XTP23_151_Paperstampedno.pdf
-
Upload
usef-usefi -
Category
Documents
-
view
216 -
download
0
Transcript of XTP23_151_Paperstampedno.pdf
-
7/28/2019 XTP23_151_Paperstampedno.pdf
1/6
8th International IFAC Symposium onDynamics and Control of Process Systems
STABLE INDIRECT ADAPTIVE FUZZY CONTROL BASED ON TAKAGI-SUGENO MODEL
Ruiyun Qia,*, Mietek A. Brdys
a
aDepartment of Electronic Electrical & Computer Engineering, University of Birmingham, Edgbaston,Birmingham B15 2TT, UK
Abstract: This paper presents an indirect adaptive fuzzy control scheme for nonlinear
uncertain stable plants with unmeasurable states. A discrete-time T-S fuzzy model isemployed as a dynamic model of an unknown plant. Based on this model, a feedbacklinearization controller is designed and applied to both the model and the plant.
Parameters of the model are updated on-line to allow for partially unknown and time-varying plants. Stability analysis shows that the adaptive controller guarantees theboundedness of all the closed-loop signals and achieves bounded tracking error. In the
ideal case where there is no modelling error and the signal for parameter learning ispersistently exciting, perfect tracking is ensured. The effectiveness of the method isverified by simulation examples. Copyright 2007 IFAC
Keywords: nonlinear control, adaptive control, fuzzy control,.
1. INTRODUCTION
Adaptive control has become very popular in manyfields of control engineering and attracted much
attention in developing advanced applications.Adaptive control is based on feedback of signals in acontrolled system for control adaptation to
effectively handle system uncertainties (Tao, 2003).
The adaptive control can be divided into two classes:direct adaptive control and indirect adaptive control.The former tunes the parameters of the controllerwhile the latter updates the parameters of the model.
This paper focuses on indirect adaptive control,which means the T-S fuzzy model is adapted on-line.Design of provably stable indirect adaptive fuzzycontroller posed more difficult theoretical questions.
Some of the early adaptive control schemes usingcontinuous-time T-S fuzzy model (Takagi &Sugeno,
1985) were presented in the papers (Wang, 1994;*
Corresponding author. Tel: 44-1214143150; fax: 44-1214144291.
E-mail address: [email protected] (Ruiyun Qi),[email protected](M.A. Brdys)
Spooner 1996; Yoo, 1998) with stability results.Wangs control law is composed of a certaintyequivalence control and a supervisory controllerwhich is used to force the tracking error not to
exceed certain error ball. Spooner and Passinosapproach consists of a certainty equivalence controlterm, a bounding control term used to restrict the
plant output trajectory, and a sliding mode control
term to compensate the approximation errors. Theirparameter adaptation algorithms are chosen to makethe derivative of a certain Lyapunov equationnonpositive. Recent adaptive fuzzy control schemes
for continuous-time system can be found in the
papers (Golea2003; Park & Cho, 2004). In Park
and Cho (2004), an adaptive law for updatingparameters is designed based the Lyapunov theory
and with the on-line parameter estimator, any type offuzzy controllers works adaptively to the parameterperturbations.All the above adaptive control schemes assume thatfull states of the controlled plant are available for
measurement. Adaptive output feedback control fornonlinear systems when full states are notmeasurable remains one of the outstanding problems.Kulawski and Brdys (2000) present an adaptive
Preprints Vol.2, June 6-8, 2007, Cancn, Mexico
385
mailto:[email protected]:[email protected]:[email protected]:[email protected] -
7/28/2019 XTP23_151_Paperstampedno.pdf
2/6
control scheme for uncertain nonlinear plants withunmeasurable states based on dynamic neuralnetworks. In their approach, a recurrent neural
network is employed as a dynamical model of theplant and the control law is designed by using thestates of the neural network instead of the plant. Our
proposal in this paper is inspired by the aboveconsideration and applies it to discrete-time adaptivefuzzy control system. A discrete-time T-S fuzzy
model is trained to approximate the input-outputbehaviour of an unknown continuous-time plant byusing I/O data from the plant. Then an appropriatecontrol law is calculated completely based on themodel, using the model structure, parameters andstates. The control input is applied to both the plant
and the model. The parameters of the model areupdated on-line by recursive least square estimation(RLSE) method, which is fast and is able to ensure
the boundedness of parameters (Tao, 2003). Withthis approach, at least the model can be stabilizedand provides bounded control input without extrasupervisory controller. The sufficient conditions for
the convergence of tracking error are derived forstable plant.The paper is organized as follows. The controlalgorithm including parameter adaptation law ispresented in Section 2. Stability of the closed-loop
system and convergence of the tracking error areproven in Section 3. Section 4 shows the simulation
results on control of a single-link robot manipulatorwith parameter disturbance. Finally, Section 5concludes the paper.
