Neural network learning and generalization for performance improvement of industrial robots

Abstract

In this article, we present an approach for improving the trajectory tracking performance of industrial robots using multilayer feedforward neural networks. The controller design based on this approach consists of a PID control and a neural network. The function of the neural network is to complement the PID controller for improving the performance of the system over time. The proposed approach has been implemented on an industrial robot-the CRS Robotics A460. Experiments are conducted to investigate the learning and generalization ability of neural networks in complementing the PID method in robot trajectory tracking. The results of this work suggest that neural networks could be added to existing PID-controlled industrial robots for performance improvement.