† Corresponding author. E-mail:
In the conventional chaos synchronization methods, the time at which two chaotic systems are synchronized, is usually unknown and depends on initial conditions. In this work based on Lyapunov stability theory a sliding mode controller with time-varying switching surfaces is proposed to achieve chaos synchronization at a pre-specified time for the first time. The proposed controller is able to synchronize chaotic systems precisely at any time when we want. Moreover, by choosing the time-varying switching surfaces in a way that the reaching phase is eliminated, the synchronization becomes robust to uncertainties and exogenous disturbances. Simulation results are presented to show the effectiveness of the proposed method of stabilizing and synchronizing chaotic systems with complete robustness to uncertainty and disturbances exactly at a pre-specified time.
In recent decades, chaos synchronization has attracted attention of many researchers. The aim of the synchronization is that a slave system follows completely the master system, despite existence of uncertainties and exogenous disturbances in both systems and difference in their initial conditions. So far various control methods have been utilized for synchronizing chaotic systems, e.g. back stepping method,[1,2] linear feedback control,[3,4] nonlinear feedback control,[5] adaptive control,[6,7] adaptive feedback,[8] fuzzy logic control,[9] LMI-based control,[10] observer-based control,[11,12] active control,[13–16] Lyapunov stability theory,[17–20] and so on. In Refs. [21] and [22], active controls have been used under the assumption of known parameters and Lyapunov stability theory has been used for establishing parameter update laws when parameters are uncertain or fully unknown.
The sliding mode control (SMC) is the other control method that has been also utilized effectively for chaos synchronization. Although robustness to uncertainty and disturbances in sliding phase is one of the most important advantages of SMC, neither uncertainty nor disturbances have been considered in Refs. [23]–[29]. Uncertainty and disturbances have been considered in Refs. [30]–[34] but these studies are not completely robust due to the existence of the reaching phase. However, this problem has been resolved in Refs. [35]–[38] by establishing a reaching law and in Ref. [39] by eliminating the reaching phase. All the papers, which have been so far mentioned, have asymptotical stability. As a result, in these papers, synchronization will be realized at infinity analytically.
Finite-time stability for chaos synchronization, stabilization and control has been achieved in some studies such as Refs. [40]–[48] with control methods other than the sliding mode one. Sun et al. in Ref. [49] have proposed an adaptive combination sliding mode controller based on using the nonsingular terminal sliding mode control technique. In Ref. [50], finite-time synchronization has been accomplished by introducing a new nonsingular terminal sliding surface. Similarly, Aghababa and Aghababa in Ref. [51] have managed to synchronize two chaotic systems finite-timely by introducing a novel nonsingular terminal sliding surface. In some papers such as Refs. [52]–[54], finite-time stability is available only for either the reaching phase or sliding phase. Sun et al. in Ref. [55] not only have succeeded in finite-time synchronization but also have made synchronization completely robust by eliminating the reaching phase. Nevertheless, in all these papers, only an upper bound of synchronization time has been determined and it depends directly on the initial error. Hence, the greater the differences, the more time the controller must take to synchronize chaotic systems. For reducing this time and subsequently the energy consumption, initial conditions of two chaotic systems must be made nearer to each other at the initial time as far as we can.
Effati et al. in Ref. [56] could accomplish the chaos synchronization at a pre-specified time by an optimal control design. In their work, the proposed controller is only applicable to chaotic Chen systems, uncertainty and disturbances must not exist and more important than the others is that we have to solve a two-point boundary value problem (TPBVP).
Motivated by the above discussion, the aim of this paper is to resolve the above problems by using the sliding mode technique. To the best of our knowledge, this is the first work in synchronizing and stabilizing a large class of chaotic system in the presence of uncertainty and disturbances precisely at a time when that time has been set in advance. Indeed, what is intended to be carried out is that two chaotic systems with uncertainty and disturbances are synchronized exactly at any time when we want, is needed or asked for free of worrying about how much two chaotic systems are far from each other at the initial time. We will show that it can be achieved remarkably without requiring the solution of a time-consuming TPBVP by employing time-varying switching surfaces. Moreover, these surfaces are determined in such a way that the reaching phase is eliminated such that the synchronization process is completely robust to uncertainty and disturbances.
It is assumed that master and slave systems are modeled as Eqs. (
Synchronization will be realized when the difference between two chaotic systems, which is called the synchronization error henceforth, becomes zero. Accordingly, the synchronization error is defined as follows:
Now if equations (
The sliding mode control procedure is composed of two stages. The first stage is to determine a sliding surface. The next one is to establish a control law, insuring that the state trajectory converges to the sliding surface and eventually reaches it and stays on it forever. Accordingly, in the following subsection, the general form of switching surfaces is described. In the next one, the control law is established by Lyapunov stability theory. Finally, an example of terminal functions whose general properties are specified in the following subsection is presented.
The switching surfaces are considered as follows:
As the synchronization error follows the terminal functions completely and it will be shown later that the derivatives of the terminal functions exist in control signal formula, they must be differentiable. In general, two distinct manners can be adopted for determining the terminal functions. We can determine them dependently or independently.
As mentioned earlier, if there is a controller to force the synchronization error to follow the terminal functions completely, then robust synchronization will be realized exactly at a pre-specified time. Mathematically, the switching surfaces (
It is necessary for the terminal functions to be differentiable and satisfy two conditions (
In this section, effectiveness and applicability of the proposed method is illustrated by some examples. The first-order differential equations are numerically approximated by the Euler method. Numerical simulations are implemented in MATLAB software. For having as accurate simulation results as possible, the step size 0.0001 is taken.
Two non-identical Lu[61] and Liu[62] chaotic systems are considered as master and slave systems and their dynamics are respectively as follows:
Firstly, simulation is performed without considering uncertainty nor disturbances. Therefore, ki = 0 for i = 1,2,3. The synchronization time has been set at T = 2 s. The controller is applied to the slave system (
Now, to show the ability of the proposed method of synchronizing two chaotic systems at any pre-specified time, the synchronization time is set at T = 1 s. Figure
Here, it is assumed that uncertainty and disturbances are present in both master and slave systems as follows:
A very simple comparison between Fig.
Synchronizing chaotic system exactly at a pre-specified time in the presence of uncertainty and disturbances is achieved for a large class of chaotic systems using time-varying sliding mode control without requiring to solve a TPBVP for the first time. The complete robustness of the synchronization process to uncertainty and disturbances is also demonstrated. This approach could be applied to other types of synchronization such as projective, function projective or modified function projective synchronization. Central limitation of this approach is the need for exact knowledge of the values of initial conditions. Further work can be carried out about relaxing the assumption of exact knowledge and introducing more suitable terminal functions that do not lower practical applications of this approach.
1 | |
2 | |
3 | |
4 | |
5 | |
6 | |
7 | |
8 | |
9 | |
10 | |
11 | |
12 | |
13 | |
14 | |
15 | |
16 | |
17 | |
18 | |
19 | |
20 | |
21 | |
22 | |
23 | |
24 | |
25 | |
26 | |
27 | |
28 | |
29 | |
30 | |
31 | |
32 | |
33 | |
34 | |
35 | |
36 | |
37 | |
38 | |
39 | |
40 | |
41 | |
42 | |
43 | |
44 | |
45 | |
46 | |
47 | |
48 | |
49 | |
50 | |
51 | |
52 | |
53 | |
54 | |
55 | |
56 | |
57 | |
58 | |
59 | |
60 | |
61 | |
62 | |
63 | |
64 | |
65 |