Research On The Fuzzy Neural Network Synchronous Control Of H-type Motion Platform

With its advantages of high precision and high efficiency,H-type motion platform is widely used in semiconductor field,precision machining field and other manufacturing industries.The H-type motion platform consists of three Permanent Magnet Linear Synchronous Motor(PMLSM).PMLSM is susceptible to external disturbance,end effect,nonlinear friction and other uncertain factors,resulting in the position tracking error on the single axis of the H-type motion platform system during operation.Meanwhile,the torsion force generated by the reciprocating motion of PMLSM on the X-axis will affect the parallel motor on the Y-axis,resulting in synchronization error between the two axes.The purpose of this thesis is to reduce the position tracking error of H-type motion platform and improve the synchronization control accuracy between the two axes.Firstly,the development status and main control strategies of H-type motion platform at home and abroad are introduced.Considering that PMLSM is vulnerable to uncertainties such as end effect and nonlinear friction,as well as the influence of torsion force generated by reciprocating motion of X-axis PMLSM on the Y-axis parallel motor,based on Lagrange equation,the mathematical model of H-type motion platform is established from the viewpoint of kinetic energy and potential energy.Secondly,in view of H-type motion platform system by the thrust fluctuation,the influence of nonlinear friction and other uncertain factors lead to the position tracking error,existing in the system of fractional order sliding mode control method is used to reduce the single shaft position tracking error,can make the system in finite time convergence,and through its full performance characteristics can effectively weaken chattering.Lyapunov stability theorem is used to prove the stability of the designed control system,and through simulation analysis,fractional-order sliding mode control is compared with sliding mode control,and it is obtained that fractional-order sliding mode control can effectively reduce the position tracking error on each single axis and improve the robustness of the system.Finally,in order to improve the synchronous control performance of Y-axis double-linear motor in H-type motion platform and reduce the synchronization error between the two axes.In this thesis,probabilistic fuzzy neural network compensator is proposed.In order to further improve the approximation ability of membership function,probability distribution function is used to approximate random environment,the fuzzy logic system is used to deal with the nonlinear factors,and the dynamic approximation ability of the neural network is used to deal with the time-varying factors.The probabilistic fuzzy neural network controller is compared with the neural network controller and the controller is simulated and analyzed by MATLAB/Simulink.