A General Fuzzified CMAC Neural Network and Its Simulation Study

Aiming at conventional cerebellar model articulation controller (CMAC), and combining CMAC addressing schemes with fuzzy logic idea, a general fuzzified CMAC (GFAC) is proposed, in which the fuzzy membership functions are utilized as the receptive field functions. The mapping of receptive field functions, the selection law of membership and the learning algorithm are presented in the paper. By using GFAC, the approximation of complex functions can be obtained which is more continuous than that by conventional CMAC. The simulation results show that GFAC has good generalization, comparatively high approximating accuracy, and ability to calculate function output differential