BNN-based fuzzy logic controller for flexible-link manipulator

Abstract

As is well recognized, rule acquisition has been regarded as a bottleneck for implementation of fuzzy logic controller. Moreover, defuzzification is a time consuming procedure. Though Roger Jang's adaptive neuro-fuzzy and Sugeno's fuzzy systems eliminated those shortcomings but both require a set of input-output data, which may not be available always. This paper reports on a backpropagation neural network based fuzzy logic controller, where neural network is trained using linguistic description. © 2006 Springer-Verlag Berlin Heidelberg.