Trajectory Tracking Control Method For Magnetic Shape Memory Alloy Based Actuator

By exploiting the magnetic shape memory effect of the magnetic shape memory alloy(MSMA)material,the MSMA-based actuator can produce deformation under the action of a magnetic field,resulting in micro-and nano-scale output displacement,which renders it a high-precision micro-positioning mechanism with unparalleled superiority over conventional actuators.The main factors that give the MSMA-based actuator advantages over its counterparts are the large stroke,high displacement resolution,and great thrust,providing it with a prospective wide range of applications in precision manufacturing processing,micro-and nano-positioning technology,biomedicine and etc.,such as MSMA-based actuator driven valveless pumps,AFM positioning systems.As a kind of non-differentiable and non-one-to-one mapping complex nonlinearity,hysteresis widely exists in the smart materials,and is featured with frequency dependence,force dependence,and susceptibility to environmental factors.However,the complex dynamic hysteresis nonlinearity inherent in the MSMA material substantially inflates the difficulty of system modeling and control,which seriously impedes the application of the MSMA-based actuator.This thesis mainly studies the hysteresis modeling method of the MSMA-based actuator,designs the effective controller to realize the high-precision tracking control of MSMA-based actuator,so as to provide the theoretical and methodological basis for further application of the MSMA-based actuator in the field of the micro-positioning and micro-driving.The main research content of this thesis is as follows:Firstly,the deformation mechanism of the MSMA material is introduced,and the working principle and hysteresis characteristics of the MSMA-based actuator are described.In light of the fact that the hysteresis output of the MSMA-based actuator is sensitive to the input signals and external conditions,the effects of the input signal frequency,amplitude,ambient temperature and load on the hysteresis characteristics of the MSMA-based actuator are analyzed experimentally.Meanwhile,the impacts of the above mentioned elements on the maximum output displacement of the MSMA-based actuator are also discussed.These experimental analyses of the MSMA-based actuator serve as a foundation for establishing the mathematical model to describe the hysteresis nonlinearity of the MSMA-based actuator.The complex hysteresis nonlinearity of the MSMA-based actuator is extremely complicated to describe by traditional hysteresis modeling methods,since the hysteresis characteristics of the MSMA-based actuator are markedly distinguished from those of other smart material-driven actuators.To tackle this problem,a nonlinear auto-regressive moving average with exogenous inputs(NARMAX)model based on a neural network is developed to depict the complex hysteresis behavior.To improve the capability of characterizing the multi-valued mapping of the hysteresis loop,the Play operator is adopted as the exogenous variable function of the NARMAX model.In addition,a neural network is applied to construct the nonlinear function of the NARMAX hysteresis model,and in this way,the established model can update the model parameters online and accommodate the dynamic variation of the system.To verify the effectiveness of the proposed model,a series of comparison experiments are implemented.The experimental results show that the proposed NARMAX model based on the neural network exhibits excellent modeling performance.Based on the decent model established above,a neural network control method based on the NARMAX model is proposed for the trajectory tracking control problem of the MSMA-based actuator.In comparison with the classical feedforward control method by resorting to the inverse model,the proposed control strategy gets rid of the reliance on the inverse model and eradicates the adverse effects on the control performance arising from the inaccuracy in solving the inverse model.In addition,the effectiveness of the proposed control strategy is also validated from the experimental studies.To address the issues of traditional control schemes with severe model dependence and inferior dynamic performance,a neural network based iterative learning control(ILC)method is proposed to eliminate the influence of the hysteresis on the positioning accuracy of MSMA-based actuator.In the classical ILC method,the parameters are fixed and the preconditions that the initial state of the system is the same for each iteration and the iteration lengths are identical need to be fulfilled.These requirements constrain the application of the ILC scheme in practical systems.In this thesis,the neural network is integrated into the ILC scheme such that the generalization as well as the control performance of the ILC will be enhanced.Furthermore,the actual system suffers from the problems such as non-strict repetition of the initial state and varying iteration length.In the controller design procedure,this thesis studies the convergence of the neural network based ILC under the influence of above mentioned adverse factors,and obtains the convergence conditions of the constructed controller,which makes up for the drawbacks of the classical ILC.The experimental results show that the proposed neural network based ILC method can ensure the MSMA-based actuator to track the desired signal accurately.The presence of the time delay often becomes a potential reason for the system instability and may yields a large steady-state error,which deteriorates the control performance of the system.A neural network adaptive control approach is proposed to control the MSMA-based actuator in this thesis.The designed control scheme takes the effect of the time delay into account so as to improve the positioning accuracy of the MSMA-based actuator.To characterize the MSMA-based actuator,we fuse the NARMAX model with the canonical form of the nonlinear system.Then,a neural network adaptive control strategy is developed to address the impact of the unknown time delay and achieve the better positioning precision of the MSMA-based actuator.By using Lyapunov theory,the tracking error is proven to be asymptotic convergence.Plenty of experiments are conducted on the MSMA-based actuator and experimental studies are provided to validate the effectiveness of the proposed control scheme.