21r

D D22r) represents a nominal system that can be chosen as a subsystem including the

equilibrium point of the original system

The state equation, the controlled output and the output equation are defined as follows:

)

(

x = r1λ(x(t)){(A i ΔA i)x(t) (A di ΔA di)x(t h) (B1i ΔB1i)w(t)

=

)}, ( ) Δ (B2i+ B2i u t

+ )

(

z = 1λ(x(t)){(C1i ΔC1i)x(t) (C1di ΔC1di)x(t h) (D11i ΔD11i)w(t)

r

=

)}, ( ) Δ (D12i+ D12i u t

+ )

(

y = 1λ(x(t)){(C2i ΔC2i)x(t) (C2di ΔC2di)x(t h) (D21i ΔD21i)w(t)

r

=

)}

( ) Δ (D22i+ D22i u t

+

(22)

where λ i ( t x()) is defined in (3) and satisfies (4) Our problem is to find a control u for the (⋅)

system (22) given the output measurements y such that the controlled output )(⋅) z (⋅

satisfies (5) for a prescribed scalar γ>0 Using the same technique as in the previous

sections, we have an equivalent description for (22):

)

(t

x = (A rA)x(t)+(A drA d)x(th)+(B1rB1)w(t)+(B2rB2)u(t)

), ( ) ( ) ( )

( Δ 1Δ 2Δ

)

(t

z = (C1rC1)x(t)+(C1drC1d)x(th)+(D11rD11)w(t)+(D12rD12)u(t)

), ( )

( )

( )

( 1 Δ 11Δ 12Δ

Δ

)

(t

y = (C2rC2)x(t)+(C2drC2d)x(th)+(D21rD21)w(t)+(D22rD22)u(t)

) ( )

( )

( )

( 2 Δ 21Δ 22Δ

Δ

(23)

where ΔA = d H1F~(t)A~d, ΔC1d=H1F~(t)C~1d, ΔC2d=H1F~(t)C~2d,and other uncertain matrices

are given in (7) As we can see from (23) that uncertain Takagi-Sugeno fuzzy time-delay

system (22) can be written as an uncertain linear time-delay system Thus, robust control

problems for uncertain fuzzy time-delay system (22) can be converted into those for an

uncertain linear time-delay system (23) Solutions to various control problems for an

uncertain linear time-delay system have been given(for example, see Gu et al., 2003;

Mahmoud, 2000) and hence the existing results can be applied to solve robust control

problems for fuzzy time-delay systems

5 Conclusion

This chapter has considered robust H∞ control problems for uncertain Takagi-Sugeno fuzzy

systems with immeasurable premise variables A continuous-time Takagi-Sugeno fuzzy

system was first considered Takagi-Sugeno fuzzy system with immeasurable premise

variables can be written as an uncertain linear system Based on such an uncertain system

representation, robust stabilization and robust H∞ output feedback controller design method

was proposed The same control problems for discrete-time counterpart were also

Trang 9

considered For both continuous-time and discrete-time control problems, numerical examples were shown to illustrate our design methods Finally, an extension to fuzzy time-delay systems was given and a way to robust control problems for them was shown Uncertain system approach taken in this chapter is applicable to filtering problems where the state variable is assumed to be immeasurable

6 References

Assawinchaichote, W.; Nguang, S.K & Shi, P (2006) Fuzzy Control and Filter Design for

Uncertain Fuzzy Systems, Springer

Boyd, S.; El Ghaoui, L.; Feron, E & Balakrishnan, V (1994) Linear Matrix Inequalities in

Systems and Control Theory, SIAM

Cao, G.; Rees, W & Feng, G (1996) H∞ control of nonlinear continuous-time systems based

on dynamical fuzzy models, International Journal of Systems Science 27, pp.821-830

Chen, C.L.; Feng, G.; Sun, D & Guan, X.P (2005) H∞ output feedback control of

discrete-time fuzzy systems with application to chaos control, IEEE Transactions on Fuzzy Systems 13, pp.531-543

