Robot trajectory planning based on GA-RBF neural network

Abstract

Abstract A robot trajectory planning method combining RBF neural network and genetic algorithm is proposed based on the robot motion model to address the accuracy and smoothness of the end effector motion trajectory of robots in uncertain environments. This article establishes the D-H matrix and kinematic model and optimizes the network structure, connection weights, and width of the RBF neural network through a genetic algorithm to accurately track the trajectory of the robot. By comparing with other trajectory planning algorithms through simulation, the results show that the GA-RBF neural planning method has small errors, strong stability, and can meet the expected requirements of industrial robot trajectory planning. This provides a certain reference for the development of robot trajectory planning methods.