Robust H
∞
Filtering for Stochastic Networked Control Systems
WU Junli
1
, KARIMI Hamid Reza
2
, TIAN Siqing
1
1. Information and Electronic Technology Institute, Jiamusi University, Jiamusi, 154002, P. R. China
E-mail: wfdwjl@gmail.com
2. Department of Engineering, Faculty of Engineering and Science, University of Agder, Grimstad, N-4898, Norway
E-mail: hamid.r.karimi@uia.no
Abstract: In this paper, the problem of robust H
∞
filtering for stochastic networked control systems is investigated. The pa-
rameter uncertainties are time-varying norm-bounded and appear in both the state and input matrices. In the networked control
systems (NCSs), the problems of measurement quantization, signal transmission delay and data packet dropout are considered.
H
∞
filter is designed such that the filtering error system is asymptotically stable in the mean square, and the L
2
-induced gain from
the noise signal to the estimation error is guaranteed to be less than a prescribed level. Finally, an example is given to show the
effectiveness of the theoretical results.
Key Words: Robust H
∞
filtering, NCSs, Stochastic sytems
1 Introduction
The objective of filtering is to estimate the state vari-
ables of a system using output measurements. In the past
decades, a great deal of interest has been devoted to the s-
tudy of filtering problem, mainly including Kalman filtering
[1, 2], H
∞
filtering [3–5], H
2
-H
∞
filtering [6]. Considering
the unknown statistics of the external noises, H
∞
filtering
has been studied by many researchers owing to the fact the
noise sources are arbitrary signals with bounded energy in
H
∞
filtering. In [7], the H
∞
filtering problem for linear con-
tinuous singular systems was studied using the linear matrix
inequalities technique. in [8], the author studied the problem
of delay-dependent energy-to-peak filter design for a class of
stochastic time-delay systems.
On another research frontier, network-based control has
gained considerable research interests in control area since
the network technique is rapidly developed and the com-
puter network is widely used in modern control systems.
Networked control systems (NCSs) are a class of systems
communicating with sensors and actuators over a commu-
nication network. NCSs have found successful application-
s in a wide range of areas, however, there exit three typi-
cal problems of quantization, delays and packet dropouts in
the network, which inevitably degrade control performance.
Therefore, more and more attention has been paid to address
these problems. To mentioned a few, In [9], the authors
firstly investigated the problem of H filtering with simultane-
ous consideration of signal transmis- sion delay, data pack-
et dropout, and signal quantization. In [10], the distribut-
ed finite-horizon filtering problem was concerned with for
a class of time-varying systems over lossy sensor networks,
in which the time-varying system (target plant) is subject to
randomly varying nonlinearities (RVNs) caused by environ-
mental circumstances.
In most existing work, it is assumed that the physical plant
are deterministic for filter design in the network environmen-
t. However, it is necessary to form some practical systems
as stochastic systems models such that the precise results of
the closed-loop system will be obtained. This has motivated
This work is supported by Jiamusi University Research Foundation un-
der Grant Lq2011-027 and National Natural Science Foundation (NNSF) of
China under Grant 61203052.
the researches on filter design for stochastic systems in the
network environment and many results have been reported in
the literature. In [11], the exponential synchronization prob-
lem was investigated for an array of N linearly coupled com-
plex networks with Markovian jump and mixed time-delays.
In the existing literature, the robust H
∞
filtering problem for
a stochastic systems in the network environment involving
quantization, delays and packet dropouts has not yet gained
sufficient research attention. The purpose of this paper is to
study this problem.
In this paper, we study the problem of robust H
∞
filter-
ing for stochastic networked control systems. The parameter
uncertainties are time-varying norm-bounded and appear in
both the state and input matrices. In the networked control
systems (NCSs), the problems of measurement quantization,
signal transmission delay and data packet dropout are con-
sidered. H
∞
filter is designed such that the filtering error
system is asymptotically stable in the mean square, and the
L
2
-induced gain from the noise signal to the estimation error
is guaranteed to be less than a prescribed level. Finally, an
example is given to show the effectiveness of the theoretical
results.
Notation: The notation used throughout the paper is fairly
standard. R
n
denotes the n-dimensional Euclidean space and
the notation P > 0 (≥ 0) means that P is real symmetric and
positive definite (semi-definite). In symmetric block matri-
ces or complex matrix expressions, we use an asterisk (∗)to
represent a term that is induced by symmetry and diag{···}
stands for a block-diagonal matrix. E{x} means the expec-
tation of x. The space of square-integrable vector function-
sover[0, ∞) is denoted by L
2
[0, ∞), and for w = {w(t)}∈
L
2
[0, ∞), its norm is given by
w
2
=
∞
t=0
|
w(t)
|
2
dt .
2 Problem Formulation
The filtering problem for the stochastic networked control
systems is shown in Fig. 1. Suppose the stochastic plant is
given by
dx(t)=[A(t)x (t)+B(t)v(t)]dt + E(t)x(t)d
ω
(t),
dy(t)=[Cx(t)+Dw(t)]dt ,
z(t)=Lx(t),
x(t)=
φ
(t), t ∈ [−
τ
, 0], (1)
Proceedings of the 33rd Chinese Control Conference
Jul
28-30, 2014, Nan
in
, China
4331