A Method for Improving the Dynamic Accuracy of a Robot Performing a Repetitive Task

Abstract

In many applications it is desirable to improve the dynamic accuracy of robots. In this paper a simple scheme for improv ing the accuracy is presented whereby the robot improves its performance each time the task is performed. The method makes use of the discrete time internal model principle. The performance of the algorithm is confirmed by computer sim ulation studies using a full nonlinear model of a 3-degree-of- freedom robot. The studies indicate that a dramatic improve ment in dynamic accuracy is achievable with the method.