Tuning of multivariable fuzzy logic controller

Abstract

A systematic method tuning for a multivariable fuzzy logic controller (MFLC) for unknown, stable processes is proposed. The tuning is based on the steady-state gain and approximate knowledge of the time constants and the delays of the process. These can be calculated from open-loop step responses. The result is a controller which has a proportional and integral mode. The tuning method is presented for a two-input, two-output process to restrict the size of the rule base. Two simulation examples illustrate the tuning procedure.< >