Inverse kinematics problem of industrial robot based on PSO-RBFNN

Abstract

In order to solve the kinematics solution of industrial robots, the solution is slow and has low precision. In this paper, the Particle Swarm Optimization algorithm(PSO) is introduced into the Radial Basis Function Neural Network(RBFNN) to optimize, and a high precision PSO-RBFNN algorithm for industrial robots is proposed. The algorithm uses 3-layer RBFNN to solve the inverse kinematics of industrial robots, and combines the kinematics model of industrial robots with PSO to optimize the network structure and connection weight of RBFNN. It realizes the nonlinear mapping from the working space pose of the industrial robot to the joint angle, thus replacing the cumbersome formula derivation and improving its solving speed. Moreover, the training success rate and the accuracy of the solution of the PSO-RBFNN algorithm are improved.