The formulation of a simple fuzzy model tuning predictive controller for MIMO systems

Abstract

The author (1993) studied a model rule form that describes the cause-effect relation between past control actions and process responses, in which the control actions are defined by three fuzzy variables, long past, medium past and short past actions, in order to formulate a simple fuzzy model tuning predictive controller (FMTPC) for SISO systems. In this paper, the formulation of a simple FMTPC for MIMO systems is investigated by applying the same representation of control actions as for SISO systems, the rule form of partial pairings of input and output relation, and fuzzy linear functions for the calculation of model output. The model representation proposed is in analogy with the way human operators perceive their past control actions and resulting process responses linguistically. Due to these treatments, the whole process of controller design has been simplified considerably without sacrificing controller performance, and consequently a controller applicable for practical uses has been realized.<>