Hindawi Publishing Corporation

Abstract and Applied Analysis

Volume 2013, Article ID 670878, 11 pages
http://dx.doi.org/10.1155/2013/670878

Research Article
Chaos Synchronization Based on Unknown Input Proportional
Multiple-Integral Fuzzy Observer

T. Youssef,1,2 M. Chadli,1 H. R. Karimi,3 and M. Zelmat2

1
Laboratory of Modeling, Information & Systems (MIS), University of Picardie Jules Verne (UPJV), 33 rue Saint Leu,
80039 Amiens Cedex 1, France
2
Laboratory of Automatic Applied (LAA), M’hamed Bougara University of Boumerdès (UMBB), 35000 Boumerdès, Algeria
3
Department of Engineering, Faculty of Engineering and Science, University of Agder, 4898 Grimstad, Norway

Correspondence should be addressed to M. Chadli; [email protected] and H. R. Karimi; [email protected]

Received 1 April 2013; Revised 17 June 2013; Accepted 18 June 2013

Academic Editor: Zidong Wang

Copyright © 2013 T. Youssef et al. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This paper presents an unknown input Proportional Multiple-Integral Observer (PIO) for synchronization of chaotic systems
based on Takagi-Sugeno (TS) fuzzy chaotic models subject to unmeasurable decision variables and unknown input. In a secure
communication configuration, this unknown input is regarded as a message encoded in the chaotic system and recovered by the
proposed PIO. Both states and outputs of the fuzzy chaotic models are subject to polynomial unknown input with kth derivative
zero. Using Lyapunov stability theory, sufficient design conditions for synchronization are proposed. The PIO gains matrices are
obtained by resolving linear matrix inequalities (LMIs) constraints. Simulation results show through two TS fuzzy chaotic models
the validity of the proposed method.

1. Introduction systems, system identification, and biological systems; see, for

example, [5–10].
It is well known that the chaotic systems have a complex Recently, a particular attention has been paid to the syn-
dynamical behavior and their fundamental characteristic chronization and control problems for dynamical networks
is the chaos. The chaotic systems are highly sensitive to due to their extensive application in fields of science and
parameters variation and to initial conditions because the engineering (see, e.g., [11–13]). In [11] the authors propose
chaos is a source of oscillation and instability. Moreover, in a novel concept of bounded H ∞ synchronization and state
the long term the behavior of the chaotic systems becomes
estimation to handle the time-varying nature of an array
difficult to predict which can lead systems to instability
stochastic complex network in discrete-time domain over a
and undesirable performance [1]. On the other hand, the
finite horizon. The synchronization problem in [12] is con-
chaotic systems constitute a good platform to investigate
sidered for a new class of continuous-time neural networks
the nonstandard control problems including synchronization
and stabilization. of neutral type with parameters, discrete-time delays, and
unbounded distributed time delays being all dependent on
Since Pecora and Carroll [2, 3] have introduced in 1990
the concept of chaotic synchronization between two chaotic the Markovian jumping mode. The sampled-data synchro-
dynamical systems based on the Lorenz’s chaotic system nization control scheme in [13] is studied for a class of dynam-
[4], the synchronization and control of chaotic systems ical networks with stochastic sampling. In the formwork of
attract more and more attention from various disciplines. state estimation, fault detection, and filtering for a class of
A great deal of chaos applications have been developed in nonlinear systems with sensor networks, we can mention
engineering fields such as secure communication, physical the works of [14–17]. Indeed, in [14] the finite-horizon
2 Abstract and Applied Analysis

distributed H ∞ state estimator design scheme is proposed input observers for chaotic synchronization and robust to
for a class of discrete time-varying nonlinear systems with channel noise. The unknown input observer in [37] is
both stochastic parameters and stochastic nonlinearities. The presented with unknown constant disturbance of parameters
problem of designing the distributed H ∞ filters in [15] is and unknown input to be recovered as messages in the
considered for a class of polynomial nonlinear stochastic master-slave configuration. The robust adaptive high-gain
systems which are represented in a state-dependent linear- fuzzy observer is designed in [32] for chaotic systems where
like form. The distributed finite-horizon filter is proposed in their parameters are assumed unknown and their states
[16] for a class of time-varying systems subject to randomly unavailable. The author deals with, in [38], the unknown
varying nonlinearities over lossy sensor networks that involve input observer design for fuzzy systems with application to
both the quantization errors and successive packet dropouts. chaotic system reconstruction within both domains, contin-
The robust fuzzy fault detection filter in [17] is designed uous and discrete time, where the sufficient conditions have
for a class of uncertain discrete-time Takagi-Sugeno fuzzy been derived in terms of linear matrix inequality constraints
systems with successive packet dropouts which involve both by using Lyapunov stability theory. Then, in [38, 39] the pole
the stochastic multiple time-varying discrete delays and the assignment in a LMI region is considered in order to improve
infinite distributed delays. the observer performance.
Different methods and techniques for chaos synchroniza-
In the context of the unmeasurable decision variable,
tion and control have been investigated including impul-
the synchronization problem for chaotic systems character-
sive control [18], feedback control [19], adaptive control
ized by TS fuzzy models was not addressed by the above
[20], lag synchronization [21], sliding mode control [22],
works. Our main contribution in this paper is to develop
and fuzzy control [23]. Particularly, in the chaotic secure
a synchronization procedure which takes into account the
communication problems, some messages can be masked
unmeasurable decision variables. The effects due to unmea-
efficiently and securely [24] by chaotic signals since the
surable decision variable and the unknown input on the
chaos has the characteristic of broadband like a noise and is
overall synchronization system (chaotic system and observer)
consequently difficult to predict. The message in the secure
are compensated with additional parameter. In addition, in
communications which is recovered by the response system
the present study the estimation of the unknown input as a
should synchronize with the drive system [25]. In framework
message to be reconstructed within a secure communication
of a secure communication, many approaches have been
concept is considered.
addressed such as chaos modulation [26], chaos shift key
[27], and chaos masking [28]. Among different methods In the framework of a secure communication, the design
of synchronization and control for chaotic systems, the TS of unknown input PIO is addressed in this work for two
fuzzy systems have received much attention from various chaotic systems, Lorenz’ system and Rossler’s system, charac-
researches fields since the pioneering work of Takagi and terized by chaotic TS fuzzy models. These models are subject
Sugeno (TS) [29]. Indeed, TS fuzzy model can approximate to unmeasurable decision variable and polynomial unknown
a highly nonlinear analytical relation of chaotic system by input where its 𝑘th derivative is zero. The proposed PIO
fuzzy IF-THEN rules where the implications describe local estimates both the states of the considered chaotic systems
dynamics as linear models. Then, the nonlinear behavior of and the polynomial unknown input. This latter is considered
the chaotic system can exactly be obtained as an aggregation as a message to encode by the chaotic system and then to
of local linear models with nonlinear activation functions. reconstruct it by the PIO. Furthermore, the integral action
TS fuzzy models are widely used as a tool of analysis and included in the observer structure contributes to reduce the
design of synchronization and control schemes because of results conservatism due to quadratic function. Indeed, this
the mathematical analysis simplicity of their simple structure parameter allows introducing an additional degree of free-
with local dynamics, for example, [30–32]. dom to be determined. By utilizing Lyapunov stability theory,
Since because of, many practical control problems the sufficient conditions are derived to design the polynomial
states are partially or fully unavailable, the state observer unknown input PIO. Then, the PIO gains parameters are
methods can be used to estimate the measurements of resolved in terms of linear matrix inequalities constraints.
unavailable or failed sensors. For this reason, it is important Moreover, when the decision variables are measurable we
to design the observers for state estimation. In relation to that also discuss this particular case in our work. Finally, we
state estimation observers there are many works to that deal present simulation results to illustrate the effectiveness of the
with stability analysis and stabilization of TS fuzzy models by proposed approach of synchronization and reconstruction.
applying Lyapunov theory and derive stability conditions in The rest of this paper is organized as follows. In Section 2,
terms linear matrix inequalities constraints [33]; of among the considered unknown input TS fuzzy model structure is
this works we can mention the results developed in [34]; described. This unknown input affects both the dynamics
When the decision variable is chosen as unmeasurable or of the TS fuzzy model and the output signal. The unknown
unavailable state in activation functions, for example, in [35], input PIO is designed in Section 3. In Section 4, simulation
the robust observer is designed for unknown inputs TS fuzzy results for two chaotic systems, Lorenz’ system and Rossler’s
models. Recently, in secure communication field the design system, are given. An unknown input, assumed as a message
problem of unknown input observer has been investigated to be received in a secure communication configuration,
in [36–39]. The authors propose in [36] a new secured is perfectly reconstructed. Finally, a conclusion and further
transmission scheme based on smooth adaptive unknown works end this paper.
Abstract and Applied Analysis 3

