𝐻
∞
Tracking Control for Time-Delay Systems with
Markovian Jumping Parameters
Baoyong Zhang Shengyuan Xu
School of Automation
Nanjing University of Science and Technology
Nanjing 210094, Jiangsu, China
Email: baoyongzhang@yahoo.com.cn
Yongmin Li
School of Science
Huzhou Teachers College
Huzhou 313000, Zhejiang, China
Email: ymlwww@yahoo.com.cn
Abstract—This paper is concerned with the 𝐻
∞
tracking
control problem for linear time-delay systems with Markovian
jumping parameters. The objective is to design state-feedback
mode-dependent controllers such that the resulting closed-loop
system is stochastically stable and its state follows a reference
signal. Based on the Lyapunov-Krasovskii functional method
together with the delay-partitioning technique, delay-dependent
conditions for the existence of the desired controllers are obtained
in terms of linear matrix inequalities. In addition, a new stability
criterion for Markovian jump systems with constant delays is
also given, which is less conservative than the recent ones in the
literature. Numerical examples are provided to demonstrate the
effectiveness of the proposed methods.
I. INTRODUCTION
The objective of the tracking control problems is to design a
controller such that the state or output of the controlled system
follows a reference signal. Such problems have strong back-
grounds in the engineering applications, and they have also
been involved in some newly developed areas. For example, a
growing attention has been paid to the tracking control prob-
lems for multi-agent networks. This topic involves a number
of interesting issues, such as consensus tracking, coordinated
tracking and swarm tracking [19]. Thus, the study of tracking
control for dynamic systems is of great importance especially
due to the growing demands from the newly developed areas.
The 𝐻
∞
tracking aims at attenuating the effect of dis-
turbances on the corresponding tracking error. This problem
has been investigated for a few of systems, such as robotic
systems [4], nonlinear input affine systems [1], T-S fuzzy
systems [5], [17] and networked control systems [13], [21].
The problems of 𝐻
∞
tracking control for continuous-time and
discrete-time linear systems with Markovian jump parameters
were considered in [3] and [18], respectively. For Markovian
jump systems with time delays, the robust 𝐻
∞
tracking control
problem was investigated in [6], where sufficient conditions
for the existence of desired controllers were obtained in terms
of linear matrix inequalities (LMIs). It should be pointed
out that the results obtained in [6] are independent from the
size of delays, which are often more conservative than delay-
dependent conditions [22]. Therefore, it is necessary to develop
new approaches for the 𝐻
∞
tracking control problem of time-
delay systems with Markovian jump parameters.
In this paper, we deal with the problem of 𝐻
∞
state
tracking control for linear systems with constant delays and
Markovian jump parameters. Our research efforts are focused
on developing a new approach to solve the addressed problem
rather than applying the existing methods directly. Firstly,
a novel Lyapunov functional is constructed based on the
ideas of delay-partitioning technique [11]. Secondly, delay-
dependent conditions for the tracking performance analysis of
the resulting closed-loop augmented system are obtained. This
result implies a new stability criterion for delayed Markovian
jump systems. A numerical example is provided to show
the less conservatism of the obtained stability criterion in
comparison with those in [9] and [23]. Finally, sufficient
conditions for the existence of desired state-feedback mode-
dependent controllers are obtained in terms of LMIs. The
effectiveness of the proposed design method is demonstrated
through a numerical example.
Notations. Throughout this paper, for real symmetric ma-
trices 𝑋 and 𝑌 , the notation 𝑋 ≥ 𝑌 (respectively, 𝑋>𝑌)
means that the matrix 𝑋 − 𝑌 is positive semi-definite (re-
spectively, positive definite). ℝ
𝑛
denotes the 𝑛 dimensional
Euclidean space and ℝ
𝑚×𝑛
denotes the set of all 𝑚 × 𝑛 real
matrices. 𝐼 denotes an identity matrix of appropriate dimen-
sion. 𝐼
𝑛
denotes a 𝑛-dimensional identity matrix. 𝑂
𝑛
and 𝑂
𝑛,𝑚
denote 𝑛 ×𝑛-dimensional zero matrix and 𝑛 ×𝑚-dimensional
zero matrix, respectively. The superscript ‘𝑇 ’ represents the
transpose. The notation ‘∗’ is used as an ellipsis for terms
that are induced by symmetry. diag(...) stands for a block-
diagonal matrix. The notation ∣⋅∣ refers to the Euclidean vector
norm. 𝔼 {⋅} stands for the mathematical expectation operator
with respect to some given probability measure. sym(𝐴) is
defined as 𝐴 + 𝐴
𝑇
. Matrix dimensions, if not explicitly stated,
are assumed to be compatible for algebraic operations.
II. P
ROBLEM FORMULATION
Given a probability space (Ω, 𝔽, ℙ), we consider the follow-
ing linear system with a constant delay and Markovian jumping
parameters [2], [23], [26]:
˙𝑥 (𝑡)=𝐴 (𝑟
𝑡
) 𝑥 (𝑡)+𝐴
𝑑
(𝑟
𝑡
) 𝑥 (𝑡 − 𝑑)
+𝐵
𝑢
(𝑟
𝑡
) 𝑢 (𝑡)+𝐵
𝜔
(𝑟
𝑡
) 𝜔 (𝑡) ,
𝑥 (𝑡)=𝜑 (𝑡) , ∀𝑡 ∈ [−2𝑑, 0] ,
(1)
978-1-4673-1872-3/12/$31.00 ©2012 IEEE 1501
Fr33.2
2012 12th International Conference on Control, Automation, Robotics & Vision
Guan
zhou, China, 5-7th December 2012 (ICARCV 2012)