Stabilization of Dynamic Quantized System with Faults
DUAN Kai
1
, CAI Yunze
1
, ZHANG Weidong
1
1. Department of Automation, Shanghai Jiao Tong University, and Key Laboratory of System Control and
Information Processing, Ministry of Education of China, Shanghai 200240
E-mail: yzcai@sjtu.edu.cn
, wdzhang@sjtu.edu.cn
Abstract: In this paper, we mainly focus on the problem of quantized feedback stabilization for a discrete-time linear system
with faults. The dynamic uniform quantizer is used here, and quantized measurements of the state are available. The research
contents are divided into two parts. And the feedback stabilization of the system with sensor or actuator faults is studied
respectively.
The sufficient conditions that the quantizer should satisfy are given. Numerical examples show that the results are
effective.
Key Words: Stabilization, networked control systems, state quantization, faults
1 Introduction
Networked control systems (NCSs) have been one of the
main research focuses in theory as well as in practice for
many years. Compared with traditional point-to-point
feedback control systems, the NCSs introduce many new
challenges in control system design, including time delays,
packets dropout, data quantization, etc[1-3].
When the transmission bandwidth is finite, the
quantization problem of the NCSs arises. There are two
types of quantization policies. For the static quantizer,
researches can be seen in [4,5]. For the dynamic quantizer,
[6] designed stabilizing uniform quantizers for linear
systems and non-linear systems. [7] studied the problem of
quantized feedback stabilization for the stochastic system
with multiplicative noises. The dynamic quantizer scales the
quantization levels dynamically to improve the steady-state
performance and it is used in this paper.
In most literatures concerning NCSs, the consecutive
signal can always be obtained by the controller and plant.
However, affected by aging and disturbances, temporary
failure of the sensors or actuators may happen[8]. To the
best of our knowledge, there are few papers dealing with the
reliable stabilization for general NCSs, not to mention the
data quantization is considered at the same time. This
motivates us to investigate the quantized feedback
stabilization problem when there are sensor or actuator
faults.
This paper is organized as follows: Section 2 gives the
problem statement. Section 3 presents the main results of
this paper, where the state quantization is considered.
Numerical examples are designed in Section 4. Some
conclusions are given in Section 5.
Notation:
1.
R
,
n
R
,
mn
R
u
and
Z
stand for the set of real numbers,
the
n
-dimensional Euclidean space, the set of all real
matrices with
m
rows and
n
columns and the set of
integers, respectively.
*
This work is supported by the National Natural Science Foundation
(NNSF) of China under Grant 61473183, 61221003, 61374160, Program of
Shanghai Subject Chief Scientist 14XD1402400, and SJTU M&E Joint
Research Foundation YG2013MS04.
2. Given a matrix
M
,
T
M
and
1
M
denote its transpose and
inverse respectively, if it exists.
3. Given a matrix
N
,
max
()N
O
and
min
()N
O
represent its
maximum eigenvalue and minimum eigenvalue,
respectively.
4. We will denote by
x
the standard Euclidean norm of a
vector
n
xR
and by
A
the induced norm of a matrix
nn
AR
u
.
5. The notation
()0P !t
for
nn
PR
u
, means that
P
is a
symmetric positive definite (positive semi-definite) matrix.
6.
x
«»
¬¼
denotes the largest integer
k
such that
kxd
, i.e.,
max{ : }xkZkx d
«»
¬¼
.
7.
x
ªº
«»
denotes the smallest integer
k
such that
kxt
, i.e.,
min{ : }xkZkx t
ªº
«»
.
2 Problem statement
2.1 System description
In this paper, the system we focus on is the discrete-time
linear system
(1) () ()xt Axt Bvt
(1)
() () ()v t Nu t NKLx t
(2)
Where
()
n
xt R
is the system state,
()
m
vt R
is the
actuator output,
()
m
ut R
is the control input.
A
and
B
are appropriately dimensional real-valued matrices.
^`
1
,,
ms
LdiagL L
and
^`
1
,,
ma
NdiagN N }
are
diagonal matrixes and are called the fault matrixes. Here
0L z
and
0N z
. Wherein ms and ma are the number of
the sensors and actuators respectively.
i
L
is the fault status
of the
th
i
sensor,
i
N
is the fault status of the
th
i
actuator,
‘1’ for normal and ‘0’ for fail.
K
is the controller feedback
gain.
Proceedings of the 34th Chinese Control Conference
Jul
28-30, 2015, Han
zhou, China
6541