Modeling And Compensation Of Hysteresis In Piezoelectric Actuator Based On Least Squares Support Vector Machine

Piezoelectric actuators(PEAs)are widely used in precision positioning technology as micro-displacement devices due to the merits of high positioning resolution,rapid response speed,small size,and large driving force.However,the inherent hysteresis of PEA degrades its positioning accuracy.In order to improve the control precision of piezoelectric-based micro-displacement system,this paper focuses on the study of the hysteresis modeling and compensation method based on inverse model of PEA.For hysteresis modeling,a dynamic Least Squares Support Vector Machine(LSSVM)model is presented using Nonlinear Auto Regressive Exogenous structure,the parameters of which are optimized by using particle swarm optimization.In order to overcome the defects that LSSVM is more sensitive to outliers and less robust,an adaptive weighting method is proposed.In this method,smaller weights are assigned to those sample data with larger errors and zero weights are assigned when the errors exceed a certain threshold.The experimental results show that the proposed method can achieve accurate prediction for samples with different frequencies.Meanwhile,the proposed model is able to eliminate the effects of outliers,reduce the sample size and improve the robustness and sparseness.To compensate for the hysteresis behavior,two feedforward control schemes according to different inputs of LSSVM inverse model are proposed.The test results show that the positioning accuracy of two separate feedforward methods are poor.To revise the model input,a hybrid feedforward controller combining two kinds of feedforward control schemes is put forward,which effectively weakens the influence of hysteresis nonlinearity on the control accuracy.To further improve the robustness and tracking performance,the hybrid feedforward-feedback control scheme is constructed based on the hybrid feedforward control and the PID feedback control.The comparative study with the traditional PID feedback control and Preisach feedforward control reveals the superior tracking performance of the proposed control strategy.Moreover,the hybrid feedforward-feedback controller is capable of tracking different testing waveforms with negligible errors,which confirms the generalization ability of the proposed approach.