RST-based system design of hybrid intelligent control

This paper proposes a RST-based fuzzy rules extraction program, which uses information entropy and closeness degree as heuristic information for attribute selection to enhance search speed of minimal attributes set, and takes full use of a priori knowledge and experiences. Based on this program, design method of reduced dimension multilayer fuzzy rules is constructed, in which every layer´s fuzzy inference rules dimension is no more than three, in favor of understanding and correcting rules. The rough-fuzzy controller is combined with conventional PID forming hybrid intelligent controller with better control quality.