RBF neural network adaptive sliding mode control based on genetic algorithm optimization

Abstract

RBF neural network adaptive sliding mode control strategy based on genetic algorithm optimization is proposed for the non-linearity and the parameter uncertainty of the robot manipulator and the influence of disturbance, friction and other factors. The sliding mode variable structure control is used to overcome the uncertainty of the system, considering the influence of the additional uncertain disturbance, the non-linearity part of the fraction, parameter variation and modeling errors and so on. RBF neural network online self-learning optimized by genetic algorithm for dynamic model of robot manipulator raises the global optimization efficiency and effectively solve the problems of the number of neurons nodes in the hidden layer and the value election of each parameter. The Simulink is carried out taking the trajectory tracking of the robot manipulator as an example. The results show that this method can effectively compensate the modeling errors, realize adaptive control of the robot manipulator with no accurate model, improve the system robustness to the external uncertain disturbance and effectively decrease the chattering of control system only using a sliding mode variable structure control. The method is effective and feasible.