2. CONTROL ALGORITHM
2.1 Modelling for adaptive controlIt has been shown that T-S fuzzy systems with
piecewise linear rule consequents are universalapproximators to approximate any continuousnonlinear system to an arbitrary accuracy (Ying,
1998 & 1999) if all fuzzy membership functions areGaussian and the t-norm is the multiplication. In thispaper, the T-S fuzzy model is used to express a real
nonlinear plant. That is, the nonlinear plant can berepresented by the following fuzzy rules:
1 1
0 1 1 1
: If ( ) is and and ( ) is , then
( 1) ( ) ( ) ( )
i i i
n n
i i i i i
n n n n
R x k M x k M
x k x k x k u kq q q q ++ = + + + +
L
L
(1)
where iR ( 1, 2, ,i N= L ) denotes the ith fuzzy rule
andNis the number of rules.1( ), , ( )
nx k x kL are the
premise variables which are the past plant output
measurements and ( )u k is the control input. iM
( 1, ,i n= L ) are fuzzy sets which are described byGaussian functions
2
2
( )( | , ) exp
2( )
i
j ji i i
j j j j i
j
x cM x c s
s
- = -
(2)
The model parameter values ij
q ,i
jc andis ,
1, ,i N= L , 1, ,j n= L are unknown. The T-S fuzzy
model approximates a nonlinear system bycombining a group of local affine models. The
premise of a fuzzy rule defines a local operatingregion and each consequent describes a local input-
output relation in the local operating region. Let
1 2( ) [ ( ), ( ), , ( )] [ ( 1), ,T
nk x k x k x k y k n= = - +L Lx
( 1), ( )]Ty k y k- , then the overall T-S fuzzy model is
calculated as
1
( 1) ( | , ) 1 ( ) ( )
( ( ) | , )
( ) ( )
NT
i i T
n
i
T
n
x k k u k
k
y k x k
w=
+ =
= Y Q
=
i ix c x
x c (3)
where
{ }{ }
( ) ( )
1
1
1
0 1
1 2
( | , ) ( | , ),
( | , ) ( | , ) ( | , ) ,
| 1, , ; 1, ,
| 1, , ; 1, ,
, ,
( ) 1 ( ) ( ) ,
( | , ) .
ni i i i
j j j j
jN
i i i
ii
j
i
j
TT TT
i i i N
n
TTe
TT T T N T
e e e
M x c
c i N j n
i N j n
k k u k
w s
w w w
s
q q
w w w
=
=
+
=
=
= = =
= = =
= Q =
=
Y =
L L
L L
L L
L
i i
i i
x c
x c x c x c
c
x x
x c x x x
(4)
An assumption is made here.
Assumption 1. The modelling error is negligible and
there exist parameter values ( , )c and Q that make
the model (3) to become a perfect representation ofthe real plant.From now on, the nonlinear plant is completelyrepresented by (3). Since all the parameters in (3) are
unknown, we define an estimation model as
1
( 1) ( | , ) 1 ( ) ( )
( ( ) | ( ), ( )) ( )
( ) ( )
NT
i i T
ni
T
n
x k k u k
k k k k
y k x k
w= + =
= Y Q
=
i i
x c
x
x c (5)
where ( ( ), ( ))k kc and ( )kQ are the estimates of
( , )c and Q at time k. ( ) nk x is the model state
vector.
