International Journal of Control, Automation, and Systems (2014) 12(6):1-9
DOI
ISSN:1598-6446 eISSN:2005-4092
http://www.springer.com/12555
H
∞
Control of LPV Systems with Randomly Multi-Step Sensor Delays
Yilian Zhang, Fuwen Yang*, and Qing-Long Han
Abstract: This paper is concerned with the H
∞
control problem for a class of linear parameter-varying
(LPV) systems with randomly multi-step sensor delays. A mathematical model which describes the
randomly multi-step sensor delayed measurements for LPV systems is established. An improved Lya-
punov functional is proposed to determine the asymptotically mean-square stability of the closed-loop
system which depends on the varying parameters. The obtained full-order parameter-dependent dy-
namic feedback controller guarantees the considered system to be asymptotically mean-square stable
and to satisfy the modified H
∞
performance for all possible delayed measurements. An extended cone
complementarity linearization method (CCLM) is used to solve the constrained linear matrix inequality
(CLMI). Simulation results illustrate the effectiveness of the proposed method.
Keywords: CCLM, H
∞
control, LPV systems, randomly multi-step sensor delays.
1. INTRODUCTION
The LPV technique has received much research
attention in control area because of its effectiveness in
dealing with nonlinear systems. It is based on the LPV
system whose state-space matrices depend on a prior
unknown but measurable time-varying parameter vector
and it allows linear control methods to be applied to
nonlinear systems, such as magnetic bearing systems,
aerospace systems and vehicle systems [1-3]. In recent
decades, significant advances have been made in the
study under the LPV framework and a number of
publications have been reported [4-7].
The randomly sensor delay is a common phenomenon
that often occurs in practical systems, such as networked
systems, remote monitoring and chemical process. It may
be induced by the intermittent sensor failures, the
asynchronous time-division-multiplexed network and the
randomly congestion of packet transmission, etc [8]. In
real-world applications, filtering and control methods
designed for systems without considering the randomly
sensor delay may not be able to satisfy the stability
requirements and the performance constraints. Yaz and
Ray [9] performed the first study on linear unbiased state
estimation for dynamic systems with randomly sensor
delay. The past decade has witnessed the emergence and
development of the control and filtering methods for
systems with randomly sensor delays [10-13]. Typically,
it is assumed that the delay is one-step sensor delay,
which is defined by the time-stamped measurements. A
stochastic parameter is used to describe the probability of
the occurrence of the one-step sensor delay. However, in
practical systems, the length of a delay that the
measurements may suffer is normally unknown in
advance. Thus, it is more acceptable to assume that the
measurements are with multi-step sensor delays, which
represents the case that the randomly delays are different
sample steps long at different time instant. Moayedi, Foo
and Soh [14] studied the filtering problems for
networked control systems (NCSs) with single/multiple
measurement packets subject to multi-step sensor delays
and multiple packet dropouts. Sun and Xiao [15]
concerned with the optimal linear estimation problem for
linear discrete-time stochastic systems with possible
randomly sensor delays which are limited within a
known bound. Nevertheless, there are still few
publications on control for systems with randomly multi-
step sensor delays despite their practical importance.
Recently, the researches on LPV systems have
involved systems with time delays and theoretical results
have been published. Wu and Grigoriadis [4] presented
parameter-dependent state feedback controllers for LPV
systems with time-delays. Sun, Du and Yuan considered
a delay-dependent H
∞
control problem for LPV systems
in [5]. Mohammadpour and Grigoriadis [6] analyzed the
stability and performance of state-delay LPV systems.
However, in the references mentioned above, it is
implicitly assumed that the considered LPV systems
contain only state delays but no sensor delays.
© ICROS, KIEE and Springer 2014
__________
Manuscript received January 30, 2014; revised April 22, 2014;
accepted April 28, 2014. Recommended by Editor Ju Hyun Park.
This work was supported in part by the National Natural
Science Foundation of China under Grant 61174064, in part by
the Australian Research Council Discovery Project under Grant
DP1096780, in part by the RDI Merit Grant Scheme Project under
Grant RSH/2028 at Central Queensland University, Australia, and
in part by China Scholarship Council.
Yilian Zhang is with the School of Information Science and
Engineering, East China University of Science and Technology,
Shanghai 200237, China (e-mail: zyl1030@126.com).
Fuwen Yang was with the School of Information Science and
Engineering, East China University of Science and Technology,
Shanghai 200237, China. He is now with the Centre for Intelligent
and Networked Systems, Central Queensland University, Rock-
hampton, QLD 4702, Australia (e-mail: f.yang@cqu.edu.au).
Qing-Long Han is with the Centre for Intelligent and Network-
ed Systems, Central Queensland University, Rockhampton, QLD
4702, Australia (e-mail: q.han@cqu.edu.au).
* Corresponding author.