An approach to the design of MIMO fuzzy controllers in cases of incomplete expert knowledge

The paper presents an approach for the establishment of the fuzzy rule bases when the knowledge of the system to be controlled is incomplete and/or for studying the feasibility of the control law of physically unavailable systems. The tools we define, in the fuzzy logic framework, allow one to build control laws starting from the knowledge of only a few particular states of the controlled system. Thus one can cope with the lack of expert knowledge and reduce the size of the knowledge bases. Our approach is illustrated by the control of two nonlinear systems: an inverted pendulum and a rotary T/sub 1/O/sub 2/ kiln.