Self-learning continuous controllers

In this paper different approach of adaptation mechanism for fuzzy model reference learning control (FMRLC) method will be introduced. In contrast to original method the proposed procedure guarantees continuity of the fuzzy controller that in reward results in its smooth input-output behaviour (mapping) that eliminates undesirable abrupt changes in the control signal. Additionally, the proposed method shows larger robustness to responses of initial conditions and various reference signals. The advantages of the proposed modification are presented on control of magnetic suspension system. From result comparing rule bases after adaptation there is absolutely smooth control surface in case of proposed method FMRLC in comparison with original FMRLC. With proposed procedure it was achieved similar result regulation to original FMRLC but with more than 4x smaller adaptation gain.