Adaptive Control for Industrial Welding Robot with Muti-degree of Freedom Using the Improved Neural Network PID

Abstract

A24 at the problem that the traditional control algorithm is not effective to track control of industrial robot in multi-degree of freedom (multi-DOF), a improved adaptive control method based on neural network PID was proposed. On the basis of the backward error propagation of BP algorithm, the method adjusted the BP network weight and thresholds corresponding to PSGA(Predator Search Genetic Algorithm). The PSGA is taken to choose the initial weights and threshold of network, which will improve the speed and precision of network training. The simulation results show that method overcomes the difficulty of traditional PID control by optimizing the dynamic process and reducing the steady-state error of the system.