Using a hidden mathematical model generated by input-output data. the reliability of a neuro-fuzzy controller for a servodrive
Dan S. Mihai · International Conference on Mathematical Methods and Computational Techniques in Electrical Engineering · 2006
The paper deals with the synthesis of a neuro-fuzzy controller for a servodrive. Both position and speed loops are monitored by a single Sugeno fuzzy controller. The fuzzy solution is analyzed comparatively with the standard structure that generated it by training the neural network. Different training conditions are taken into account. The influence of the motor parameters, load characteristics, reference value and tuning parameters are considered. The proposed solution is simple and ensures a good and robust behavior.