Abstract

In this paper a simple generalization method of an iterative learning control scheme is proposed. A look-up table implementation of a Takagi-Sugeno fuzzy logic controller is developed based on a self-organizing control scheme. For that purpose, a first order penalty equation is used as a reinforcement in an iterative learning run, modifying some elements of the table recorded as responsible of the actual bad performance. The schema does not need a model of the process, but experimental process knowledge is required to tune the penalty equation.