Design and stability of self-organizing fuzzy control of high-order systems

Abstract

The paper describes a model reference-based self-organizing fuzzy control scheme which is suitable for high-order systems of well-known structure. The conditions for stability of the learning process, which may be used for determination of the learning coefficient values, have been worked out. The way of improving the system response in the initial stage of learning by using the referent model-based method of presetting a fuzzy control surface is described. The proposed fuzzy control method has been tested in the angular speed control loop of a DCM servo drive. The results obtained have shown the feasibility and marked effectiveness of the self-organizing fuzzy control scheme indicating a better performance of the system with the preset fuzzy controller than with a blank fuzzy rule-table in the initial stage of learning.