Application of neural networks sliding-mode control in tracking control of robot manipulators

Abstract

Sliding-mode control for trajectory tracking in robot application constantly occurs buffeting,with the problem proposing the neural sliding mode control method based on the filter.According to sliding mode variable structure control theory and neural networks,combining the filter,the two joint robot trajectory tracking control had been simulated and studied.Through the filter and simple sliding mode control method,Radial Basis Function neural worker sliding mode equivalent controlled and the method of RBF neural sliding mode reaching law,the three kinds of method had been compared with and simulatedly analysed.The results show that the design of neural network sliding mode reaching law control system has good tracking,robustness and a high control precision,simple achievement,and effectively eliminated buffeting.