𝑗
2. Unknown Input TS Fuzzy Model Structure where 𝐾𝑃𝑖 ∈ 𝑅𝑛×𝑛𝑦 , 𝐾𝐼𝑖 ∈ 𝑅𝑛V ×𝑛𝑦 , and 𝐾𝐼𝑖 ∈ 𝑅𝑛V ×𝑛𝑦 correspond
to proportional and integral gains, respectively. Due the effect
The TS fuzzy model subject to unknown input is considered of the unmeasurable decision variables, the variables 𝑧𝑥 (𝑡),
as follow: 𝑧V (𝑡), and 𝑧V𝑗 (𝑡) are introduced in the PIO.
𝑟
𝑥̇ (𝑡) = ∑𝜇𝑖 (𝑥) (𝐴 𝑖 𝑥 (𝑡) + 𝐵𝑖 𝑢 (𝑡) + 𝑑𝑖 + 𝐹𝑖 V (𝑡)) , According to Hypothesis 1, TS fuzzy model (1) and
𝑖=1 (1) unknown input PIO (4) can be rewritten, respectively, as
𝑦 (𝑡) = 𝐶𝑥 (𝑡) + 𝐹V (𝑡) , follow:

where 𝑥(𝑡) ∈ 𝑅 represents the state vector, 𝑢(𝑡) ∈ 𝑅𝑛𝑢

𝑛 𝑟

corresponds to known input vector, V(𝑡) ∈ 𝑅𝑛V shows the 𝑥𝑎̇ (𝑡) = ∑𝜇𝑖 (𝑥) (𝐴𝑖 𝑥𝑎 (𝑡) + 𝐵𝑖 𝑢 (𝑡) + 𝑑𝑖 ) ,
𝑖=1 (5)
unknown input, and 𝑦(𝑡) ∈ 𝑅𝑛𝑦 is the output vector. 𝐴 𝑖 ∈
𝑅𝑛×𝑛 are the state matrices, 𝐵𝑖 ∈ 𝑅𝑛×𝑛𝑢 are the input matrices, 𝑦 (𝑡) = 𝐶𝑥𝑎 (𝑡) ,
𝑑𝑖 ∈ 𝑅𝑛 is a vector system dependent, 𝐹𝑖 ∈ 𝑅𝑛×𝑛V and 𝐹 ∈
𝑅𝑛𝑦 ×𝑛V are the unknown input matrices, and 𝐶 ∈ 𝑅𝑛𝑦 ×𝑛 is the 𝑟

output matrix. The activation functions 𝜇𝑖 (𝑥) depend on the ̂̇ 𝑎 (𝑡) = ∑𝜇𝑖 (𝑥)
𝑥 ̂ (𝐴𝑖 𝑥̂𝑎 (𝑡) + 𝐵𝑖 𝑢 (𝑡) + 𝑑𝑖 + 𝐾𝑖 (𝑦 (𝑡) − 𝑦̂ (𝑡)))
𝑖=1
state 𝑥(𝑡) of the system and satisfy the following conditions:
𝑟
+ 𝑧 (𝑡) ,
∑𝜇𝑖 (𝑥) = 1, ∀𝑡 ≥ 0,
𝑖=1 (2)
𝑦̂ (𝑡) = 𝐶𝑥̂𝑎 (𝑡) ,
0 ≤ 𝜇𝑖 (𝑥) ≤ 1, ∀𝑖 ∈ {1, . . . , 𝑟} , (6)
where 𝑟 is the number of local models.
where
Hypothesis 1. Unknown input V(𝑡) has a polynomial form of
𝑘 − 1 degree in time whose 𝑘th derivative is equal to zero.
𝑥 (𝑡) 𝑥̂ (𝑡)
Let the following notations be introduced: [ V (𝑡) ] [ ̂V (𝑡) ]
V̇ (𝑡) = V1 (𝑡) , [ ] [ ]
𝑥𝑎 (𝑡) = [ ]
[ V1 (𝑡) ] , 𝑥̂𝑎 (𝑡) = [ ]
[ ̂V1 (𝑡) ] ,
[ ⋅⋅⋅ ] [ ⋅⋅⋅ ]
V̇1 (𝑡) = V2 (𝑡) ,
[V𝑘−1 (𝑡)] [ ̂V𝑘−1 (𝑡) ]
.. (3) (7a)
. 𝑧𝑥 (𝑡)
[ 𝑧V (𝑡) ]
V̇𝑘−1 (𝑡) = V𝑘 (𝑡) , [ ]
𝑧 (𝑡) = [
[ 𝑧V1 (𝑡) ]
]
[ ⋅⋅⋅ ]
V𝑘 (𝑡) = 0.
[ 𝑧V𝑘−1 (𝑡) ]
Note that the polynomial form allows considering a wide
variety of unknown inputs. with

