Research Of Hysteresis Nonlinearity Modeling And Sliding Mode Control Method For Magnetically Controlled Shape Memory Actuators

As a typical smart material, magnetically controlled shape memory alloy(MSMA) is generally considered as one of the ideal materials which is very suitable for fabricating precise positioning actuators for its advantages of high frequency response, large strain and high energy density. The actuators based on MSMA possess some distinct advantages over traditional actuators, but the hysteresis nonlinearty of MSMA actuators serving as a main obstacle prevents their further applications. Therefore, it is critical to design effective control methods to eliminate hysteresis effect of MSMA actuators. For the purpose of eliminating hysteresis nonlinearity of MSMA actuators, this paper concentrates on the hysteresis nonlinear modeling and the corresponding control methods. The research on the mathematical model of hysteresis nonlineary of MSMA actuators is conducted firstly in this paper. Subsequently, the effective control methods based on the proposed hysteresis nonlineary model are designed.Nonlinear auto regressive moving average exogenous(NARMAX) model is used to serve as the hysteresis model for MSMA actuators which is able to describe the complex hysteresis nonlinearity. To obtain the unidentified parameters of the NARMAX model, the radial basis function neural network(RBFNN) and the diagonal recurrent neural network(DRNN) are utilized, and the precise hysteresis model is given by using the identified parameters. The simulations of the NARMAX model identified by two neural networks are conducted to demonstrate effectiveness and feasibility of the proposed hysteresis model and the proposed identification methods.By transforming the exogenous function, the NARMAX inverse model is able to be obtained easily. Subsequently, the unidentified parameters of the NARMAX inverse model are identified by RBFNN. Based on the identified NARMAX inverse model, the feedforward controller with pre-compensation is designed which is an important component of theopen-loop control method for MSMA actuators. The inverse-model-based feedforward controller can eliminate the hysteresis effect of MSMA actuators and improve the control accuracy in some degree. However, the disturbance rejection ability of the feedforward control system has poor performance which tremendously reduce control accuracy. In order to resolve this problem, the hybrid control method is proposed in this paper which is composed of the adaptive PID feedback control method and the inverse-model-based feedforward control method. The simulation results of the hybrid control method show that the proposed hybrid control method can improve the control accuracy of MSMA actuators efficiently which demonstrates the effectiveness of the proposed hybrid control method.To make further improvement from two aspects of the control accuracy and the robustness ability for MSMA actuators, the adaptive sliding mode control(ASMC) is proposed in this paper. By designing the appropriate adaptive law and the sliding mode reaching law, ASMC is derived which can improve the disturbance rejection ability, the tracking performance and the control accuracy of MSMA actuators significantly and eliminate hysteresis effect effectively. For the purpose of further improving the tracking performance and the control accuracy of MSMA actuators, the adaptive back-stepping sliding mode control(ABSMC) is proposed in this paper. And the global stabilities of the ASMC system and ABSMC system of MSMA actuators are guaranteed by using Lyapunov stability criterion. The simulation results show that the designed sliding mode controllers are able to eliminate the negative effect of the external disturb signal on the control system of MSMA actuators, and the control system of MSMA actuators can obtain preferable control performance. But, ABSMC possesses faster convergence speed, better tracking performance and stronger robustness in comparison with ASMC, which can be concluded from the simulation results.