Robust self-learning fuzzy logic controller

Abstract

It is known that the self-organizing fuzzy logic controller proposed by Procyk is sensitive to the external signals such as set-point changes and/or disturbances. Also, this difficulty may be encountered in other fuzzy learning controllers that have a learning algorithm to minimize the cost function of the error. To solve this problem, a new robust self-learning fuzzy logic controller is proposed based on the principle of sliding mode control. Computer simulation shows that the proposed method is robust to the set-point changes and the disturbances. Also, to show the applicability to the tracking control of MIMO systems, at is applied to a two-link robot manipulator.