3. Unknown Input Proportional 𝑒𝑎 (𝑡) = 𝑥𝑎 (𝑡) − 𝑥̂𝑎 (𝑡) , 𝑒𝑎𝑦 = 𝑦 (𝑡) − 𝑦̂ (𝑡) , (7b)
Multiple-Integral Observer Design
𝐴𝑖 𝐹𝑖 0 0 ⋅⋅⋅ 0
The unknown input PIO is considered as follow: [0 0 𝐼𝑛V 0 ⋅⋅⋅ 0]
[ ]
𝑟 𝐴𝑖 = [
[0 0 0 𝐼𝑛V ⋅⋅⋅ 0] ],
[⋅ ⋅ ⋅ ⋅⋅⋅ ⋅⋅⋅ ⋅⋅⋅ ⋅⋅⋅ 𝐼𝑛V ]
̂̇ (𝑡) = ∑𝜇𝑖 (𝑥)
𝑥 ̂ (𝐴 𝑖 𝑥̂ (𝑡) + 𝐵𝑖 𝑢 (𝑡) + 𝑑𝑖 + 𝐹𝑖 ̂V (𝑡)
𝑖=1 [0 0 0 0 0 0]
+ 𝐾𝑃𝑖 (𝑦 (𝑡) − 𝑦̂ (𝑡))) + 𝑧𝑥 (𝑡) , 𝐵𝑖 𝑑𝑖
[ 0 ] [ 0 ]
𝑦̂ (𝑡) = 𝐶𝑥̂ (𝑡) + 𝐹̂V (𝑡) , [ ] [ ]
𝐵𝑖 = [
[ 0 ],
] 𝑑𝑖 = [
[ 0 ],
] (7c)
𝑟 [ ⋅⋅⋅ ] [ ⋅⋅⋅ ]
̂V̇ (𝑡) = ∑𝜇𝑖 (𝑥)
̂ 𝐾𝐼𝑖 (𝑦 (𝑡) − 𝑦̂ (𝑡)) + ̂V1 (𝑡) + 𝑧V (𝑡) ,
𝑖=1
[ 0 ] [ 0 ]
𝑟
𝐾𝑃𝑖
̂V̇𝑗 (𝑡) = ∑𝜇𝑖 (𝑥)
𝑗
̂ 𝐾𝐼𝑖 (𝑦 (𝑡) − 𝑦̂ (𝑡)) + ̂V𝑗+1 (𝑡) [ 𝐾𝐼𝑖 ]
𝑖=1 [ 1 ]
𝐾𝑖 = [
[ 𝐾𝐼𝑖
],
] 𝐶 = [𝐶 𝐹 0 ⋅ ⋅ ⋅ 0 ] ,
+ 𝑧V𝑗 (𝑡) , for 𝑗: 1, . . . , 𝑘 − 1 [ ⋅⋅⋅ ]
𝑘−1
if 𝑗 = 𝑘 − 1, ̂V𝑗+1 (𝑡) = 0, [ 𝐾𝐼𝑖 ]
(4) where 𝐼𝑛V is an identity matrix.
4 Abstract and Applied Analysis

