International Journal of Trend in Scientific Research and Development (IJTSRD)

International Open Access Journal | www.ijtsrd.com

ISSN No: 2456 - 6470 | Volume - 3 | Issue – 1 | Nov – Dec 2018

Exponential State Observer Design for a Class of

Uncertain Chaotic and Non-Chaotic
Chaotic Systems
Yeong-Jeu Sun
Professor, Department off Electrical Engineering, I-Shou
Shou University, Kaohsiung, Taiwan

ABSTRACT
In this paper, a class of uncertain chaotic and nonnon-
chaotic systems is firstly introduced and the state
observation problem of such systems is explored.
Based on the time-domain
domain approach with integra
integral and
differential equalities, an exponential state observer
for a class of uncertain nonlinear systems is
established to guarantee the global exponential
stability of the resulting error system. Besides, the
guaranteed exponential convergence rate can be
calculated correctly. Finally, numerical simulations
are presented to exhibit the feasibility and
effectiveness of the obtained results.
Key Words: Chaotic system, state observer design,
uncertain systems, exponential convergence rate
integral and differential equalities,
equalities a new state
observer for a class of uncertain nonlinear systems
will be provided to ensure the global exponential
stability of the resulting error system. In addition, the
guaranteed exponential convergence rate can be
precisely calculated. Finally, some numerical
simulations will be given to demonstrate the
effectiveness of the main result.
ult.
This paper is organized as follows. The problem
formulation and main results are presented in Section
2. Several numerical simulations are given in Section
3 to illustrate the main result. Finally, conclusion
remarks are drawn in Section 4. Throughout this
paper, ℜn denotes the n-dimensional
dimensional real space,
x := x ⋅ x denotes the Euclidean norm of the column
T

1. INTRODUCTION
vector x, and a denotes the
he absolute value of a real
In recent years, various chaotic systems have been
widely explored by scholars;; see, for example, [1 [1-8] number a.
and the references therein. Frequently, chaotic signals
are often the main cause of system instability and 2. PROBLEM ROBLEM FORMULATION AND MAIN MAI
violent oscillations. Moreover, chaos often occurs in RESULTS
various engineering systems and applied physics
physics; for In this paper, we explore the following uncertain
instance, ecological systems, secure communicati
communication, nonlinear systems:
and system identification. Based on practical x&1 (t ) = ∆f1 (x1 (t ), x2 (t ), x3 (t )), (1a)
considerations, not all state variables of most systems x& 2 (t ) = ax2 (t ) + f 2 (x1 (t ), x3 (t )), (1b)
can be measured. Furthermore, the design of the state x&3 (t ) = f 3 (x3 (t )) + f 4 (x1 (t )), (1c)
estimator is an important work when the sensor fails. y (t ) = bx3 (t ), ∀ t ≥ 0, (1d)
Undoubtedly, the state observer design of systems
with both chaos and uncertainties tends to be more
[x1 (0) x2 (0) x3 (0)] = [x10 x20 x30 ] , (1e)
T T

difficult than those without chaos and uncertainties.

For the foregoing reasons, the observer design of where x(t ) := [x1 (t ) x2 (t ) x3 (t )]T ∈ℜ3 is the state vector,
y (t ) ∈ℜ is the system output, [x10 x20 x30 ] is the
T
uncertain chaotic systems is really significant and
essential. initial value, ∆f1 is uncertain function, and a, b ∈ ℜ
indicate the parameters of the systems, system with a < 0 and
In this paper, the observability
lity problem for a class of b ≠ 0 . Besides, in n order to ensure the existence and
uncertain nonlinear chaotic or non-chaotic
chaotic systems is uniqueness of the solution, we assume that all the
investigated. By using the time-domain
domain approach with