2.2Control synthesis
( 1)ry k+( 1)y k+
( 1)y k+
( )u k
( 1)e k+
( )kx
( )k
M
1-
( )ke
( )v k
Tk
1-
1-
Fig. 1 Adaptive control based on T-S fuzzy model
The control objective is to make the output of the
nonlinear plant (3) tracking a specified referencetrajectoryr
y .
Assumption2. The reference ( )r
y k satisfies
( )r k Uy (6)
386
-
7/28/2019 XTP23_151_Paperstampedno.pdf
3/6
where [ ]( ) ( 1) ( )T
r r rk y k n y k = - +y L and U is
a known bound.The overall control structure is shown in Fig. 1. Itconsists of three parts: a nonlinear uncertain plant, an
adaptive T-S fuzzy model and a model basedfeedback linearization controller.
The control law is designed based on the model (5)whose evolution of output can be expressed as
1
1
1
( 1) ( ( ) | , ) ( ) 1 ( ) ( )
( ( ) | , ) ( ) 1 ( )
( ( ) | , ) ( ) ( )
( ( ) | ( )) ( ( ) | ( )) ( )
NT
i i T
i
NT
i i T
f
i
Ni i
g
i
y k k k k u k
k k k
k k u k
f k k g k k u k
w
w
w
=
=
=
+ =
=
+
= +
i i
i i
i i
x c x
x c x
x c
x p x p
(7)where
{ }
0 1
1
1
1
( ) ( ) ( ) ( ) ,
( ) ( ), ( ) ( ), ( ), ( )
( ( ) | ( )) ( ( )) ( ) 1 ( ) ,
( ( ) | ( )) ( ( )) ( ).
Ti i i i
f n
i i
g n
NT
i i T
f
i
Ni i
g
i
k k k k
k k k k k k
f k k k k k
g k k k k
q q q
q
w
w
+
=
=
=
= = Q
=
=
L
p c
x p x x
x p x
(8)
To apply feedback linearization technique to design
the control law ( ( ) | ( ))g k kx p is required to be
invertible.
Assumption 3. ( ( ) | ( ))g k k e>x p where e is a
small real positive number, which implies therelative degree of the T-S fuzzy model is equal toone.With the assumption 3, the control law can becalculated as
1 ( ) [ ( | ( )) ( )] ( | ( ))
u k f k v k g k
= - +x px p
(9)
where ( )v k is a new input chosen as
( ) ( 1) ( | ( ))Tr
v k y k k k = + + k e p (10)
where [ ( | ( )) ( 1| ( 1))k k e k n k n= - + - + Le p p
]( | ( ))T
e k kp and ( | ( ))e k kp is the model tracking
error ( | ( )) ( ) ( | ( ))
re k k y k y k k = -p p (11)
and1
( , , )Tn
k k= L k be such that the zeros of the
polynomial 11
( ) n nn
h z z k z k -= + + +L lie within the
unit circle centered at the origin of the plane.The control law (9) applied to (7) results in it beingdecoupled and linear with respect to the new input
( )v k
( 1) ( )y k v k+ = (12)
2.3Online parameter adaptation
The initial T-S fuzzy model can be obtained by eitheroff-line or on-line identification methods. In thispaper, the initial model is obtained by the on-line
identification method proposed in ((Qi and Brdys,
2005). The initial model is good enough for themodel based controller to produce proper controlsignal in the beginning of closed-loop control.
However, the plant may have external disturbancesand parameters disturbances so that the initial modelneeds to be updated. Therefore, the model should be
adaptive when plant changes. The parameters of themodel can be adjusted on-line to provide an adaptivemodel.
For fixed parameters
( ( ), ( ))k kc
, Q can beestimated by recursive least square estimation(RLSE).
