Aportaciones al control inverso con modelo de referencia basado en lógica borrosa, redes neuronales y algoritmos genéticos

Abstract

Debido al interés que existe actualmenten en combinar el control de prealimentación -Feed Forward (FF)- con el clásico de realimentación -Feed Back (FB)-, el presente trabajo tiene como fnalidad marcar una estrategia de diseño de un sistema que incorpora estos dos controles.

The aim of this work is to develop a system design strategy that is composed of two types of control: Feed Forward Control (FFC) and Feed Back Control (FBC). The FFC implements an approximate inverse of the plant, by neural networks, and the FBC makes the system more robust, being designed as a Fuzzy Proportional Integral Controller (FPIC). A genetic algorithm is included to tune the neural network weights plants and the final control system gives a better performance than other techniques published.