Inverse learning-based heterogeneous fuzzy data fusion for hybrid intelligent control

In order to make the most use of the heterogeneous fuzzy data (HFD) available for hybrid intelligent controller (HIC) design, we classify the HFD into four categories: direct data (DD), direct rules (DR), indirect data (ID) and indirect rules (IR). Based on inverse learning, we convert the ID, IR, DD, and DR into a uniform domain; therefore some data fusion theory can be used to integrate HFD with direct and indirect forms. Some conjunctive (T-norms), disjunctive (T-conorms) and compromise (mean) fusion operators have been selected and compared for the above proposed HFD fusion methods. The validity of the proposed methods is verified through hybrid intelligent control of a biped walking robot. The simulation results show that the gait can be improved by the proposed inverse learning-based HFD fusion methods