An Adaptive Convergence Ratio Generation Algorithm for Iterative Learning Control

Abstract

Iterative learning control is a technique for improving the accuracy of industrial machining. Its convergence ratio significantly affects the learning efficiency, and is usually set based on user experience. However, this paper develops an algorithm for two-axis machining that automatically generates an appropriate convergence ratio at each learning iteration. This algorithm can rapidly reduce root-mean-square contour errors and the total learning iterations required for machining. The proposed algorithm is integrated into and implemented in LinuxCNC. Experimental results for the tested machining paths show that the proposed algorithm's convergence ratio quickly improves machining accuracy in fewer learning iterations.