@ IJTSRD | Available Online @ www.ijtsrd.com | Volume – 3 | Issue – 1 | Nov-Dec

Dec 2018 Page: 1158
 International Journal of Trend in Scientific Research and Development (IJTSRD) ISSN: 2456-6470
2456
functions of ∆f1 and f i (⋅), ∀ i ∈ {2,3,4} are smooth and From (1c), it can be readily obtained that
the inverse function of f 4 exists. The chaotic sprott E f 4 ( x1 (t )) = x&3 (t ) − f 3 (x3 (t )) .
system is a special case of systems (1) with ∆f1 = x2 x3 ,
f 2 = x12 , f 3 = 1 , f 4 = −4 x1 , and a = −1 . It is a well
well-known It results that
x1 (t ) = f 4−1 ( x&3 (t ) − f 3 (x3 (t )))
fact that since states are not always available for direct
1 1  (5)
measurement, particularly in the event vent of sensor = f 4−1  y& (t ) − f 3  y (t ) 
failures, states must be estimated. The he aim of this b b 
paper is to search a novel state observer for the
uncertain nonlinear systems (1) such that the global Thus, one has
exponential stability of the resulting error systems can e1 (t ) = x1 (t ) − z1 (t )
be guaranteed. 1 1 
= f 4−1  y& (t ) − f 3  y (t ) 
b b 
Before presenting
nting the main result, the state
1 1 
reconstructibility is provided as follows. − f 4−1  y& (t ) − f 3  y(t ) 
b b 
Definition 1 = 0, ∀ t ≥ 0, (6)
The uncertain nonlinear systems (1) are exponentially
state reconstructible if there exist a state observer in view of (5) and (2a). In addition,
addition from (1b), (1c),
f (z , z&, y ) = 0 and positive numbers κ and α such that (4), and (6), it is easy to see that
e&2 (t ) = x&2 (t ) − z&2 (t )
e(t ) := x(t ) − z (t ) ≤ κ exp(− α t ), ∀ t ≥ 0 ,
= [ax2 (t ) + f 2 (x1 (t ), x3 (t ))]
− [az2 (t ) + f 2 (z1 (t ), z3 (t ))]
where z (t ) represents the reconstructed state of
= [ax2 (t ) + f 2 (x1 (t ), x3 (t ))]
systems (1). In this case, the positive number α is
− [az2 (t ) + f 2 (x1 (t ), x3 (t ))]
called the exponential convergence rate.
= a[x2 (t ) − z 2 (t )]
Now, we are in a position to present the main results = ae2 (t ), ∀ t ≥ 0.
for the exponential state observer of uncertain systems
(1). It follows that
e&2 (t ) − ae2 (t ) = 0, ∀ t ≥ 0.
Theorem 1.
The uncertain systems (1) are exponentially state Multiplying with exp(− at ) yields
reconstructible. Moreover,, a suitable state observer is e&2 (t ) ⋅ exp(− at ) − ae2 (t ) ⋅ exp(− at ) = 0, ∀ t ≥ 0.
given by
1 1 
z1 (t ) = f 4−1  y& (t ) − f 3  y (t ) , (2a) Hence, it can be readily obtained that
tha
b b  d [e2 (t ) ⋅ exp(− at )]
= 0, ∀ t ≥ 0.
z&2 (t ) = az 2 (t ) + f 2 ( z1 (t ), z3 (t )), (2b) dt

z3 (t ) = y (t ), ∀ t ≥ 0 .
1
(2c) Integrating the bounds from 0 and t , it results
b
d [e (t ) ⋅ exp(− at )]
t t
In this case, the guaranteed exponentia
exponential convergence
∫0 2 dt dt = ∫ 0 dt = 0, ∀ t ≥ 0 .
rate is given by α := −a . 0

This implies that

Proof. For brevity, let us define the observer error e2 (t ) = e2 (0) ⋅ exp(at ), ∀ t ≥ 0 . (7)
ei (t ) := xi (t ) − zi (t ), ∀ i ∈ {1,2,3} and t ≥ 0 . (3)
Thus, from (4), (6), and (7),, we have
From (1d) and (2c), one has e(t ) = e12 (t ) + e22 (t ) + e32 (t )
e3 (t ) = x3 (t ) − z3 (t )
= e2 (0) ⋅ exp(at ), ∀ t ≥ 0.
= y (t ) − y (t )
1 1
Consquently, we conclude that the t system (2) is a
b b
suitable state observer with the guaranted exponential
= 0, ∀ t ≥ 0. (4)
convergence rate α = −a . This completes the proof. □

