Research On Adaptive Control Method Of Uncertain Robot Based On Neural Network

The application of traditional robot control algorithms is based on the known model parameters.With the increasing requirements for robot performance,there is a lot of uncertainty for the robot model.Therefore,these problems are difficult to solve by traditional control methods.In this paper,the robot control system is taken as the research object.Based on the traditional methods,the adaptive control strategy of robot based on neural network is proposed to solve the system instability problems caused by the uncertainty of model parameters and external disturbance in robot trajectory tracking control.The composite controller is designed by neural network control algorithm,sliding mode control algorithm,disturbance observer method and other algorithms to suppress the influence of uncertain factors on the control system.The specific research work is as follows:Firstly,aiming at the uncertainty of robot,a sliding mode adaptive controller based on neural network is designed.A new sliding mode surface is selected to reduce the impact of sliding mode chattering on the control system.On this basis,the neural network is used to approximate the uncertain terms in the model,which significantly improves the adaptive ability of the system.The stability theory is used to verify the scheme,and the effectiveness of the control scheme is verified by comparing the simulation results.Secondly,in order to solve the problem of large workload of computation when neural network deals with system uncertainty,a neural network anti-interference controller based on backstepping method is designed.By using backstepping method and selecting appropriate Lyapunov function,the adaptive control law is constructed systematically.The design of disturbance observer not only reduces the workload of neural network online calculation,but also improves the accuracy of the controller.Finally,the simulation is carried out to verify the scheme.Finally,to improve the adaptive ability of the robot system,a composite controller was designed by combining neural network,sliding mode control and disturbance observer.The chattering is reduced by designing the sliding approach law and the dynamic performance of the system is improved.The scheme verifies the stability performance of the disturbance observer,thereby better guaranteeing the global stability of the system.Simulation shows that the controller has good stable performance and faster convergence speed when dealing with robot uncertainty.