Experimental verification of accelerated norm-optimal iterative learning control

Abstract

Accelerated Norm-Optimal Iterative Learning Control (NOILC) is a recently developed method to improve the convergence performance of the well known NOILC algorithm. This paper investigates the effectiveness of this method experimentally on a gantry robot facility, which has been extensively used to test a wide range of linear model based ILC algorithms. The results obtained confirm that the accelerated algorithm outperforms NOILC algorithm and in particular, the improvements at initial stage can be substantial, which is of great interest in practical applications.