Study On Hysteresis Modeling And Control Of Piezoelectric Micro-manipulation Stage

Piezoelectric micro-manipulation stage was widely applied in integrated circuit manufacturing,medical science,micro-mechanical manufacturing,optical processing and other frontier fields,which require high precision,fast response,good stability and other properties.The nonlinear characteristics of micro-manipulation stage will be occurred due to the hysteresis of piezoelectric actuator,which has seriously affected the motion accuracy and dynamic performance of the stage.The establishment of the hysteresis model and the research of the control strategy are the key problems to be solved urgently in the micro-manipulation stage.In this paper,1-dimensional micro-manipulation stage as the object,respectively,from the composite control,single neuron PID control and fuzzy control for motion tracking control research.The main contents are as follows:Hysteresis model of piezoelectric actuator(PEA)is established.In order to improve the control precision of piezoelectric actuator,a hybrid modeling method was proposed to describe the hysteresis characteristics based on Preisach and support vector machine(SVM).The regression model of the hysteresis loop of the PEA is established by using support vector machine theory,and the output displacement corresponding to any voltage sequence can be predicted by the Preisach model.The grid search method based on cross validation was applied to optimize penalty parameter c and kernel function parameter g,which has a great influence on the fitting accuracy of regression model.To illustrate the established model can accurately reflect the hysteresis nonlinearity of piezoelectric actuator,the experimental verification analysis was carried out as for the PEA with model,and the relative error range of PEA output displacement of the measured and predicted values are 0.6%-2.1%.The experiment results showed that the proposed modeling method is feasible and effective.The compound control algorithm is designed to improve the positioning accuracy of micro-manipulation stage.In order to solve the problem of the hysteresis nonlinearity of micro-manipulation stage,a compound control algorithm of discrete Preisach inverse model combined with PID feedback is proposed.The theoretical model reflecting the hysteresis phenomenon is established by combining discrete Preisach model and support vector machine.The inverse model of discrete Preisach is obtained by adopting the iterative search method based on the hysteresis model.The feedforward compensation of the stage is carried out based on the inverse model.In order to correct the deviations by Preisach inverse model and the external environment,the PID control is used to feedback regulation.In order to setting of PID parameters,the transfer function of the micro-manipulation stage is obtained by using experimental modal method.In order to illustrate the feasibility of the proposed control method,the experiment is carried out.The results show that the proposed method has better control accuracy and fast response.The single neuron PID control strategy based on RBF neural network is proposed for motion tracking control.Although compound control has advantages of higher accuracy and better rapidity,hysteresis modeling complex and the model is not adaptive,so single neuron PID control strategy is proposed.Dynamic model of micro operation platform,is used to determine the initial value of connection weights of the single neuron PID control can be determine based on the model.The gradient information of the micro-manipulation stage can be obtained online by using RBF neural network identifier,and online adjustment information of PID connection weights can be obtained.The learning algorithm of the single neuron network is applied to achieve the online self-tuning of the PID parameters in order to achieve the adaptive motion tracking control of micro-manipulation stage.The experimental results showed displacement error range of single neuron PID based on RBF neural network is [-0.5 ~ 0.5] ?m,and adjustment time is 0.1 s.It showed that the proposed control method has better control accuracy and response speed,and has stronger adaptabilityPosition precision compensation method of a micro-manipulation stage is based on fuzzy control.For dynamic hysteresis characteristics of piezoelectric micro-manipulation stage,a position accuracy compensation method is proposed based on the fuzzy control strategy to get rid of the dependence on hysteretic model.As for a one dimensional micro-manipulation stage,the position deviation and deviation rate of the stage is fuzzy input,and the input voltage change of piezoelectric actuator is fuzzy output.A method of developing fuzzy rules is presented based on experiment data of PID control in order to acquire experience.In order to illustrate the feasibility of the proposing method,analysis platform tracking position errors when the stage is tracking the sine signals with different frequencies.The experimental results showed that the proposed fuzzy control method can make the stage have higher position tracking accuracy and faster tracking speed,and has better adaptability.