( ) ( ) ( 1) ( 1) ( )
1 ( ) ( ) ( )TS k k k
k kk S k k
eY +Q + = Q -
+Y Y(13)
( ) ( ) ( ) ( )( 1) ( )
1 ( ) ( ) ( )
T
T
S k k k S k S k S k
k S k k
Y Y+ = -
+Y Y(14)
where ( )kY denotes ( | ( ), ( ))k kY x c and
( 1) ( 1) ( 1) ( ) ( ) ( 1)Tk y k y k k k y k e + = + - + = Y Q - + .
The initial conditions are chosen as
1 2 (0) (0) (0) (0) , (0)
TT T T
N S Iq q q Q = = W L
where1 2
(0) (0) (0)T T TN
q q qL are the parameters
of the initial T-S model and W is a large positiveconstant. Large W can speed up parameterconvergence, but too large may cause instability. It
needs to be chosen properly during on-lineadaptation.
3. STABILITY ANALYSIS
3.1Convergence of model tracking error
Applying (9) to (7) , we get
( 1) ( 1) ( | ( ))Tr
y k y k k k+ = + + k e p (15)
Since ( 1| ( )) ( 1) ( 1| ( 1))r
e k k y k y k k + = + - + +p p ,
we have the following error dynamics
2 1
( 1| ( 1)) ( 1 | ( 1))
( 1| ( 1)) ( | ( ))n
e k k k e k n k n
k e k k k e k k
+ + = - - + - +- - - - -Lp p
p p(16)
which can be formulated in matrix form
( 1| ( 1)) ( | ( ))k k k k + + = Le ep p (17)
where
1 1
0 1 0 0 0 0
0 0 1 0 0 0
,
0 0 0 0 0 1
n nk k k-
L = - - -
L
L
M M M M O M M
L
L L L L
(18)
Since L is a stable matrix
( 11
n n
nI z k z k--L = + + +L is stable), the system
(17) is asymptotically stable, which implies
( | ( )) 0k k e p when k . The model tracking
error will converge to zero.
3.2Boundedness of parameters and control signalAssumption 4. In order to deal with the T-S model
bring linear in parameters when performing tracking
387
-
7/28/2019 XTP23_151_Paperstampedno.pdf
4/6
error boundedness analysis, we shall assume that the
premise parameters of the model ( , )c are the same
as the plant parameters ( , )c .
However, updates of all paramters will be carried
during the control simulation studies.
Lemma 1. The adaptive algorithm (13)-(14)
guarantees that ( )kQ% and ( )kQ are bounded.
Lemma 2.If the signal ( )kY is persistently exciting,then the least-squares algorithm (13)-(14) ensures
that lim ( ) 0kk
Q =
% .
These follow from standard properties of the least-square estimation methods applied to linearlyparameterized models.
Applying (9) to (7) yields
( 1) ( 1) ( )Tr
y k y k k+ = + + k e (19)
From assumption 2, ( 1)r
y k+ is bounded and since
( )ke is bounded. Thus, equation (19) implies theboundedness of ( 1)y k+ . Since ( )kx is a vector
consisting of past outputs, ( )kx is bounded.
Bounded ( )kx , bounded parameters ( )kQ and
bounded ( ( ))i kw x imply bounded ( ( ))f kx and
bounded ( ( ))g kx in (8). Thus, from (9) we obtain
the boundedness of ( )u k .
3.3Boundedness of plant tracking error
From (3), we have the plant output evolution.
1
1 1
( 1) ( ( )) 1 ( ) ( )
( ) 1 ( ) ( ) ( ) ( )
( ( )) ( ( )) ( )
N
Ti i T
i
N NT
i i T i i
f g
i i
y k k k u k
k k u k
f k g k u k
w
w w
=
= =
+ =
= +
= +
x x
x x x
x x
(20)Applying (9) to (20) and after some manipulations,we have
( 1) ( ( )) ( ( )) ( )
( ) ( ) ( ) ( ( ) ( ) ( ))
( 1) ( )Tr
y k f k g k u k
f g u k f g u k
y k k
+ = +
= + - +
+ + +
x x
x x x x
k e
(21)
Moving ( 1)r
y k+ to the left side of (21) and after
some manipulation, we get
( 1) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
T T T
T T T
e k k k k k
k k k k
+ = Y Q -Y Q-
= Y Q +Y Q-% %
k e
k e
(22)
where ( )kY denotes ( ( ) | , )kY x c and ( )kY
denotes ( ( ) | ( ), ( ))k k kY x c .
