A high precision fuzzy-neural controller based on data remodification

Fuzzy logic and neural networks have wide applications in intelligent systems. The research describes a high precision fuzzy neural controller design method in which a feedforward network learns fuzzy rules offline while employing an analytical approach in place of the conventional error back propagation method which can be time consuming to implement. Through repeated modifications of the system inputs, the proposed technique restores valuable information often lost during fuzzification. As a result, interpolation and data remodification properties inherently rooted in the controller significantly improve the system response