Iterative Learning Control for Minimum Time Path Following

Abstract

The technique of iterative learning control (ILC) is frequently applied to improve the tracking performance of those systems operating repetitively. This paper extends the ILC task description to handle path following problems by relaxing assumptions on fixed motion profiles and trial length. The removal of these design constraints yields significant control design flexibilities to choose an admissible solution of the problem so that an extra performance index is optimized. A two stage design framework is considered to find the minimum path following time, and guarantee a feasible solution of the path following problem under system constraints using ILC. The solutions of the two stages are provided by a norm optimal ILC update and the bisection method with implementation guidelines, which are combined to form a comprehensive implementation algorithm. The effectiveness of the proposed algorithm is evaluated using a case study on a gantry robot model.