A neuro-fuzzy approach to design adaptive control systems for PM motor drives

The paper deals with the theoretical development and the practical implementation of an adaptive speed regulator scheme based on the Model Reference Adaptive Control approach, employing fuzzy laws for the adaptive process. A neuro-fuzzy approach for the optimization of the fuzzy adaptive laws is used to face one of the major drawbacks of the fuzzy control, the lack of a systematic optimization procedure. A performance comparison between the new controller and a conventional MRAC scheme is carried out by mean of simulations confirming the superiority of the proposed fuzzy adaptive regulator. Experimental tests are presented in order to practically evaluate the features of the proposed adaptive control scheme.