Adaptive fuzzy-neuro control of DC drive system

Speed control of electrical drives is often made difficult by nonlinearities, difficulties in proper system parameters identification, unpredictable parameter variations and external load disturbances. As a possibility a model reference adaptive speed control using on-line trained fuzzy neural network (FNN) is presented. In this control method fuzzy-logic controller is equipped with additional option for on-line tuning its chosen parameters. In the paper PI-type fuzzy logic controller is used as the direct speed controller, whose connective weights are trained on-line according to the error between the states of the plant and the reference model. The FNN speed controller is on-line tuned to preserve a favourable model-following characteristics under various operating conditions. The analysis, design, simulation and experimental implementation and comparison of the proposed control schemes are described.