@ IJTSRD | Available Online @ www.ijtsrd.com | Volume – 3 | Issue – 1 | Nov-Dec

Dec 2018 Page: 1159
 International Journal of Trend in Scientific Research and Development (IJTSRD) ISSN: 2456-6470
2456
3. NUMERICAL SIMULATIONS chaotic Lur’e systems,” ,” Nonlinear Analysis:
Consider the uncertain nonlinear system systems (1) with Hybrid Systems, vol. 29,, pp. 303-321, 2018.
∆f1 = ∆c ⋅ x2 x3 , f 2 = x1 , f 3 = 1 ,
2
(8a) 3. J. Kim and H. Ju, “Hausdorff
Hausdorff dimension of the
f 4 = −4 x1 , a = −1 , b = 2 , −1 ≤ ∆c ≤ 1 . (8b) sets of Li-Yorke
Yorke pairs for some chaotic dynamical
systems including A-coupled
coupled expanding systems,”
system
By Theorem 1,, we conclude that the uncertain Chaos, Solitons & Fractals,
Fractals vol. 109, pp. 246-251,
systems (1) with (8) is exponentially state 2018.
reconstructible by the state observer
4. F.H. Hsiao, “Chaotic
Chaotic synchronization
−1
z1 (t ) = y& (t ) + ,
1
(9a) cryptosystems combined with RSA encryption
8 4
algorithm,” Fuzzy Sets and Systems,
Systems vol. 342, pp.
z&2 (t ) = − z 2 (t ) + z12 (t ), (9b) 109-137, 2018.
z3 (t ) = y (t ), ∀ t ≥ 0.
1
(9c) 5. J. Park and P. Park, “ H ∞ sampled-state feedback
2
control for synchronization of chaotic Lur’e Lur
The typical state trajectory of the uncertain systems systems with time delays,” ,” Journal of the Franklin
(1) with (8) is depicted in Figure 1. Furthermore
Furthermore, the Institute, vol. 355, pp. 8005-8026,
8005 2018.
time response of error states is depicted in Fig
Figure 2. 6. A.K. Shikha, “Chaotic
Chaotic analysis and combination-
combination
From the foregoing simulations results, it is seen that combination synchronization of a novel
the uncertain systems (1) with (8) are exponentially hyperchaotic
aotic system without any equilibria,”
equilibria
state reconstructible by the state observer oof (9), with Chinese Journal of Physics,
Physics vol. 56, pp. 238-251,
the guaranted exponential convergence rate α = 1 . 2018.
4. CONCLUSION 7. R. Zhang, D. Zeng, S. Zhong, K. K Shi, and J. Cui,
In this paper, a class of uncertain chaotic and non
non- “New
New approach on designing stochastic sampled-
sampled
chaotic systems has been introduced and the state data controller for exponential synchronization of
observation problem of such system systems has been chaotic Lur’e systems,” ,” Nonlinear Analysis:
studied. Based on the time-domain
domain approach with Hybrid Systems, vol. 29,, pp. 303-321, 2018.
integral and differential equalities,, a novel state 8. J. Kim and H. Ju, “Hausdorff
Hausdorff dimension of the
observer for a class of uncertain nonlinear systems has sets of Li-Yorke
Yorke pairs for some chaotic dynamical
been constructed to ensure the global exponential systems including A-coupled
coupled expanding systems,”
systems
stability of the resulting error system. Moreover, the Chaos, Solitons & Fractals, vol. 109, pp. 246-251,
guaranteed exponential convergence rate can be 2018.
precisely calculated. Finally, numerical simulations
have been presented to exhibit the effectiveness and 9. D. Astolfi, L. Marconi, L.L Praly, and A.R. Teel,
feasibility of the obtained results. “Low-power peaking-freefree high-gain
high observers,”
Automatica, vol. 98, pp. 169
69-179, 2018.
ACKNOWLEDGEMENT 10. S. Xiao, Y. Zhang, and B. Zhang, “Event-
The author thanks the Ministry of Science and triggered network-based
based state observer design of
Technology of Republic of China for supporting this positive systems,” Information Sciences,
Sciences vol. 469,
work under grants MOST 106-2221 2221-E-214-007, pp. 30-43, 2018.
MOST 106-2813-C-214-025-E,, and MOST 107-2221-
11. D. Bernal, “State observers in the design of
E-214-030. Besides, the author is grateful to Chair
eigenstructures for enhanced sensitivity,”
Professor Jer-Guang Hsieh for the useful comments.
Mechanical Systems and Signal Processing,
Processing vol.
110, pp. 122-129, 2018.
REFERENCES
1. S. Xiao and Y. Zhao, “A A large class of chaotic 12. H. Hammouri, F.S. Ahmed,
Ahmed and S. Othman,
sensing matrices for compressed sensing
sensing,” Signal “Observer
Observer design based on immersion technics
Processing, vol. 149, pp. 193-203,, 201
2018. and canonical form,” Systems & Control Letters,
Letters
vol. 114, pp. 19-26, 2018.
2. R. Zhang, D. Zeng, S. Zhong, K. Shi, and J. Cui,
“New
New approach on designing stochastic sampled
sampled- 13. S. Li, H. Wang, A. Aitouche, and N. Christov,
Christov
data controller for exponential
ntial synchronization of “Sliding
Sliding mode observer design for fault and

@ IJTSRD | Available Online @ www.ijtsrd.com | Volume – 3 | Issue – 1 | Nov-Dec

Dec 2018 Page: 1160
 International Journal of Trend in Scientific Research and Development (IJTSRD) ISSN: 2456-6470
2456
disturbance estimation using
ing Takagi
Takagi-Sugeno,” 14. A. Sassi, H.S. Ali, M. Zasadzinski, and K.
European Journal of Control, vol. 44
44, pp. 114-122, Abderrahim, “Adaptive observer design for a class
2018. of descriptor nonlinear systems,” European
Journal of Control, vol. 44, pp. 90-102,
90 2018.
8

2
x1(t); x2(t); x3(t)

-2

-4 x1: the Blue Curve

x2: the Green Curve
-6 x3: the Red Curve

-8
0 50 100 150
t (sec)

Figure 1: Typical state trajectory es of the uncertain nonlinear systems (1) with (8).

1.5
e1(t); e2(t); e3(t)

1 e3

0.5

e1=e2=0
0

-0.5
0 5 10 15
t (sec)

Figure
ure 2: The time response of error states.

@ IJTSRD | Available Online @ www.ijtsrd.com | Volume – 3 | Issue – 1 | Nov-Dec

Dec 2018 Page: 1161