Systems & Control Letters 78 (2015) 19–24
Contents lists available at ScienceDirect
Systems & Control Letters
journal homepage: www.elsevier.com/locate/sysconle
Convergence domain for time-varying switched systems
Jinghui Suo, Jitao Sun
∗
Department of Mathematics, Tongji University, Shanghai 200092, PR China
a r t i c l e i n f o
Article history:
Received 13 December 2013
Received in revised form
19 September 2014
Accepted 10 January 2015
Available online 18 February 2015
Keywords:
Switched systems
Convergence domain
Multiple Lyapunov functions
a b s t r a c t
In this paper, we investigate the convergence domain for time-varying switched systems. We construct
multiple Lyapunov functions and each Lyapunov function is decreasing outside a ball whose radius is time-
varying. We give a relatively accurate convergence domain for a general time-varying switched system.
Finally, a numerical example, a mass–spring–damper system, is provided to show the effectiveness of the
theoretical results.
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
Stability is an important issue for dynamical systems. However
sometimes there is no need to achieve stability in practice. In cer-
tain engineering applications, it is more natural to identify whether
the solutions of dynamical systems are bounded, see [1–9] and ref-
erences therein.
Lyapunov analysis is very useful to show boundedness of the
solutions. When we observe the derivative of a Lyapunov function
along the solution of a differential dynamical system, it is common
that this derivative is not always negative but is indeed negative
when the state trajectory satisfies some conditions. [2,3,5,10]
considered the case that the derivative of a Lyapunov function was
negative when the norm of the state was greater than or equal to
a constant and provided a theorem about ultimate boundedness of
the solution. Recently, Zhao et al. [4] investigated a more general
case that the derivative of a Lyapunov function was negative when
the norm of the state was greater than or equal to a certain function
of time and gave a convergence domain for a general class of time-
varying nonlinear systems.
A switched system is a dynamical system consisting of a family
of subsystems and a switching signal that determines the switch-
ing between them. In recent years, switched systems have gained
considerable attention in science and engineering and a number
of criteria have been derived, see [7–19] and references therein.
In view of the interaction between continuous dynamics and dis-
crete dynamics, switched systems may have very complicated be-
haviors. For instance, switching may destabilize a switched system
∗
Corresponding author. Tel.: +86 21 65983241x1307; fax: +86 21 65981985.
E-mail address: sunjt@sh163.net (J. Sun).
even if all individual subsystems are stable, whereas suitably con-
strained switching between unstable subsystems may give rise to
stability. Motivated by this, some scientists have focused on the
boundedness of the solutions of switched systems [7–10]. How-
ever, these researches do not provide the relatively accurate con-
vergence domain for a general switched system. In this paper, we
study the convergence domain for time-varying switched systems.
Since a common Lyapunov function may not exist for a switched
system, we use multiple Lyapunov function method and assume
that for each Lyapunov function, its derivative is negative when
the norm of the state is no less than a function of time. To the best
of our knowledge, there has been no result of such research. There-
fore the study in this aspect is meaningful and challenging.
This paper is organized as follows. In Section 2, we give the main
results about convergence domain for a time-varying switched
system. An application example, a mass–spring–damper system, is
provided to illustrate the theoretical results in Section 3, following
by conclusions in Section 4.
Notation: ∥ · ∥ refers to the Euclidean norm for vectors. I
represents the identity matrix. B(
¯
x, r) = {x| ∥x −
¯
x∥ ≤ r} denotes
a closed ball with radius r > 0 centered at a point
¯
x. A function β is
said to be of class K if it is a continuous, strictly increasing function
satisfying β(0) = 0.
2. Main results
We consider the following switched system
˙
x = f
σ (t)
(t, x), (1)
where x ∈ R
n
, σ : [0, ∞) → P = {1, 2, . . . , m} is a switching
signal, m is the number of subsystems. σ is a piecewise constant
http://dx.doi.org/10.1016/j.sysconle.2015.01.007
0167-6911/© 2015 Elsevier B.V. All rights reserved.