A case study on multiple controller adaptive control

Abstract

This document reports work on the application of a Multiple Controller Adaptive Control algorithm based on fictitious reference signals. The algorithm has to select a controller in a small set of candidate controllers. The work shows steps towards a good selection of a cost functional and an analysis, based on recent stability results, shows why a cost function that is usual in other control methods is not suitable. With appropriate cost functional and set of controllers, simulations show the method is able to control a time varying process in the presence of measurement noise and loop delay, even outside the design interval for the parameter variation.