J Control Theory Appl 2010 8 (2) 229–232
DOI 10.1007/s11768-010-7248-6
Nonfragile control for a class of
uncertain switching fuzzy time-delay systems
Yi LIU, Jun ZHAO
(
Key Laboratory of Integrated Automation of Process Industry, Ministry of Education, Northeastern University,
Shenyang Liaoning 110004, China)
Abstract: A sufficient condition for the existence of nonfragile state-feedback controllers and a switching law for a
class of uncertain switching fuzzy time-delay systems are proposed based on the multiple Lyapunov functions technique,
under the assumption that the controller gain to be designed has variations. The corresponding results are all formulated in
terms of linear matrix inequalities (LMIs). Simulation results show the effectiveness of the design method.
Keywords: Switching system; Fuzzy control; Nonfragile control; Multiple Lyapunov functions
1 Introduction
For a fuzzy system, if the system includes discrete part,
such as discrete switching signal, the system is a switching
fuzzy system. For fuzzy systems, many results are available;
for example, the stability problem was studied in [1], the
robust output feedback control was studied based on fuzzy
observer in [2], and the modelling and robust H
∞
control
problem of the fuzzy system were considered in [3]. For
switching systems, the construction method of Lyapunov
function and the stability analysis and controller design
were addressed in [4∼8]. However, the research results of
the switching fuzzy systems are very limited compared with
those of switching systems and fuzzy systems. The smooth-
ness condition of switching fuzzy systems was studied in
[9]. The controller design and stable condition of switching
fuzzy systems were obtained based on switching Lyapunov
function method in [10]. The controller switching problem
of the fuzzy systems with uncertainties was addressed in
[11].
On the other hand, controllers often suffer from distur-
bances and/or variations, which may deteriorate the system
performances or even destabilize the system. Therefore, the
nonfragile control strategy has been attracting a lot of atten-
tion in the area of control study. For nonswitching systems,
nonfragile control has been well explored, and many results
are available in the literature (see, for example, [12]). For
switching systems, very few results have appeared on the
topic. In [13], an H
∞
nonfragile controller was designed
via Riccati inequalities. A sufficient condition for the ex-
istence of nonfragile state feedback controllers was given
in [14] by using the multiple Lyapunov functions technique
and the average dwell time method. However, to the best of
our knowledge, no results have been reported on nonfragile
control for switching fuzzy systems, which motivates the
present study.
In this paper, we study nonfragile control for switch-
ing fuzzy systems with delays. A sufficient condition for
the existence of nonfragile state-feedback controllers and a
switching law are proposed by using the multiple Lyapunov
function technique, under the assumption that the controller
gain to be designed has variations. The corresponding re-
sults are all formulated in terms of linear matrix inequalities
(LMIs).
2 Problem statement
Consider the switching fuzzy time-delay system with un-
certainties described as follows:
R
i
σ
: If z
1
(t) is M
i
σ1
, ··· ,z
p
(t) is M
i
σp
, then
˙x(t)=(A
σi
+ΔA
σi
)x(t)+(A
hσi
+ΔA
hσi
)x(t−h)
+(B
σi
+ΔB
σi
)u
σ
(t)],
x(t)=Ψ(t),t∈ [−h, 0],i=1, 2, ··· ,N
σ
. (1)
where M
i
σj
represent fuzzy subsets; z(t)=[z
1
(t),z
2
(t),
··· ,z
p
(t)] is the premise vector. σ ∈ M = {1, 2, ··· ,l}
is a piecewise constant function representing the switching
signal; x(t) ∈ R
n
is the system state vector; u
σ
(t) ∈ R
m
is the system input vector; A
σi
,A
hσi
,B
σi
are the system
matrices; ΔA
σi
, ΔA
hσi
, ΔB
σi
are time-varying matrices
of appropriate dimensions, which represent uncertainties of
the system; h denotes the constant delay factor in the sys-
tem.
We consider the state-dependent switching signal of the
form σ = σ(x(t)). Suppose {
˜
Ω
1
,
˜
Ω
2
, ··· ,
˜
Ω
l
}is a partition
of R
n
\{0}, i.e.
l
i=1
˜
Ω
i
= R
n
\{0}, and Ω
i
Ω
j
= ∅,i =
j, ∅ denotes empty set. We will design a switching law of
the form:
σ(x(t)) = r, when x(t) ∈
˜
Ω
r
. (2)
Define a function by
ν
r
(x(t)) =
1,x(t) ∈
˜
Ω
r
,
0,x(t) /∈
˜
Ω
r
.
Received 5 November 2007; 23 June 2009.
This work was supported by the National Natural Science Foundation of China (No.60874024, 60574013, 90816028), and the Specialized Research
Fund for the Doctoral Program of Higher Education of China (No.200801450019).
c
South China University of Technology and Academy of Mathematics and Systems Science, CAS and Springer-Verlag Berlin Heidelberg 2010