Feng, G.; Cao, G & Rees, W (1996) An approach to H∞ control of a class of nonlinear

systems, Automatica 32, pp.1469-1474

Gu, K; Kharitonov, V and Chen, J (2003) Stability of Time-Delay Systems, Boston: Birkhauser

Guerra, T.M.; Kruszewski, A.; Vermeiren, L & Tiemant, H (2006) Conditions of output

stabilization for nonlinear models in the Takagi-Sugeno’s form, Fuzzy Sets and Systems 157, pp.1248-1259

Hong, S.-K & Langari, R (1998) Synthesis of an LMI-based fuzzy control system with

guaranteed optimal H∞ performance, Proc of IEEE International Conference on Fuzzy Systems, pp.422-427

Katayama, H & Ichikawa, A (2002) H∞ control for discrete-time Takagi-Sugeno fuzzy

systems, International Journal of Systems Science 33, pp.1099-1107

Ma, X.J.; Sun, Z.Q & He, Y.Y (1998) Analysis and design of fuzzy controller and fuzzy

observer, IEEE Transactions on Fuzzy Systems 6, pp.41-50

Mahmoud, M.S (2000) Robust Control and Filtering for Time-Delay Systems, New York-Basel:

Mercel-Dekker, Inc

Takagi, T & Sugeno, M (1985) Fuzzy identification of systems and its applications to

modeling and control, IEEE Transactions on System, Man, Cybernetics 15, pp.116-132

Tanaka, K & Sugeno, M (1992) Stability analysis and design of fuzzy control systems,

Fuzzy Sets and Systems 45, pp.135-156

Tanaka, K & Sano, M (1994) On the concepts of regulator and observer of fuzzy control

systems, Proc of IEEE International Conference on Fuzzy Systems, pp.767-772

Tanaka, K.; Ikeda, T & Wang, H.O (1996) Robust stabilization of a class of uncertain

nonlinear systems via fuzzy control: quadratic stabilizability, H∞ control theory,

and linear ,atrix inequalities, IEEE Transactions on Fuzzy Systems 4, pp.1-13

Tanaka, K.; Ikeda, T & Wang, H.O (1998) Fuzzy regulators and fuzzy observers: relaxed

stability conditions and LMI-based designs, IEEE Transactions on Fuzzy Systems 6,

pp.1-16

Tanaka, K & Wang, H.O (1997) Fuzzy regulators and fuzzy observers: a linear matrix

inequality approach, Proc of IEEE Conference on Decision and Control, pp.1315-1320

Trang 10

Tanaka, K & Wang, H.O (2001) Fuzzy Control Systems Design and Analysis: A Linear Matrix

Inequality Approach, Wiley-Interscience

Xie, L (1996) Output feedback H∞ control of systems with parameter uncertainty,

International Journal of Control 63, pp.741-759

Yoneyama, J.; Nishikawa, M.; Katayama, H & Ichikawa, A (2000) Output stabilization of

Takagi-Sugeno fuzzy systems, Fuzzy Sets and Systems 111, pp.253-266

Yoneyama, J.; Nishikawa, M.; Katayama, H & Ichikawa, A (2001a) Design of output

feedback controllers for Takagi-Sugeno fuzzy systems, Fuzzy Sets and Systems 121,

pp.127-148

Yoneyama, J.; Nishikawa, M.; Katayama, H & Ichikawa, A (2001b) H∞-Control for

Takagi-Sugeno fuzzy systems, International Journal of Systems 32, pp.915-924

Yoneyama, J (2008a) Robust H∞ output feedback control for uncertain fuzzy systems with

immeasurable premise variables, Advances in Fuzzy Sets and Systems 3, pp.99-113

Yoneyama, J (2008b) H∞ output feedback control for fuzzy systems with immeasurable

premise variables: discrete-time case, Applied Soft Computing 8, pp.949-958

Ngày đăng: 21/06/2014, 10:20