Fuzzy tuning method for simulation of dynamic processes

Abstract

A new MIMO type (multi-input, multi-output) fuzzy inference procedure for successive identification of a dynamic process of first order is proposed and investigated. It consists of two fuzzy rule bases adjusting the gain K and the time constant T respectively. A special "soft switching" procedure called fuzzy switch has been introduced. It calculates at each time instant the activation degree of tuning needed for both fuzzy reasoning procedures. A test plant example has been used for extensive evaluation of the practical applicability of the proposed fuzzy reasoning method.