3.1. Unmeasurable Decision Varaiables. The dynamic error Proof. The proposed quadratic function of Lyapunov is
𝑒𝑎 (𝑡) of state estimation is given by 𝑉(𝑡) = 𝑒𝑎𝑇 (𝑡)𝑃𝑒𝑎 (𝑡) with 𝑃 = 𝑃𝑇 > 0. The conditions ((11a),
(11b), and (11c)) guarantee the asymptotic stability of the
𝑟
dynamic error of state estimation (8). The proof is partially
𝑒𝑎̇ (𝑡) = ∑𝜇𝑖 (𝑥)
̂ A𝑖 𝑒𝑎 (𝑡) + Δ𝐴𝑥𝑎 (𝑡) + Δ𝐵𝑢 (𝑡) + Δ𝑑 − 𝑧 (𝑡) ,
given in the appendix and for more details see [40].
𝑖=1
(8)
3.2. Measurable Decision Variables Case. The design of
where unknown input PIO with measurable decision variables
represents the particular case of our study developed in this
𝑟 𝑟
section. In this condition the unknown input PIO is as follow:
A𝑖 = 𝐴𝑖 − 𝐾𝑖 𝐶, Δ𝐴 = ∑𝜇𝑖 𝐴𝑖 , Δ𝐵 = ∑𝜇𝑖 𝐵𝑖 ,
𝑖=1 𝑖=1 𝑟
(9) ̂̇ (𝑡) = ∑𝜇𝑖 (𝑥) (𝐴 𝑖 𝑥̂ (𝑡) + 𝐵𝑖 𝑢 (𝑡) + 𝑑𝑖 + 𝐹𝑖 ̂V (𝑡)
𝑥
𝑟
𝑖=1
Δ𝑑 = ∑𝜇𝑖 𝑑𝑖 , ̂ .
𝜇𝑖 = 𝜇𝑖 (𝑥) − 𝜇𝑖 (𝑥)
𝑖=1 +𝐾𝑃𝑖 (𝑦 (𝑡) − 𝑦̂ (𝑡))) ,
Remark 1. Since the activation functions satisfy the convex 𝑦̂ (𝑡) = 𝐶𝑥̂ (𝑡) + 𝐹̂V (𝑡) ,
sum property, we can write −1 < 𝜇𝑖 < 1, and the
𝑟
variables matrices Δ𝐴, Δ𝐵, Δ𝑑 are bounded and the following
̂V̇ (𝑡) = ∑𝜇𝑖 (𝑥) 𝐾𝐼𝑖 (𝑦 (𝑡) − 𝑦̂ (𝑡)) + ̂V1 (𝑡) , (12)
conditions hold: 𝑖=1
𝑟
󵄩󵄩 󵄩󵄩 󵄩󵄩 󵄩󵄩 𝑟
󵄩󵄩Δ𝐴󵄩󵄩 ≤ 𝛿1 , 𝛿1 = ∑𝛿1𝑖 , 󵄩󵄩Δ𝐵󵄩󵄩 ≤ 𝛿2 , ̂V̇𝑗 (𝑡) = ∑𝜇𝑖 (𝑥) 𝐾𝐼𝑖 (𝑦 (𝑡) − 𝑦̂ (𝑡)) + ̂V𝑗+1 (𝑡) ,
𝑗
󵄩 󵄩 󵄩 󵄩
𝑖=1 𝑖=1
(10)
𝑟 𝑟
󵄩󵄩 󵄩󵄩 for 𝑗 : 1, . . . , 𝑘 − 1
𝛿2 = ∑𝛿2𝑖 , 󵄩󵄩Δ𝑑󵄩󵄩 ≤ 𝛿3 , 𝛿3 = ∑𝛿3𝑖
󵄩 󵄩
𝑖=1 𝑖=1
if 𝑗 = 𝑘 − 1, ̂V𝑗+1 (𝑡) = 0,
with 𝛿1𝑖 > 0, 𝛿2𝑖 > 0 and 𝛿3𝑖 > 0, are the Euclidian norms of where all variables and matrices are defined in relations ((7a),
𝐴𝑖 , 𝐵𝑖 , and 𝑑𝑖 , respectively. (7b), and (7c)), and the activation functions 𝜇𝑖 (𝑥) depend on
the measurable states.
Theorem 2. The dynamic error (8) is asymptotically stable if The dynamics of the augmented state estimation error
there exist a common positive definite matrix 𝑃 = 𝑃𝑇 , matrices 𝑒𝑎 (𝑡) between the TS fuzzy model (1) and PIO (12) becomes
𝑀𝑖 and the positive scalars 𝛼 and 𝛼0 , for all 𝑖 ∈ {1, . . . , 𝑟}, such
𝑟
that:
𝑒𝑎̇ (𝑡) = ∑𝜇𝑖 (𝑥) (𝐴𝑖 − 𝐾𝑖 𝐶) 𝑒𝑎 (𝑡) . (13)
𝑖=1
𝑇 𝑇 𝑇
2 Theorem 3. The dynamic error (13) is asymptotically stable if
[𝑃𝐴𝑖 + 𝐴𝑖 𝑃 − 𝑀𝑖 𝐶 − 𝐶 𝑀𝑖 + 𝛼0 𝛿1 𝐼 𝑃 ] < 0, (11a)
𝑃 −𝛼𝐼 there exist a symmetric matrix 𝑄 > 0 and matrices 𝑁𝑖 such
that the following conditions hold, for all 𝑖 ∈ {1, . . . , 𝑟}:
where the matrices and parameters 𝐴𝑖 , 𝐶, 𝛿1 are defined in (7c),
and (10), respectively. 𝑇 𝑇 𝑇
𝑄𝐴𝑖 + 𝐴𝑖 𝑄 − 𝑁𝑖 𝐶 − 𝐶 𝑁𝑖 < 0. (14a)
The parameters of unknown input PIO (4) are obtained by
The parameters of the unknown input PIO (12) are given by:
𝐾𝑖 = 𝑃−1 𝑀𝑖 , (11b)
𝐾𝑖 = 𝑄−1 𝑁𝑖 . (14b)
󵄨 󵄨
𝑧 = 0, if 󵄨󵄨󵄨󵄨𝑒𝑎𝑦 󵄨󵄨󵄨󵄨 < 𝜀,
Proof. Consider the Lyapunov quadratic function 𝑉(𝑡) =
𝑥̂𝑇 𝑥̂ 𝑇
𝑇
𝑢 𝑢 −1 𝑇 𝑒𝑎𝑇 (𝑡)𝑄𝑒𝑎 (𝑡), where 𝑄 = 𝑄𝑇 > 0. The time derivative of 𝑉(𝑡)
𝑧= 𝜎1 𝛿21 𝑇𝑎 𝑎 𝑃−1 𝐶 𝑒𝑎𝑦 + 𝜎2 𝛿22 𝑇 𝑒
𝑃 𝐶 𝑒𝑎𝑦 allows writing
2𝑒𝑎𝑦 𝑒𝑎𝑦 2𝑒𝑎𝑦 𝑎𝑦
(11c)
𝑟
𝑇
1 𝑇 󵄨 󵄨 𝑉̇ = ∑𝜇𝑖 (𝑥)
̂ 𝑒𝑎𝑇 ((𝐴𝑖 − 𝐾𝑖 𝐶) 𝑄 + 𝑄(𝐴𝑖 − 𝐾𝑖 𝐶)𝑖 ) 𝑒𝑎 . (15)
+ 𝜎3 𝛿32 𝑇 𝑒
𝑃−1 𝐶 𝑒𝑎𝑦 , if 󵄨󵄨󵄨󵄨𝑒𝑎𝑦 󵄨󵄨󵄨󵄨 ≥ 𝜀 𝑖=1
2𝑒𝑎𝑦 𝑎𝑦
̇ < 0 is satisfied if
The stability condition 𝑉(𝑡)
with variables 𝑥̂𝑎 , 𝑧, 𝑒𝑎𝑦 and the parameters 𝛿2 , 𝛿3 are
𝑇
described in (7a), (7b), (10), respectively, and 𝜎1 = (𝛼0 /𝜆), (𝐴𝑖 − 𝐾𝑖 𝐶) 𝑄 + 𝑄(𝐴𝑖 − 𝐾𝑖 𝐶)𝑖 < 0; (16)
𝜎2 = ((𝛼𝛼0 𝜆 3 )/(𝛼(𝛼0 + 𝜆 3 (1 + 𝜆)) − 𝛼0 𝜆 3 )), 𝜎3 = 𝜆 3 where
𝜆, 𝜆 3 are positive scalars arbitrarily fixed and 𝜀 is a very small with variables change 𝑁𝑖 = 𝑄𝐾𝑖 ; we obtain the linear matrix
positive threshold. inequalities ((14a) and (14b)). The proof is completed.
Abstract and Applied Analysis 5

2.5

30
20 2
10
x2 (t)

0
−10 1.5
−20
−30
60 1
40 20
10
20 0
x3 ( 0 −10
t) −20 x 1(t) 0.5
0 5 10 15 20 25 30
x2 (x1 , x3 ) t (s)

Figure 1: Chaotic behavior of Lorenz fuzzy system.


̂

The unknown input PIO gains 𝐾𝑖 = 𝑄−1 𝑁𝑖 are deter- Figure 2: The unknown input and its estimated.
mined by resolving these constraints. In the following section,
a simulation example is given through two chaotic systems in Table 1
order to validate this proposed approach.
𝜆 = 2 × 103 𝛼 = 5.505 × 104 𝛼0 = 0.001

