Abstract

This paper presents a new approach for tuning Fuzzy PID controllers using Iterative Learning and Least Squares with Regularization. The controller is tuned so that, without knowing a model of the plant and for a given reference signal, the closed loop system will follow a desired model reference as close as possible. The methodology given in this work allows a global design of a fuzzy controller applied to linear systems. Furthermore, some simulations show the attractive characteristics, like performance and its easy implementation, associated to this approach. Copyright (C) 2000 IFAC.