Design of a neurofuzzy controller with simplified architecture.

Abstract

This work presents the design of a neurofuzzy controller with simplified architecture that minimizes the processing time used in several stages associated with systems and processes modelling. The basic procedures of fuzzification and defuzzification are very simplified, whereas the inference procedures are computed in a direct way. The simplified architecture has allowed a fast and easy configuration of the neurofuzzy controller, as consequence, the control rules that define the control actions are obtained automatically. To validate the proposed approach, this neurofuzzy system is used in an industrial application for fluid flow control.