Contradiction sensitive fuzzy model-based adaptive control Original Research Article

Fuzzy model-based adaptive control, unlike traditional fuzzy control, extracts expert knowledge from data by using model identification techniques. In this paper, we propose an analysis of the contradiction in this knowledge as an additional search criterion during the model identification process. In order to do so, we first define a measure of contradiction between fuzzy rules. Then, we assume that a minimum degree of consistency exists in the rule base, and a process of attenuation is carried out between the rules that do not comply with this degree. This allows conflicting situations to be detected without having to wait for the learning dataset to reveal it. Finally, the results obtained from the method proposed are compared with those of a fuzzy model-based adaptive controller which is not sensitive to the contradiction.