Network-based H
∞
Control for Stochastic Time-delay Systems
via Average Dwell Time Approach
FAN Rongrong
1
, ZHANG Xiaomei
2
, LU Guoping
1,3
1. School of Electrical Engineering, Nantong University, Nantong 226019, P. R. China
E-mail: fanrongrong86@163.com
2. School of Electronics and Information, Nantong University, Nantong 226019, P. R. China
E-mail: zhang.xm@ntu.edu.cn
3. Jiangsu College of Information Technology, Wuxi 214153, P. R. China
Corresponding author. Email: lu.gp@ntu.edu.cn
Abstract: This paper investigates a problem of network-based H
∞
control for stochastic time-delay systems. Taking into
account network-induced delay and data packet dropout, the networked control system is modeled as a switched stochastic time-
delay system. By applying an average dwell time approach, a sufficient condition is developed for the mean-square exponential
stability of the resulting closed-loop system with a weighted H
∞
performance index. Finally, a numerical example is provided
to illustrate the effectiveness of the proposed method.
Key Words: Network-based control, stochastic system, time-delay system, switched system, average dwell time
1 Introduction
Since stochastic systems play a more and more important
role in many branches of science and engineering applica-
tions, a plenty of relevant research results have been pre-
sented during the past decades[1–3]. In [1], sliding mode
control for uncertain stochastic systems with time-varying
delay was concerned. In [2], the robust H
∞
filtering prob-
lem for a class of nonlinear discrete time-delay stochastic
systems was studied. In [3], H
∞
filtering problem for a
class of uncertain nonlinear time-delay stochastic systems
was studied, where nonlinearities were sector-bounded. In
[4], the stability and performance analysis, stabilization and
H
∞
control problems for continuous-time switched stochas-
tic systems were concerned.
Because of the network channel, there usually exist
time delay and data packet dropouts when transmitting the
data[5–9]. For example, the probability distribution of the
random delay and packet dropout is utilized to deal with
some problems of networked systems, such as stabilization
[7], robust H
∞
control [8] and reliable control [9]. Moti-
vated by [6, 10], this paper intends to study the H
∞
con-
trol problem for stochastic time-delay systems with network-
induced delay and data packet dropout. The networked
closed-loop system is modeled as a switched system con-
sists of two subsystems with small time-varying delay and
large time-varying delay, respectively.
Notations: Throughout this paper, P>0(< 0, ≥ 0, ≤)
stands for a symmetrical positive(negative, semi-positive,
semi-negative) definite matrix P . diag{· · ·} stands for a
block-diagonal matrix. L and E {·} denote the infinites-
imal operator and the expectation operator with respect to
some probability measure P , respectively. The notation ∗
in a symmetric matrix denotes the symmetric block in the
matrix. λ
max
(Q)(respectively, λ
min
(Q)) stand for the max-
imum (respectively, minimum) eigenvalue of the matrix Q.
This work is supported by National Natural Science Foundation
(NNSF) of China under Grants no. 61174065, no. 61174066 and
no. 61004027, Natural Science Foundation from Jiangsu Province no.
BK2010275.
·denotes the Euclidean norm of vectors or the spec-
tral norm of matrices. Matrices, if not explicitly stated, are
assumed to have compatible dimensions for algebraic oper-
ations.
2 Problem Statements and Preliminaries
Consider the following plant
dx(t)=[Ax(t)+A
τ
x(t − τ)+Bu(t)+Dv(t)]dt
+[Ex(t)+E
τ
x(t − τ)]dω(t)
z(t)=Cx(t)+C
τ
x(t − τ),
x(t)=φ(t),t∈ [−τ,0] (1)
where x(t) ∈ R
n
, u(t) ∈ R
m
, z(t) ∈ R
q
are the state,
control input and controlled output, respectively; v(t) ∈
R
p
is the disturbance input and belongs to L
2
[0, +∞);
ω(t) is a scalar Brownian motion defined on a complete
probability space (Ω, F , P) with a filtration {F
t
}
t≥0
;
A, A
τ
,B,C,C
τ
,D,E,E
τ
are known constant matri-
ces and τ is the constant delay. The initial state φ(t) is a
continuous function on [−h, 0].
In the networked control system under consideration, it is
supposed that the sensor is clock driven while the controller
and the actuator are event driven. Moreover, we assume that
zero-order hold is adopted and the data are transmitted with
a single packet.
Similar to [5], the control input u(t) is given by
u(t)=Kx(i
k
h),t∈ [i
k
h + τ
k
,i
k+1
h + τ
k+1
) (2)
where K is a known constant matrix, h the sampling pe-
riod, i
k
(k =1, 2,...) some integers and {i
1
,i
2
,i
3
,...}⊂
{1, 2, 3,...}. Time-delay τ
k
denotes the time from the in-
stant i
k
h when sensor nodes sample sensor data from a
plant to the instant when actuators transfer data to the plant.
Clearly,
+∞
k=1
[i
k
h + τ
k
,i
k+1
h + τ
k+1
)=[t
0
, +∞),t
0
=
i
1
h + τ
1
≥ 0.
Let d(t)=t − i
k
h, then system (1) can be rewritten as
dx(t)=[Ax(t)+A
τ
x(t − τ)+BKx(t − d(t))
3URFHHGLQJVRIWKHVW&KLQHVH&RQWURO&RQIHUHQFH
-XO\+HIHL&KLQD
1645