Recurrent Wavelet Fuzzy Neural Network Control For H-type Platform Driven By Linear Motor

The H-type motion platform has been applied in more and more areas in recent years,due to its parallel system of the double-linear motor with high precision and high dynamic performance.In this thesis,a two-axis mathematical model including beam interference force is established to solve the uncertain disturbance and its double-axis mechanical coupling of H-type motion platform.The global sliding mode controller for single-axis and recurrent wavelet fuzzy neural network compensator for double-axis of the H-type motion platform is designed to improve the tracking accuracy and synchronous accuracy of the platform.Firstly,the developed status,structural characteristics and working principle of permanent magnet linear synchronous motor and H-type motion platform is introduced.Due to the application of the H-type motion platform driven by linear motor with the uneven force in the Y direction dual motors caused by the reciprocating motion of X-axis load,the mathematical model of the H-type motion platform considering the varying beam interference and uncertainty is established.Then,according to the problem of load change and disturbance in permanent magnet linear synchronous motor,the single-axis global sliding mode controller is designed to improve the tracking accuracy of the system and suppress the high frequency chattering.Then the Lyapunov theorem is used to verify the stability of the designed global sliding mode controller.In order to improve the two-axis synchronous accuracy of the motion platform,the single-axis tracking error is combined with the double-axis synchronous error,and the double-axis cross-coupling controller is designed according to the combined mixing error,and the mixing error is used as the input of the single-axis controller.The designed control system is simulated by MATLAB/Simulink,and the designed method can be used to reduce the tracking error and synchronous error of the system.Finally,in order to further reduce the double-axis synchronous error of the motion platform and improve the stability of the system,the wavelet transform capable of time-varying signal analysis and the recurrent structure with dynamic capability are combined with the fuzzy neural network.The translation and dilation factors in the wavelet function,the mean and standard deviation in the fuzzy structure,and the connection weights between the layers are learned and adjusted by the neural network.The recurrent wavelet fuzzy neural network compensator is designed to replace the cross-coupling synchronous controller according to the double-axis synchronous error.The designed compensator was simulated by MATLAB/Simulink and compared with the simulation results using the cross-coupled controller to analyze the simulation results of the two controllers.