RBF Neural Network Control Based on Parameter Adjustable CNF for Robot Manipulators Behaviors

In order to solve the influence of uncertain factors on the robot system and realize the ability of the uncertain robot system to accurately track the reference input, the strategy of combining adaptive RBF neural network control and parameter adjustable composite nonlinear feedback (CNF) control is studied. A robot CNF controller based on adaptive RBF neural network compensation is proposed. The core is to use the adaptive RBF neural network control for online approximation. As the compensation item of CNF controller, the advantages of the two control methods are fully utilized to reduce the influence of uncertainty factors on system performance. The convergence of the closed-loop system is proved by the feedback linearization technique and Lyapunov theory. The simulation results show the effectiveness of this scheme.