Improved ADRC Synchronization Control Of H-type Platform Based On Fuzzy Neural Network

The direct drive H-type motion platform researched in this thesis is driven by two parallel permanent magnet linear synchronous motors（PMLSM）in the Y-axis direction and one PMLSM in the X-axis direction on the beam.With the advantages of high positioning accuracy,large thrust and fast response speed,it is widely used in laser cutting,computerized numerical control machine tools,machining and other fields.Due to the PMLSM system removes the intermediate transmission device and is directly connected with the load,it is easy to be affected by mechanical coupling and uncertain factors such as parameter perturbation,and external disturbance,resulting in the dynamic mismatch between the two motors in the Y-axis direction and position synchronization error,which is very unfavorable to the processing accuracy of the product.If it is serious,it will cause system damage and even endanger personal safety.The purpose of this thesis is to design an improved active disturbance rejection control to improve the single axis control accuracy of the system.At the same time,for the problem of synchronization of two motors in the Y-axis direction,a fuzzy neural network controller is designed to improve the synchronization accuracy of the system.Firstly,by consulting the literature at home and abroad,this thesis elaborates the development status,structural characteristics and system control strategy of H-type platform.During the movement of H-type platform,the motor on the beam moves back and forth,resulting in the deviation of the overall mass center of the beam and the uneven stress of the two motors in the y-axis direction,based on the basic principle and dynamic model of PMLSM,the mathematical model of H-type platform is established.Then,in view of the fact that the H-type motion platform is easily affected by mechanical coupling,external disturbance and other uncertain factors,the active disturbance rejection control is designed as the single axis controller of the H-type platform to ensure the tracking performance of the system.Since the nonlinear function in traditional active disturbance rejection control is not differentiable at the inflection point,it is necessary to design a smoother differentiable nonlinear function.At the same time,in order to reduce the observation pressure of the extended state observer,the least squares support vector machine is introduced to optimize it to improve the single axis tracking accuracy of the system.Aiming at the problem that the two motors are not synchronized in the Y-axis direction,a cross coupling controller is designed to ensure the synchronization performance of the H-type platform.Cross coupling control is to mix the tracking error and synchronization error of the platform,and redistribute the mixed error to each axis to keep the two motors synchronized and improve the synchronization performance of the system.Finally,in order to further reduce the synchronization error of the system,the T-S fuzzy pi-sigma neural network synchronization compensator is designed by combining fuzzy control with neural network control.Using the strong learning ability of fuzzy neural network and the approximation ability to any nonlinear function,the synchronization error is close to zero.The gradient descent method is used to modify the parameters in the fuzzy neural network in real time to realize the dynamic compensation of the system.The proposed control strategy is modeled and simulated by MATLAB/Simulink to verify its feasibility and effectiveness.