Considering
1 1
( ) ( ) ( )
( ) ( ) ( ) ( )
T T T
N Ni iT i iT
e ei i
k k k
k kw w= =
Y Q = Y Q-Y Q
= -
%
x
x x
x
( )1
1 1
( ) ( ) ( )
( ) ( ) ( ) ( )
Ni i iT
e
iN N
i iT i iT
x x
i i
k
k k
w w
w w
=
= =
= -
- +
x x x
x e x e
(23)
where1 2
Ti i i i
x nq q q = L .
Substituting (23) into (22), the plant tracking error
dynamics can be obtained as follows
( )1
1 1
1
( 1) ( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( )
( ) ( ) ( )
NT i i iT
e
iN N
i iT i iT T
x x
i iN
i iT
x
i
e k k k k
k k k
k k
w w
w w
w
=
= =
=
+ = Y Q + -
- + -
= +G
%x x x
x e x e k e
x e
(24)where
( )
1
1
( ) ( ) ( ) ( ) ( ) ( )
( ) ( ) ( )
NT i iT T
x
i
Ni i iT
e
i
k k k k k
k
w
w w
=
=
G = Y Q - -
+ -
% x e k e
x x x
(25)
The error dynamics (24) can also be written into a
matrix form
1
( 1) ( ) ( ) ( )N
i i
i
k A k k w=
+ = +Ge x e (26)
where
1 2
0 1 0 0 0 0
0 0 1 0 0 0
, ( ) .( )
0 0 0 0 0 1
i
i i i
n
A kk
q q q
= G = G
0
L
L
M M M M O M M
L
L L L L
(27)
Assumption 5. There exists a positive constant a and a common positive definite symmetric matrix
Pfor all the matrices ( 1, , )iA i N= L such that
0 for 1, ,iT iA PA P aI i N- - < = L (28)
Theorem 1. Consider the closed-loop systemconsisting of plant(3), controller(9) and parameter
adaptation law (13)-(14). If the assumptions 1-5 hold,then
(i) the plant tracking error ( )ke is bounded;(ii) if in addition, the signal ( )kY is
persistently exciting, ( )ke converges tozero.
Proof. Consider the following possible Lyapunovfunction
( ( )) ( ) ( )TV k k P k =e e e (29)
Then the difference of the Lyapunov function alongthe plant tracking error trajectory (26) is calculated as
2
1
1 1
1
( 1) ( 1) ( ) ( )
( ( )) ( ) ( ) ( )
( ) ( 2 ) ( )
2 ( ) ( ) ( ) ( ) ( )
T T
Ni T iT i
iN N
i j T iT j jT i
i j i
NT i i T
i
V k P k k P k
k A PA P k
k A PA A PA P k
k P A k k P k
w
w w
w
=
= = +
=
D = + + -
= - +
+ + -
+ G +G G
e e e e
x e e
e e
x e
(30)
388
-
7/28/2019 XTP23_151_Paperstampedno.pdf
5/6
With assumption 5, we have
2 2 for all andiT j jT iA PA A PA P aI i j+ - - (31)
Applying (31) in (30), the following inequality canbe obtained
1 1
1
2 22
max
( ) ( ) ( ) ( )
2 ( ) ( ) ( ) ( ) ( )
1 2( ) ( ) ( )2
N NT i j
i j
NT i i T
i
V k a k
k P A k k P k
Ma k P k a
w w
w
l
= =
=
D -
+ G +G G
- + + G
e x x e
x e
e
(32)
where
{ }max1
( ) max , 1, ( ( ) )N
i i i
ii
M P A i N P x Al w=
= = L
Consider
1
1
( ) ( )
( ) ( ) ( ) ( )
( ) ( ) ( )
NT i iT T
x
i
Ni i i
e
i
k k
k k k
k
w
w w
=
=
G = G
Y Q + +
+ -
% x k e
x x x
(33)
Since all the elements in (33) are bounded, ( )kG is
bounded. Hence there exists positive constant rG
such that
( )k rGG (34)
We can conclude outside the ball
2
min4 2 ( )
( ) : ( )M a P
k ka
le rG
+ < =
e e (35)
the Lyapunov function difference can be bounded by
( ( )) 0 ( )V k k eD " e e (36)
The hypothesis (i) has now been proven.
