Compensated Iterative Learning Control of Industrial Robots

In this paper, a compensated iterative learning control algorithm for trajectory tracking of industrial robots is proposed, which is applicable for the spike and jitter phenomenon in the process of error convergence. Unlike previous iterative learning control algorithms, this algorithm is constructed not only by using the last input and deviation signals, but also to incorporate the control and deviation signals before last to learn more information. Based on the original D-type learning law, the input and the error terms are compensated respectively, so as to accelerate the convergence speed and effectively eliminate the jitter in the process of error convergence. The speed and accuracy of the system can be improved simultaneously. There is a certain reference value for the high precision control of the robot. Simulation and experimental results also confirm and verify the effectiveness of the proposed method.