4. Simulation Examples 𝑖 1 2
−11.471 12.445 −11.057 11.926
Two chaotic systems are characterized by TS fuzzy mod- [ ] [ ]
𝐾𝑝𝑖 [−22.084 10.838] [ 02.100 11.650 ]
els with unmeasurable decision variables and subjected to [ ] [ ]
unknown input. In a secure communication concept, the [−94.350 96.029] [ 97.909 −96.239]
unknown input is assumed as a message to be recovered 𝐾𝐼𝑖 [15.123 03.344] [21.570 −03.137]
in the PIO after being encoded in the chaotic system by
𝐾𝐼𝑖1 [83.889 −02.802] [79.394 01.612]
means of public transmission canal. These chaotic systems are
used to show the good simultaneous reconstruction of states
and message by the proposed unknown input PIO. The first
nonlinear model is the Lorenz’s system [41], and the second The Lorenz chaotic attractor is given in Figure 1.
is the fourth Rossler’s system [21]. The TS fuzzy model of the Lorenz chaotic system (17) is
2
4.1. Lorenz Chaotic System. The Lorenz chaotic system [38] is 𝑥̇ (𝑡) = ∑𝜇𝑖 (𝑥) (𝐴 𝑖 𝑥 (𝑡) + 𝐸𝑖 V (𝑡)) ,
represented by following the dynamic equations: 𝑖=1 (19)
𝑥1̇ = −10𝑥1 + 10𝑥2 , 𝑦 (𝑡) = 𝐶𝑥 (𝑡) + 𝐸V (𝑡) ,
𝑥2̇ = 28𝑥1 − 𝑥2 − 𝑥1 𝑥3 , 0 1 1
(17) with 𝐵1 = 𝐵2 = [ 0 ], 𝐸1 = [ 1 ], 𝐸2 = [ 0.5 ], 𝐸 = [ 11 ], 𝐶 =
0 1 1
8 [ 00 01 11 ].
𝑥3̇ = 𝑥1 𝑥2 − 𝑥3 .
3 The unknown input V(𝑡) is assumed as a message to be
recovered by the unknown input PIO (4). This observer plays
4.1.1. TS Fuzzy Model. The Lorenz’s system can be exactly the role of decoder and the chaotic system the encoder within
represented by TS fuzzy model with the decision variable a secure communication configuration.
𝑥1 (𝑡) ∈ [−30, 30] as follow [42]:
̇ = 𝐴 1 𝑥(𝑡),
Rule 1: 𝑥1 (𝑡) is 𝜇1 (𝑥1 (𝑡)), THEN 𝑥(𝑡) 4.1.2. Unknown Input PIO. The unknown input PIO gains are
determined by resolving the LMIs constraints ((11a), (11b),
̇ = 𝐴 2 𝑥(𝑡),
Rule 2: 𝑥2 (𝑡) is 𝜇2 (𝑥1 (𝑡)), THEN 𝑥(𝑡) and (11c)) of Theorem 2:
where 𝑥(𝑡) = [𝑥1 (𝑡), 𝑥2 (𝑡), 𝑥3 (𝑡)], 𝜇1 (𝑥1 (𝑡)) = (30 + 𝑥1 (𝑡))/60, 𝑇 𝑇
𝑇
𝜇2 (𝑥1 (𝑡)) = (30 − 𝑥1 (𝑡))/60, and 𝐾𝑖 = [𝐾𝑃𝑖 𝐾𝐼𝑖𝑇 𝐾𝐼𝑖1𝑇 ] , 𝑍 = [𝑍𝑥𝑇 𝑍V𝑇 𝑍V1
𝑇
] . (20)

−10 10 0 −10 10 0 The unknown input PIO gains are given by Table 1.
[ 28 −1 −30] [ 28 −1 30 ] The unknown input is assumed as a message to be
𝐴1 = [ , 𝐴2 = [ . (18)
−8 ] −8 ] encoded by chaotic system with the second derivative being
0 30 0 −30
[ 3 ] [ 3 ] zero, shown in Figure 2. The simulation results are obtained
6 Abstract and Applied Analysis

20 30

20
10
10

0 0

−10
−10
−20

−20 −30
0 5 10 15 20 25 30 0 5 10 15 20 25 30
t (s) t (s)

x1 x2
̂1
x ̂2
x

(a) (b)
50

40

30

20

10

0
0 5 10 15 20 25 30
t (s)

x3
̂3
x

(c)

Figure 3: (a) The state 𝑥1 (𝑡) and its estimated 𝑥̂1 (𝑡). (b) The state 𝑥2 (𝑡) and its estimated 𝑥̂2 (𝑡). (c) The state 𝑥3 (𝑡) and its estimated 𝑥̂3 (𝑡).

with the initial conditions 𝑥0 = [1 1 1] 𝑥̂0 = [0 0 0] and 4.2. Fourth Rossler Chaotic System. The fourth Rossler
with 𝜀 = 10−3 . chaotic system [21] is represented by the following dynamic
The unknown input and the estimated one are given in equations:
Figure 2. Excepted around the time origin, we obtained a
good reconstruction of the unknown input. Figures 3(a), 3(b), 𝑥1̇ = −𝑥2 − 𝑥3 ,
and 3(c) show the states 𝑥1 (𝑡), 𝑥2 (𝑡), and 𝑥3 (𝑡) and their
estimated 𝑥̂1 (𝑡), 𝑥̂2 (𝑡), and 𝑥̂3 (𝑡), respectively. 𝑥2̇ = 𝑥1 + 0,254𝑥2 + 𝑥4 ,
The dynamic errors of the states estimation are given (21)
𝑥3̇ = 𝑥1 𝑥3 + 3,
in Figure 4. The obtained simulation results show the good
reconstruction of the states and the unknown input. 𝑥4̇ = −0,5𝑥3 + 0,05𝑥4 .
Remark 4. Note that applying Theorem 3 instead of
Theorem 2 for this example leads to bad estimation. The 4.2.1. TS Fuzzy Model. The fourth Rossler’s system can be
simulation results for this example, carried out with the exactly described by TS fuzzy model with the decision
same initial conditions: 𝑥0 = [1 1 1] and 𝑥̂0 = [0 0 0], variable 𝑥1 (𝑡) ∈ [−80, 20] as follow [42]:
are shown in Figure 5. Indeed, we see clearly that the best
estimation (Figure 6) is given by Theorem 2 (dashed line) ̇ = 𝐴 1 𝑥(𝑡) + 𝑑1 ,
Rule 1: 𝑥1 (𝑡) is 𝜇1 (𝑥1 (𝑡)), THEN 𝑥(𝑡)
which takes into account the estimation of decision variables. ̇ = 𝐴 2 𝑥(𝑡) + 𝑑2 ,
Rule 2: 𝑥2 (𝑡) is 𝜇2 (𝑥1 (𝑡)), THEN 𝑥(𝑡)
Abstract and Applied Analysis 7

1.5 Table 2

𝜆 = 3 × 103 𝛼 = 6.243 × 105 𝛼0 = 6.467 × 10−5

1
𝑖 1 2
0.5 011.018 001.805 020.387 −026.055
[ ] [ ]
[ −000.343
[ 001.344 ]
]
[ 004.339
[ −013.805 ]
]
𝐾𝑝𝑖 [ ] [ ]
0 [ −078.779 109.443 ] [ 166.122 −242.987 ]
[ ] [ ]
[ 011.208 005.138 ] [ 028.076 −045.031 ]
−0.5
𝐾𝐼𝑖 [052.544 −006.911] [030.739 058.089]

−1 𝐾𝐼𝑖1 [097.452 −008.136] [066.736 083.503]

0 5 10 15 20 25 30
𝐾𝐼𝑖2 [091.450 −008.807] [058.600 089.165]
t (s)
𝐾𝐼𝑖3 [034.284 −001.174] [027.830 018.118]
Error x1
Error x2
Error x3
where 𝑥(𝑡) = [𝑥1 (𝑡), 𝑥2 (𝑡), 𝑥3 (𝑡), 𝑥4 (𝑡)], 𝜇1 (𝑥1 (𝑡)) = (80 +
Figure 4: The errors between states and their estimated. 𝑥1 (𝑡))/100, 𝜇2 (𝑥1 (𝑡)) = (20 − 𝑥1 (𝑡))/100, and
0 −1 −1 0
[1 0.25 0 1 ]
𝐴1 = [
[0 0
],
1 20 0 ]
[0 0 −0.5 0.05]