With Lemma 2, if the signal ( )kY is persistentlyexciting, then the least-squares algorithm (13)-(14)
ensures that lim ( ) 0k
k
Q % . With the assumption 4,
lim ( ) 0k
k
Q % means the model become a perfect
representation of the plant, which will make the
model state ( )kx become the same as the plant state
( )kx . Thus we have lim ( ) ( ) 0i ik
w w
- =x x . With
the control law (9), we have proven lim ( ) 0k
k
e .
Therefore, when k , ( )kG in (33) converges
to zero and (32) becomes
21( ( )) ( )
2V k a k k D - e e (37)
we obtain ( ) 0k e as k . The hypothesis (ii)
and the Theorem 1 have now been proven.
5. SIMULATION RESULTSIn this example, the validity and effectiveness of theproposed control scheme are illustrated through the
tracking control of a single-link robot arm described
by1 2
2 1 1 2 2
( ) ( )
( ) sin[ ( )] ( ) ( )
q t q t
q t q t q t bu t a a
=
= - - +
&
& (38)
where1
( ) ( )y t q t= is the arm position which is the
measured output,2( )q t is the angular velocity which
is unmeasurable and ( )u t is the input torque. The
parameters1
a and2
a depend on the mass and
length of the arm. The values of1
a ,2
a and b are
chosen as1 2
1ba a= = = . The premise variables are
selected as1( ) [( 1) )]x kT y k T= - and
2( ) ( )x kT y kT= .
The sampling time 0.01secT= . The controlobjective is to make1
q track a bounded reference
signal. The initial T-S fuzzy model for closed-loopcontrol is an input-output model obtained by the on-
line identification method proposed in Qi and Brdys(2005) which follows the structure of (38).
0 10 20 30 40 50 60
-1
0
1
2
Time (sec)
Outputs
plant output
reference
0 10 20 30 40 50 60-0.05
0
0.05
0.1
0.15
0.2
Time (sec)
PlantTrackingError
Fig. 2. Plant tracking response with parameter
disturbance (2
a changes from 1 to 1.5 at 5t s= )
0 20 40 60-15
-10
-5
0
5x 10-3
Time (sec)0 20 40 601.99
1.9952
2.0052.01
Time (sec)
0 20 40 60-1-0.998-0.996-0.994-0.992
-0.99
Time (sec)0 20 40 60-1
0
1
23x 10
-3
Time (sec)
a30 a31
a32 a33
Fig. 3. Consequent parameter adaptation of rule 3
0 5 10 15 20 25 30 35 40-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
Time (sec)
Controlinput
Fig. 4. Control signal
Fig. 2 illustrates the plant tracking response withdisturbance and tracking error and Fig. 4 shows the
control signal. The plant parameter2
a changes from
1 to 1.5 at 5t s= , which causes the tracking error to
389
-
7/28/2019 XTP23_151_Paperstampedno.pdf
6/6
increase after 5t s= . However, the controller is ableto adapt to the change and regain good tracking
performance. The tracking error decreases graduallyas the model is learning to become a goodapproximation of the plant. Parameter adaptation of
rule 3 is shown in Fig. 3. The learning rate W is
chosen as 61.0 10W = .
The reference trajectory in Fig. 5 is obtained by
passing a piecewise constant through a first-orderstable linear filter. A disturbance in
1a was
introduced at 44t s= . As a result, it can be seen thatthe tracking error increases after 44s . The parameter
adaptations are shown in Fig. 6.The control law canproduce control signal that adapts to the change and
regain good tracking performance.
