In 1990, Ott, Grebogi and Yorke presented the OGY method to control chaos [1] . After their pioneering work, chaotic control has become a focus in nonlinear problems and a lot of work has been done in the field [2] - [4] . Nowadays, many methods have been proposed to control chaos [5] [6] . Generally speaking, there are two kinds of control ways: feedback control and nonfeedback control. Feedback methods [7] - [11] are used to stabilize the unstable periodic orbit of chaotic systems by feeding back their states. Nonfeedback methods [11] -[14] are adopted to change chaotic behaviors by applying perturbations to some parameters or variables. In the paper, we use feedback method to control the dynamic behavior of Liu system. By adjusting feedback coefficient, Liu system can be stabilized at equilibrium point or limit cycle around its equilibrium. Lyapunov exponents spectrum and bifurcation diagram are adopted to analyze the dynamic behavior of the controlled system. Numerical simulations and circuit experiments show the effectiveness of this method.

Liu system [15] is described as

When, , , and, system (1) exhibits a chaotic behavior. Its attractor is shown in Figure 1. The projections of system (1)’s attractor are shown in Figure 2. System (1) has three equilibriums:, ,.

Considering the voltage restraint of practical electronic components, let

Then in the new coordinate system, system (1) will be described as

System (3) can be seemed as a reduced Liu system and the equilibriums are, ,. The circuit realization of Equation (3) is shown in Figure 3.

In Figure 3, the voltages of C₁, C₂, C₃ are used as variables. The relevant function can be described as

When we choose R₁ = 10 kΩ, R₂ = 20 kΩ, R₃ = 10 kΩ, R₄ = 100 kΩ, R₅ = 10 kΩ, R₆ = 20 kΩ, R₇ = 10 kΩ, R₈ = 80 kΩ, R₉ = 40 kΩ, R₁₀ = 100 kΩ, R₁₁ = 10 kΩ, R₁₂ = 10 kΩ, R₁₃ = 100 kΩ, R₁₄ = 10 kΩ, R₁₅ = 10 kΩ, R₁₆ = 40 kΩ, C₁ = C₂ = C₃ = 1 μF, the circuit system (4) is equivalent to system (3). The supplies of all active devices are ±18 V and the initial voltages of C₁, C₂, C₃ are random, we obtain the experiment observations of system (4) as Figure 4 (with Multisim 7.0).

Comparing Figure 2 and Figure 4, we can know that a reduced Liu system has been realized by circuit experiment. Next, we will add a feedback controller to this circuit to control chaos. Various attractors will be demonstrated not only by numerical simulations but also by the circuit experiment observations.

Suppose we want to stabilize Liu system at equilibrium and the limit cycle surrounding respectively. For convenience, choose x as feedback variable, this feedback can be added to any of the three functions of Liu system. Applying the controller to the second function, then the controlled Liu system is described as

where is feedback coefficient.

In order to study the relation between and system (5)’s behavior, we make the bifurcation diagram of system (5) with in Figure 5. stands for the largest x in every unsteady period or steady period. When system (5) is stabilized at fixed point or system (5)’s behavior is periodic, has only one value or numbered values with certain; When system (5)’s behavior is chaotic, will have numberless values with certain. According to the method presented by Ramasubramanian et al. [16] , we obtain the Lyapunov exponents spectrum of system (5) with in Figure 6. When the largest Lyapunov exponent, system (5)’s behavior is chaotic; When, system (5)’s behavior is periodic; When, system (5) is stabilized at fixed point. From Figure 5 and Figure 6, we have the following conclusions: when, system (5) is chaotic (except a very narrow zone near, where system (5) may be periodic); when, system (5) is periodic; when, system (5) is stabilized at.

We obtain the above conclusions by numerical calculation. In fact, the accurate range for to stabilize system (5) at can be obtained by theoretical calculation. Substitute the values of parameters and equilibriums, we obtain the Jacobian matrix of system (5) at:

Suppose as eigenvalue, then the characteristic equation of Equation (6) is

where, ,.

According to Routh-Hurwitz criterion, when, , and, the real parts of all the eigenvalues of Equation (6) are negative, then system (5) will be stabilized at. It’s easy to obtain the solution.

As for the reduced Liu system, it’s easy to obtain the relevant controlled system:

Obviously the above conclusions about are still available to system (8). Next we will use system (8) for numerical simulations and circuit experiments. The circuit diagram for system (8) is shown in Figure 7. The relevant function can be described as

When we choose, , all other cognominal electronic components are defined as the above, then circuit system (9) is equivalent to system (8) and we can adjust to obtain proper feedback coefficient.

Substitute the value of, we have

Choose typical value for numerical simulations, the simulation results of system (8) are shown in Figure 8. Choose, we can obtain the equivalent circuit system (9), the experiment results are shown in Figure 9 (with Multisim 7.0).

From Figure 8 and Figure 9, we can know that system (8) is equivalent to system (9) in troth. When (), the reduced Liu system is chaotic; When (), the reduced Liu system is periodic; When (), the reduced Liu system’s behavior is a limit cycle around the equilibrium; When (), the reduced Liu system is stabilized at lastly. These results accord with the conclusions in Section 3.

We study the chaotic control of Liu system with feedback method in the paper. Liu chaotic system and its control are realized not only by numerical simulations but also by circuit experiments. Computer simulation and circuit experiment results show the effectiveness of our method. Moreover, our control needs only one communication channel, which is significant in practical applications.

The work was supported by Doctor Specific Funds of Dalian University.

MingjunWang, (2016) Controlling Liu Chaotic System with Feedback Method and Its Circuit Realization. International Journal of Modern Nonlinear Theory and Application,05,40-47. doi: 10.4236/ijmnta.2016.51004