0.5 0 −1 −1 0
[1 0.25 0 1 ]
𝐴2 = [
[0 0 −80 0 ] ,
] (22)
0 [0 0 −0.5 0.05]
0
−0.5 [0]
[
𝑑1 = 𝑑2 = [ ] .
3]
[0]
−1
0 5 10 15 20 25 30 The fourth Rossler chaotic attractor is given in Figure 7.
t (s) The TS fuzzy model of the fourth Rossler system (21) is
2
Error  with Theorem 1
Error  with Theorem 2 𝑥̇ (𝑡) = ∑𝜇𝑖 (𝑥) (𝐴 𝑖 𝑥 (𝑡) + 𝐸𝑖 V (𝑡) + 𝑑𝑖 ) ,
𝑖=1 (23)
Figure 5: The errors between unknown input and its estimated.
𝑦 (𝑡) = 𝐶𝑥 (𝑡) + 𝐸V (𝑡)
0 1 1
with 𝐵1 = 𝐵2 = [ 00 ], 𝐸1 = [ 11 ], 𝐸2 = [ 0.5 1
1 ], 𝐸 = [ 1 ],
0.03 0 1 1
𝐶 = [ 00 10 01 10 ].
0.02 The unknown input V(𝑡) is assumed as a message to be
0.01 encoded by the TS fuzzy model (23).
0
4.2.2. Unknown Input PIO. The unknown PIO gains are
−0.01 obtained by resolving the LMIs constraints ((11a), (11b), and
(11c)) of Theorem 3:
−0.02
𝑇 𝑇
−0.03 𝐾𝑖 = [𝐾𝑃𝑖 𝐾𝐼𝑖𝑇 𝐾𝐼𝑖1𝑇 𝐾𝐼𝑖2𝑇 𝐾𝐼𝑖3𝑇 ] ,
−0.04
(24)
𝑇 𝑇
𝑍= [𝑍𝑥𝑇 𝑍V𝑇 𝑇
𝑍V1 𝑇
𝑍V2 𝑍V3 ] .
−0.05
17 17.5 18 18.5 19 19.5 20
The unknown PIO gains are given by Table 2.
Figure 6: The zoom on the errors between unknown input and its The unknown input is assumed as a message to be
estimated. encoded by chaotic system. The best results are obtained
8 Abstract and Applied Analysis

40 40

30 30
x4 (t)

x4 (t)
20 20

10 10
40
20 150
0 20 20
0 100 0
−20 −20 −20
x2 ( −40 −40 50 −40
t) −60 −60 x3 ( −60
−80 x 1(t) t) 0 −80 x 1(t)

x4 (x1 , x2 ) x4 (x1 , x3 )

(a) (b)

40 200

150
30
x4 (t)

x3 (t)

100
20
50

10 0
40
20 20
100 40
20 0 0
x3 ( 0 −20 −20
t) −20 x2 (t −40
−40 ) −40 −60
0 x 2(t)
−60 −60 −80 x 1(t)

x4 (x2 , x3 ) x3 (x1 , x2 )

(c) (d)

Figure 7: (a) Chaotic behavior 𝑥4 (𝑥1 (𝑡), 𝑥2 (𝑡)) of fourth Rossler fuzzy system. (b) Chaotic behavior 𝑥4 (𝑥1 (𝑡), 𝑥3 (𝑡)) of fourth Rossler fuzzy
system. (c) Chaotic behavior 𝑥4 (𝑥2 (𝑡), 𝑥3 (𝑡)) of fourth Rossler fuzzy system. (d) Chaotic behavior 𝑥3 (𝑥1 (𝑡), 𝑥2 (𝑡)) of fourth Rossler fuzzy
system.

2.5 with the fourth derivative being zero, shown in Figure 8. The
simulation results are obtained with the initial conditions
𝑥0 = [0 0 0 30] and 𝑥̂0 = [1 1 1 29] and with 𝜀 = 10−3 .
2
The unknown input and the estimated one are given in
Figure 8. Excepted around the time origin, we got a good
1.5
reconstruction of the unknown input. Figures 9(a), 9(b), 9(c)
and 9(d) represent the states 𝑥1 (𝑡), 𝑥2 (𝑡), 𝑥3 (𝑡), and 𝑥4 (𝑡) and
their estimated 𝑥̂1 (𝑡), 𝑥̂2 (𝑡), 𝑥̂3 (𝑡), and 𝑥̂4 (𝑡), respectively.
1 The dynamic errors of the states estimation are repre-
sented in Figure 10. The obtained simulation results show the
good estimation of the states and the unknown input.
0.5
0 5 10 15 20 25 30
t (s)
5. Conclusion

̂ In this paper, we have addressed the synchronization and
reconstruction problem for chaotic systems. The TS fuzzy
Figure 8: The unknown input and its estimated. models subjected to unmeasurable decision variables and
Abstract and Applied Analysis 9

50 150

0 100

−50 50

−100
0

−150
−50
−200
−100

0 5 10 15 20 25 30 0 5 10 15 20 25 30
t (s) t (s)

x1 x2
̂1
x ̂2
x

(a) (b)
300
140

250 120

200 100

150 80

60
100
40
50
20
0
0 5 10 15 20 25 30 0
0 5 10 15 20 25 30
t (s)
t (s)

x3 x4
̂3
x ̂4
x

(c) (d)
Figure 9: (a) The state 𝑥1 (𝑡) and its estimated 𝑥̂1 (𝑡). (b) The state 𝑥2 (𝑡) and its estimated 𝑥̂2 (𝑡). (c) The state 𝑥3 (𝑡) and its estimated 𝑥̂3 (𝑡). (d)
The state 𝑥4 (𝑡) and its estimated 𝑥̂4 (𝑡).

1.5 unknown inputs are employed to exactly describe the behav-

ior of two chaotic systems, Lorenz’s system and Rossler’s
1 system. Based on Lyapunov theory and LMI formulation, an
unknown input proportional integral observer to achieve the
0.5
synchronization and the unknown input reconstruction is
0
designed. To take into account a wide variety of unknown
inputs, a polynomial form with 𝑘th derivative zero is con-
−0.5 sidered. Moreover, both the measurable and unmeasurable
decision variables cases are studied. Simulation results are
−1 given to verify the effectiveness of the proposed method by
reconstructing both states and unknown inputs. In the secure
−1.5 communication field, the proposed polynomial unknown
0 5 10 15 20 25 30
t (s)
input PIO with unmeasurable decision variables presents a
good synchronization technique and messages recovering.
Error x1 Error x3 Motivated by the given results, the problem of diagnosis
Error x2 Error x4 and fault tolerant control for more complex systems will be
Figure 10: The errors between states and their estimated. considered. Moreover, to reduce the conservatism due to the
10 Abstract and Applied Analysis