0 10 20 30 40 50 60 70 80 90 100
-1
0
1
2
Time (sec)
Outputs
0 10 20 30 40 50 60 70 80 90 100
-0.2
0
0.2
0.4
0.6
0.8
Time (sec)
PlantTrackingError
plant output
reference
Fig. 5. Plant tracking response with parameter
disturbance (1
a changes from 1 to 1.5 at 44t s= )
0 50 100-0.1
-0.05
0
0.05
0.1
Time (sec)
0 50 1001.99
1.995
2
2.005
Time (sec)
0 50 100-1
-0.98
-0.96
-0.94
Time (sec)
0 50 1000
1
x 10-4
Time (sec)
30
^33
^31
32
Fig. 6. Consequent parameter adaptation of rule 2
6. CONCLUSION
The indirect adaptive fuzzy control algorithm for
uncertain nonlinear stable plants is presented in thispaper. A discrete-time T-S fuzzy model is employedas a dynamical model of an unknown continuous-
time plant with unmeasurable states. By propertraining, the T-S fuzzy model is able to achieveinput-output behaviour close to the plant and provideuseful information about the states of the plant. A
feedback linearization control is calculated entirelybased on the model, using the model structure,
parameters and states, while the parameters of themodel are updated on-line by RLSE method.Sufficient conditions stability and tracking error
convergence are derived for stable nonlinear plant. Itshould be noted here that although the errorboundedness or the closed-loop global stability are
proven under the assumption that the premiseparameters of the model are the same as in the plant,a local stability or the error boundedness are
achieved if the initial parameter values are closeenough to the plant parameter values. Simulationexamples verify the effectiveness, adaptation and
tracking performance of the proposed control scheme.
ACKNOWLEDGEMENT
This work has been partly supported by UniversitiesUK on behalf of the Department for Education and
Skills (DfES) through the UK funding councils andby the Polish State Committee for ScientificResearch under grant No. 4 T11A 008 25.
REFERENCES
Golea, N., Golea, A. and Benmahammed, K. (2003),
Stabel indirect fuzzy adaptive control, Fuzzy
Sets and Syst., vol. 137, pp. 353-366.Kulawski, G. J. and Brdys, M. A. (2000), Stable
adaptive control with recurrent networks,
Automatica, vol. 36, pp. 5-22.Park, C.-W. and Cho, Y.-W. (2004), T-S Model
Based Indirect Adaptive Fuzzy Control UsingOnline Parameter Estimation, IEEE Trans.Syst. Man. And Cyber., Part B, Vol. 34, pp.
2293-2302.Qi, R. and Brdys, M.A. (2005), Adaptive fuzzy
modelling and control for discrete-timenonlinear uncertain systems, in Proc. Of the
2005 American Control Conference, pp. 1108-1113.
Spooner, J. T. and Passino, K. M., Stable adaptivecontrol using fuzzy systems and neuralnetworks,IEEE Trans. Fuzzy Syst., vol. 4, pp.
339-359, Aug. 1996.Takagi, T. and Sugeno, M. (1985), Fuzzy
idenfication of systems and its applications to
modeling and control,IEEE Trans. Syst., Man,Cybern., vol. SMC-15, pp. 116-132.
Tao, G. (2003), Adaptive design and analysis.
Hoboken, NJ: John Wiley & Sons, Inc..Wang, L.-X. (1994), Adaptive Fuzzy Systems and
Control: Design and Stability Analysis.Englewood Cliffs, NJ: Prentice-Hall.Ying, H. (1998), General MISO Takagi-Sugeno
fuzzy systems with simplified linear ruleconsequent as universal approximators forcontrol and modeling applications, IEEETrans. Fuzzy Syst, vol. 6, pp. 582-587.
Ying, H. (1999), Comparison of necessaryconditions for typical Takagi-Sugeno andMamdani fuzzy systems as universal
approximators, IEEE Trans. Syst., Man,Cybern., vol. 29, pp. 508-514.
Yoo, B. and Ham, W. (1998), Adaptive fuzzy
sliding mode control of nonlinear systems,IEEE Trans. Fuzzy Syst., vol. 6, pp. 315-321.
390