PREDICTIVE CONTROL-BASED ON LOCAL LINEAR FUZZY MODELS

This paper deals with predictive control based on fuzzy models. A nove l algorithm (LOLIMOT) is proposed for the construction of Takagi-Sugen o fuzzy models. The rule consequents are optimized by a local orthogon al least-squares method that selects the significant regressors. The r ule premises are optimized by a tree construction algorithm which part itions the input space in hyper-rectangles. A generalized predictive c ontroller (GPC) and a dynamic matrix controller (DMC) are designed. Bo th controllers require the extraction of a linear model from the Takag i-Sugeno fuzzy model. For the GPC a new technique called local dynamic linearization is proposed that exploits the special structure of the local linear models. The DMC is based on the evaluation of a step resp onse. The effectiveness of both the identification algorithm and the p redictive controllers is shown by application to temperature control o f an industrial-scale cross-flow heat exchanged.