F. Zhao et al.
fuzzy system is a systematic representation for nonlin-
ear systems. It is well-known that, by using the T–S
fuzzy model, the global behavior of a nonlinear system
can be represented by a weighted sum of some locally
simple linear subsystems. Many issues related to the
stability analysis and control synthesis of T–S fuzzy
systems have been studied [6–11] in the past decades.
Recently, a class of fuzzy systems which is described
by the singular form have been considered [12–14],
where the model is an extension of T–S fuzzy model.
A singular system is also called a descriptor system, dif-
ferential algebraic system or a generalized state-space
system, which arises from a convenient and natural
modeling process in characterizing a class of practical
systems. In [12], the problem of stability analysis for
nonlinear singular systems with Markovian jumping
parameters and mode-dependent interval time-varying
delay was studied. New delay-dependent stability con-
ditions were derived by constructing a mode-dependent
Lyapunov function and using integral inequalities. A
robust observer-based output feedback control for sin-
gular fuzzy systems in the presence of immeasurable
states, approximation errors and uncertainty was pro-
posed in [14]. Time-delay phenomena were first discov-
ered in biological systems [40] and were later found in
various practical systems, such as networked systems
[41,42], population dynamics [43] and communica-
tion systems. They are often a source of instability and
poor control performance. In the above actual systems,
there is a phenomenon that the random and time-delay
are presented at the same time. In order to accurately
describe the dynamic characteristics of the systems, the
stochastic time-delay systems are established and have
attracted the attention of many researchers [34–37]. A
large variety of important and interesting methods have
been proposed for the analysis and design of singular
fuzzy systems with time-delay [15,16].
The H
∞
filtering problem for singular systems has
been of continuous interest because of its wide applica-
tions [17,18]. Based on the fact that T–S fuzzy model
is a powerful tool to describe a nonlinear system, some
authors used fuzzy approach to investigate the filter-
ing problem for nonlinear singular systems with time-
delay [19]. By applying T–S fuzzy model, a nonlinear
dynamic system can be transformed to a s et of lin-
ear subsystems via fuzzy rules. In this type of fuzzy
model, local dynamics in different state-space regions
are represented by linear models. So we can study the
filtering problem of nonlinear systems by employing
these methods which are used to deal with the filtering
problem for linear systems. It should be noted that the
H
∞
filtering problem of linear systems has been stud-
ied, and a great number of results have been reported
(see [20–22]). However, compared with linear systems,
the H
∞
filtering problem for nonlinear systems has not
been fully investigated although it is important in con-
trol design and signal processing [23,24].
The term of input-to-state stable (ISS) was proposed
by Sontag [25]. It plays an important role in stabil-
ity analysis and controller design of deterministic non-
linear systems. The ISS of control systems has been
widely studied, and many results have been obtained.
Meanwhile, there have been various extensions for ISS,
such as integral input-to-state stable (iISS), input-to-
state stable in mean (ISSiM), e
λt
-weighted iISSiM and
so on [26,27].
In addition, there is H
∞
filtering problem reported
for a class of special nonlinear systems [38,39]. In [38],
the H
∞
filtering problem was considered for a class of
stochastic nonlinear systems with time-delay and the
nonlinear term satisfying a Lipschitz constraint. We
know that a T–S fuzzy system is a systematic repre-
sentation for general nonlinear systems. Different from
the special nonlinear systems, the H
∞
filtering problem
is considered for general nonlinear systems which are
descried by T–S fuzzy method in this paper. The focus is
on the design of a fuzzy filter such that the correspond-
ing filtering error system is e
λt
-weighted iISSiM and
the H
∞
attenuation level from noise to estimation error
is below a prescribed scalar. Based on an auxiliary vec-
tor, an integral inequality and a linear matrix inequal-
ities (LMIs) technique, a set of sufficient conditions
is proposed to ensure that the filtering error system is
e
λt
-weighted iISSiM. The desired fuzzy filter is estab-
lished in terms of a set of linear matrix inequalities.
Three examples are provided to illustrate the effective-
ness of the proposed fuzzy filter design method.
Notations The symbols R and R
+
denote the set of
real numbers and the set of nonnegative real numbers,
respectively. R
n
denotes the n-dimensional Euclidean
space. The superscript ‘T ’ stands for matrix transpo-
sition. ε{·} denotes the expectation. The expression
α ∈ K
∞
denotes that α is a K
∞
function. L
2
[0, ∞)
is the space of square-integrable vector functions over
[0, ∞). I denotes the identity matrix. The expression
A < B means that the matrix B − A is positive defi-
nite. λ
max
(A) and λ
min
(A) are used to denote the max-
123