New Delay-Dependent Robust Stability Criteria for Uncertain
Neutral Systems with Mixed Delays
Yuebing Li, Shaosheng Zhou, Biao Zhang
School of Automation, Hangzhou Dianzi University, Hangzhou 310018, P. R. China
E-mail: sszhou65@163.com
Abstract: This paper is concerned with the problem of the delay-dependent robust stability of neutral systems with mixed delays
and time-varying structured uncertainty. The uncertainty modeled in the systems is of the linear fractional type which includes
the norm-bounded uncertainty as a special case. By constructing a new type of augmented Lyapunov-Krasovskii functional and
using some integral inequalities, a new stability criterion is obtained in terms of a linear matrix inequality. Numerical examples
are given to demonstrate the effectiveness and less conservativeness of the proposed methods.
Key Words: Neutral system; Delay-dependent stability; Linear matrix inequality
1 Introduction
During the last decade, a great deal effort has been done
to the stability analysis of time-delay systems since time de-
lay is always one of the sources of instability and poor per-
formance. The time delay often occurs in many dynamic
systems such as chemical processes, biological systems, net-
worked control systems[1,2]. Therefore, the stability analy-
sis of time-delay systems has attracted considerable atten-
tion[3]. The existing stability criteria for time-delay systems
can be diveded into two categories: delay-independent[4,5]
and delay dependent[6-12]. The delay-independent ones
which are independent of delay size are generally more con-
servative than delay-independent ones, especially when a
delay is small. On the other hand, the delay-dependent
criteria has focused on identical delays in neutral and dis-
crete terms[9,11]. He et al.[13] proposed a delay-dependent
stability criteria using the free-weighting matrices method,
which contained mixed delays and time-varying structured
uncertainties. Also, this method was used to derive a new
delay-dependent stability criterion and a stabilizing method
for uncertain neutral systems [14]. An augmented Lyapunov
functional [15] was introduced to present some new stabil-
ity criteria of neutral time-delay systems.Recently, a novel
augmented Lyapunov functional has been proposed in [16]
to derive the robust stability of uncertain neutral systems. It
seems that the augmented Lyapunov functional approach is
very effective to reduce the conservativeness. In this paper,
a new method is proposed to deal with the problem of the
delay-dependent stability of neutral systems. First, by con-
structing a new augmented Lyapunov-Krasovskii functional
which contains triple-integral terms and using some integral
inequalities, a criterion for a nominal neutral system is ob-
tained in terms of the linear matrix inequality. Then, the
criterion derived is extended to a neutral system with time-
varying uncertainties. Finally, numerical examples are given
to illustrate the effectiveness of the proposed method.
2 Notation and preliminaries
Consider the following time-varying structured uncertain
neutral system Σ:
This work is supported in part by National Natural Science Founda-
tion (NNSF) of China under Grant 61273093, but also the National Natural
Science Foundation of Province of Zhejiang under Grant LZ12F03001.
˙x(t) − C ˙x(t − τ
2
)=(A +ΔA(t))x(t)
+(B +ΔB(t))x(t − τ
1
),t>0
x(t)=φ(t),t ∈ [−τ, 0] (1)
where x(t) ∈ R
n
is the state vector; τ
1
,τ
2
>0 are constant
delays; τ := max(τ
1
,τ
2
), A, B, C ∈ R
n
are constant matri-
ces; and the spectrum radius of the matrix C, ρ(C), satisfies
ρ(C)<1. The initial condition φ(t) denotes a continuous
vector-valued initial function of t∈[−τ, 0]. The time-varying
structured uncertainties are assumed to take the following
linear fractional form:
[ΔA(t)ΔB(t)] = DΨ[E
a
E
b
], (2)
Ψ=[I − F (t)G]
−1
F (t), (3)
I − GG
T
> 0 (4)
where D, G, E
a
,E
b
are known real constant matrices with
appropriate dimensions; and F(t) is an uncertain time-
varying matrix satisfying
F (t)F
T
(t) ≤ I. (5)
First, the nominal system Σ
0
of Σ is defines to be
˙x(t) − C ˙x(t − τ
2
)=Ax(t)+Bx(t − τ
1
),t>0
x(t)=φ(t),t ∈ [−τ, 0] (6)
To obtain the main results,the following lemma is intro-
duced, which plays an important role in the development of
the main results.
Lemma 1 [17]. For any constant matrix Q=Q
T
> 0 and a
scalar τ>0 such that the following integrations are well
defined, then
−
t
t−τ
T
(ξ)Q(ξ)dξ
≤−
1
τ
(
t
t−τ
(ξ)dξ)
T
Q(
t
t−τ
(ξ)dξ),
−
0
−τ
t
t+θ
T
(ξ)Q(ξ)dξdθ
≤−
2
τ
2
(
0
−τ
t
t+θ
(ξ)dξdθ)
T
Q(
0
−τ
t
t+θ
(ξ)dξdθ
Proceedings of the 33rd Chinese Control Conference
Jul
28-30, 2014, Nan
in
, China
6114