quadratic approach, nonquadratic Lyapunov functions will References

be introduced in our further research.
[1] G. Jiang, G. Chen, and W. K. Tang, “Stabilizing unstable
equilibria of chaotic systems from a state observer approach,”
Appendix IEEE Transactions on Circuits and Systems II, vol. 51, no. 6, pp.
281–288, 2004.
By using the quadratic Lyapunov function 𝑉(𝑡) = 𝑒𝑎𝑇 (𝑡)𝑃𝑒𝑎 (𝑡) [2] L. M. Pecora and T. L. Carroll, “Synchronization in chaotic
where 𝑃 = 𝑃𝑇 > 0 and the following lemma. systems,” Physical Review Letters, vol. 64, no. 8, pp. 821–824,
1990.
Lemma A.1. For any matrices X and Y of appropriate dimen- [3] T. L. Carroll and L. M. Pecora, “Synchronizing chaotic circuits,”
sions, the following property is satisfied: IEEE Transactions on Circuits and Systems, vol. 38, no. 4, pp.
453–456, 1991.
𝑋𝑇 𝑌 + 𝑌𝑇 𝑋 ≤ 𝜆𝑋𝑇 𝑋 + 𝜆−1 𝑌𝑇 𝑌 with 𝜆 > 0, (A.1) [4] E. N. Lorenz, “Deterministic nonperiodic flow,” Journal of the
Atmospheric Sciences, vol. 20, no. 2, pp. 130–141, 1963.
the time-derivative of 𝑉(𝑡) leads [5] J. Yang, G. Hu, and J. Xiao, “Chaos synchronization in coupled
chaotic oscillators with multiple positive lyapunov exponents,”
𝑟 Physical Review Letters, vol. 80, no. 3, pp. 496–499, 1998.
𝑇 2
𝑉̇ ≤ ∑𝜇𝑖 (𝑥)
̂ (𝑒𝑎𝑇 (A𝑖 𝑃 + 𝑃A𝑖 + 𝛼0 𝛿12 𝐼 + ∝−1 𝑃 ) 𝑒𝑎 ) [6] S. Boccaletti, C. Grebogi, Y.-C. Lai, H. Mancini, and D. Maza,
𝑖=1 “The control of chaos: theory and applications,” Physics Reports,
vol. 329, no. 3, pp. 103–197, 2000.
+ 𝜎1 𝛿12 𝑥̂𝑎𝑇𝑥̂𝑎 + 𝜎2 𝛿22 𝑢𝑇 𝑢 + 𝜎3 𝛿23 𝐼 − 2𝑒𝑎𝑇 𝑃𝑧 [7] S. Boccaletti, J. Kurths, G. Osipov, D. L. Valladares, and C.
(A.2) S. Zhou, “The synchronization of chaotic systems,” Physics
Reports, vol. 366, no. 1-2, pp. 1–101, 2002.
with [8] E. M. Shahverdiev, “Synchronization in systems with multiple
time delays,” Physical Review E, vol. 70, no. 6, Article ID 067202,
𝛼0 = 𝜆 1 (1 + 𝜆) , ∝−1 = (𝜆−1 −1 −1
1 + 𝜆2 + 𝜆3 ) , 4 pages, 2004.
[9] J. Xu, L. Min, and G. Chen, “A chaotic communication scheme
𝛼0
𝜎1 = 𝜆 1 (1 + 𝜆−1 ) = ( ), based on generalized synchronization and hash functions,”
𝜆 Chinese Physics Letters, vol. 21, no. 8, pp. 1445–1448, 2004.
(A.3) [10] J. W. Haefner, Modeling Biological Systems: Principles and
𝛼𝛼0 𝜆 3
𝜎2 = 𝜆 2 = ( ), Applications, Springer, New York, NY, USA, 2005.
𝛼 (𝛼0 + 𝜆 3 (1 + 𝜆)) − 𝛼0 𝜆 3 [11] B. Shen, Z. Wang, and X. Liu, “Bounded 𝐻∞ synchroniza-
tion and state estimation for discrete time-varying stochastic
𝜎3 = 𝜆 3 . complex networks over a finite horizon,” IEEE Transactions on
Neural Networks, vol. 22, no. 1, pp. 145–157, 2011.
And taking into account (11c) we obtain [12] Y. Liu, Z. Wang, J. Liang, and X. Liu, “Synchronization of
coupled neutral-type neural networks with jumping-mode-
2 𝑥̂𝑎𝑇 𝑥̂𝑎 −1 𝑇 dependent discrete and unbounded distributed delays,” IEEE
2𝑒𝑎𝑇 𝑃𝑧 = 2𝑒𝑎𝑇 𝑃𝜎1 𝛿1 𝑇 𝑒
𝑃 𝐶 𝑒𝑎𝑦
2𝑒𝑎𝑦 𝑎𝑦
Transactions on Cybernetics, vol. 43, no. 1, pp. 102–114, 2013.
[13] B. Shen, Z. Wang, and X. Liu, “Sampled-data synchronization
𝑢𝑇 𝑢 −1 𝑇 control of dynamical networks with stochastic sampling,” IEEE
+ 2𝑒𝑎𝑇 𝑃𝜎2 𝛿22 𝑇 𝑒
𝑃 𝐶 𝑒𝑎𝑦 Transactions on Automatic Control, vol. 57, no. 10, pp. 2644–
2𝑒𝑎𝑦 𝑎𝑦 (A.4) 2650, 2012.
1 𝑇
[14] D. Ding, Z. Wang, H. Dong, and H. Shu, “Distributed 𝐻∞
+ 2𝑒𝑎𝑇 𝑃𝜎3 𝛿32 𝑇 𝑒
𝑃−1 𝐶 𝑒𝑎𝑦 state estimation with stochastic parameters and nonlinearities
2𝑒𝑎𝑦 𝑎𝑦 through sensor networks: the finite-horizon case,” Automatica,
vol. 48, no. 8, pp. 1575–1585, 2012.
= 𝜎1 𝛿12 𝑥̂𝑎𝑇 𝑥̂𝑎 + 𝜎2 𝛿22 𝑢𝑇𝑢 + 𝜎3 𝛿23 𝐼. [15] B. Shen, Z. Wang, Y. S. Hung, and G. Chesi, “Distributed 𝐻∞
filtering for polynomial nonlinear stochastic systems in sensor
Then, the relation (A.2) becomes networks,” IEEE Transactions on Industrial Electronics, vol. 58,
no. 5, pp. 1971–1979, 2011.
𝑟
𝑇 2 [16] H. Dong, Z. Wang, and H. Gao, “Distributed filtering for a class
𝑉̇ ≤ ∑𝜇𝑖 (𝑥)
̂ 𝑒𝑎𝑇 (A𝑖 𝑃 + 𝑃A𝑖 + 𝛼0 𝛿12 𝐼 + ∝−1 𝑃 ) 𝑒𝑎 . (A.5) of time-varying systems over sensor networks with quantization
𝑖=1
errors and successive packet dropouts,” IEEE Transactions on
Signal Processing, vol. 60, no. 6, pp. 3164–3173, 2012.
̇
The condition of stability 𝑉(𝑡) < 0 (for all 𝑖 = 1, . . . , 𝑟) is
[17] H. Dong, Z. Wang, J. Lam, and H. Gao, “Fuzzy-model-based
satisfied if robust fault detection with stochastic mixed time delays and
𝑇 2 2 successive packet dropouts,” IEEE Transactions on Systems,
A𝑖 𝑃 + 𝑃A𝑖 + 𝛼0 𝛿12 𝐼 + ∝−1 𝑃 + 𝛼−1 𝑃 < 0. (A.6) Man, and Cybernetics B, vol. 42, no. 2, pp. 365–376, 2012.
[18] B. Wanga, J. Wang, and S. M. Zhong, “Impulsive synchroniza-
The Schur complement of condition (A.6) with variables tion control for chaotic systems,” Procedia Engineering, vol. 15,
given in (9) allows writing the LMI (11a). pp. 2721–2726, 2011.
Abstract and Applied Analysis 11

