Hysteresis Modeling And Adaptive Control Method Of Magnetic Shape Memory Alloy Actuator

In recent years,people are more and more focused on the development of intelligent materials.Magnetic Shape Memory Alloy(MSMA)as a new type of intelligent material which is considered to be one of the ideal materials for manufacturing micro/nano actuators because of its characteristics such as high frequency,large stroke and high energy density.The actuators made of magnetically controlled shape memory alloy have incomparable advantages over traditional actuators.Therefore,the related research of MSMA materials has received extensive attention.However,due to the complex hysteresis and nonlinear characteristics between input and output of MSMA actuators,their applications in micro-positioning and micro-driving are severely hindered.Therefore,it is necessary to design a reasonable control method to eliminate the influence of hysteresis on the control precision of the system.In order to solve this problem,this paper will mainly study from two aspects.One is about how to establish an accurate mathematical model to describe the hysteresis of the MSMA actuators.Secondly,based on the establishment of the hysteresis model,an effective control strategy is designed to reduce the influence of hysteresis on the positioning accuracy of the MSMA actuators.Firstly,the Volterra static model and its principle are introduced.The Volterra hysteresis model is used to describe the static hysteresis in the MSMA actuators.In order to describe the rate-dependent hysteresis of the MSMA actuators more accurately,a series radial basis function neural network method was proposed to establish the Volterra dynamic hysteresis model based on the Volterra static sub-model.Fading memory least square method and gradient descent method were used to identify the established dynamic Volterra model.The experimental results show that the established Volterra dynamic hysteresis model can well describe the rate-dependent hysteresis of the MSMA actuators.Secondly,in order to eliminate the influence of hysteresis on the positioning accuracy of the MSMA actuators,a feedforward inverse compensation controller was designed to compensate the hysteresis of the MSMA actuators based on the established Volterra static hysteresis model.In order to eliminate the influence of rate-dependent hysteresis on the positioning accuracy of MSMA actuators and improve the control accuracy,a neural network adaptive feedback control strategy based on Volterra model series radial basis function neural network was proposed based on Volterra feed-forward inverse compensation to reduce the influence of rate-dependent hysteresis on the system accuracy.Experimental results show that compared with the feedforward control method,the neural network adaptive control method based on Volterra hysteresis compensation has higher control precision.Finally,the stability of the proposed neural network adaptive controller is proved by Lyapunov stability theory.Finally,in this paper,a radial basis function neural network direct robust adaptive control method considering saturation term is adopted to improve the control accuracy of the MSMA actuators,which is more in line with the requirements of the actual system and ensures the control performance.It is proved by the Lyapunov function that the system is stable under the designed radial basis function neural network direct robust adaptive controller considering the saturation term.Experimental results show that,compared with the neural network adaptive control method proposed in the previous chapter,the direct robust adaptive control method considering the saturation term has better control performance,which proves the effectiveness of the proposed method.