Cutoff-Frequency Phase-In Iterative Learning Control

Abstract

In this brief, a cutoff frequency phase-in iterative learning control (ILC) method is proposed to deal with initial position offset and improve tracking accuracy. In the proposed method, the cutoff frequency of the filter for tracking error is time-varying and follows a predefined profile. The cutoff frequency at initial phase of each operation is high to allow learning system suppress the initial position offset and its influence to improve tracking performance. High initial cutoff frequency results in a quick error decay in the initial phase of operation and also prevents error accumulation in later phase of operation so that good learning transient can be expected at the same time. Profile design and parameter selections are discussed. Experimental results demonstrate the effectiveness of the proposed method.