[19] C. Tao, C. Yang, Y. Luo, H. Xiong, and F. Hu, “Speed feedback adaptive unknown-input observers,” Communications in Non-
control of chaotic system,” Chaos, Solitons and Fractals, vol. 23, linear Science and Numerical Simulation, vol. 17, no. 9, pp. 3727–
no. 1, pp. 259–263, 2005. 3739, 2012.
[20] Y. Yu, “Adaptive synchronization of a unified chaotic system,” [37] M. Chen and W. Min, “Unknown input observer based chaotic
Chaos, Solitons and Fractals, vol. 36, no. 2, pp. 329–333, 2008. secure communication,” Physics Letters A, vol. 372, no. 10, pp.
[21] C. Li, X. Liao, and K.-W. Wong, “Lag synchronization of 1595–1600, 2008.
hyperchaos with application to secure communications,” Chaos, [38] M. Chadli and I. Zlinka, “Unknown input observer design for
Solitons and Fractals, vol. 23, no. 1, pp. 183–193, 2005. fuzzy systems with application to chaotic system reconstruc-
[22] M. S. Tavazoei and M. Haeri, “Determination of active sliding tion,” Computers and Mathematics with Applications, vol. 66, no.
mode controller parameters in synchronizing different chaotic 2, pp. 147–154, 2013.
systems,” Chaos, Solitons and Fractals, vol. 32, no. 2, pp. 583–591, [39] M. Chadli, A. Akhenak, J. Ragot, and D. Maquin, “State and
2007. unknown input estimation for discrete time multiple model,”
[23] J. G. Barajas-Ramı́rez, G. Chen, and L. S. Shieh, “Fuzzy chaos Journal of the Franklin Institute, vol. 346, no. 6, pp. 593–610,
synchronization via sampled driving signals,” International 2009.
Journal of Bifurcation and Chaos in Applied Sciences and [40] T. Youssef, M. Chadli, H. R. Karimi, and M. Zelmat, “Design of
Engineering, vol. 14, no. 8, pp. 2721–2733, 2004. unknown inputs proportional integral observers for TS fuzzy
[24] T. Liao and S. Tsai, “Adaptive synchronization of chaotic systems models,” Neurocomputing Journal, vol. 210, no. 163, p. 174, 2013.
and its application to secure communications,” Chaos, solitons [41] X. Meng, Y. Yu, G. Wen, and R. Chen, “Chaos synchronization
and fractals, vol. 11, no. 9, pp. 1387–1396, 2000. of unified chaotic system using fuzzy logic controller,” in
[25] C.-J. Cheng, “Robust synchronization of uncertain unified Proceedings of the IEEE International Conference on Fuzzy
chaotic systems subject to noise and its application to secure Systems (FUZZ ’08), pp. 544–547, June 2008.
communication,” Applied Mathematics and Computation, vol. [42] K. Tanaka, T. Ikeda, and H. O. Wang, “A unified approach
219, no. 5, pp. 2698–2712, 2012. to controlling chaos via an LMI-based fuzzy control system
[26] T. Yang, “Secure communication via chaotic parameter modu- design,” IEEE Transactions on Circuits and Systems I, vol. 45, no.
lation,” IEEE Transactions on Circuits and Systems I, vol. 43, no. 10, pp. 1021–1040, 1998.
9, pp. 817–819, 1996.
[27] H. Dedieu, M. P. Kennedy, and M. Hasler, “Chaos shift keying:
modulation and demodulation of a chaotic carrier using self-
synchronizing Chua’s circuits,” IEEE Transactions on Circuits
and Systems II, vol. 40, no. 10, pp. 634–642, 1993.
[28] K. M. Cuomo, A. V. Oppenheim, and S. H. Strogatz, “Synchro-
nization of Lorenz-based chaotic circuits with applications to
communications,” IEEE Transactions on Circuits and Systems II,
vol. 40, no. 10, pp. 626–633, 1993.
[29] T. Takagi and M. Sugeno, “Fuzzy identification of systems and
its applications to modeling and control,” IEEE Transactions on
Systems, Man and Cybernetics, vol. 15, no. 1, pp. 116–132, 1985.
[30] K. Lian, T. Chiang, C. Chiu, and P. Liu, “Synthesis of fuzzy
model-based designs to synchronization and secure commu-
nications for chaotic systems,” IEEE Transactions on Systems,
Man, and Cybernetics B, vol. 31, no. 1, pp. 66–83, 2001.
[31] J.-H. Kim, C.-W. Park, E. Kim, and M. Park, “Adaptive syn-
chronization of T-S fuzzy chaotic systems with unknown
parameters,” Chaos, Solitons and Fractals, vol. 24, no. 5, pp. 1353–
1361, 2005.
[32] C.-H. Hyun, C.-W. Park, J.-H. Kim, and M. Park, “Synchroniza-
tion and secure communication of chaotic systems via robust
adaptive high-gain fuzzy observer,” Chaos, Solitons and Fractals,
vol. 40, no. 5, pp. 2200–2209, 2009.
[33] S. Boyd, L. El Ghaoui, E. Feron, and V. Balakrishnan, Linear
Matrix Inequalities in System and Control Theory, SIAM Studies
in Applied Mathematics, Society for Industrial and Applied
Mathematics, Philadelphia, Pa, USA, 1994.
[34] K. Tanaka and H. O. Wang, Fuzzy Control System Design and
Analysis. A Linear Matrix Inequality Approach, John Wiley &
Sons, New York, NY, USA, 2001.
[35] M. Chadli and H. R. Karimi, “Robust observer design for
unknown inputs Takagi-Sugeno models,” IEEE Transactions on
Fuzzy Systems, vol. 21, no. 1, pp. 158–164, 2013.
[36] H. Dimassi, A. Loria, and S. Belghith, “A new secured trans-
mission scheme based on chaotic synchronization via smooth
Copyright of Abstract & Applied Analysis is the property of Hindawi Publishing Corporation
and its content may not be copied or emailed to multiple sites or posted to a listserv without
the copyright holder's express written permission. However, users may print, download, or
email articles for individual use.