Abstract
This article presents a methodology to estimate fuzzy models for fuzzy c;ontrol of dynamic systems. The basic motivation is to develop fuzzy controllers that make use of the dynamics of the system. The fuzzy modeling is accomplished by an inductive learning algorithm that uses the concept of rough set theory. The fuzzy controller is constructed from the inverse of the fuzzy model. Therefore, it tries to cancel the dynamics of the system. We show that the fuzzy models can be obtained with a minimal a priori knowledge about the system and that they are good predictors. The inverse-model-based fuzzy controllers developed here are, in general, high pass filters that may amplify the noise and cause saturations. Two simulated examples illustrate the methodology.
Get full access to this article
View all access options for this article.