New generic fuzzy algorithms-based auto-tuning implementation for DC brushless servo systems and its evaluations

This paper describes a feasible auto-tuning method of DC brushless servo systems which introduces fuzzy algorithms. This is basically considered as a technique to adjust two gains of a PI compensator under the principle providing an additional reference generator, observer and fuzzy reasoning with a general-purpose speed control loop including PI regulator. It is possible to adjust the gains so as to have an optimum level through repeating ten or more step responses in relation to variations of load inertia and mechanical vibration in addition to control characteristic fluctuations. It proves that a DC brushless servo system can be automatically tuned to provide an optimum response, even if it is at a completely unstable state, on the basis of experimental results