A comparison of some adaptive-predictive fuzzy-control strategies

Abstract

A comparison of some fuzzy model-based adaptive-predictive control strategies is provided. The considered strategies are based on the minimization of a predefined long-range cost functional reflecting the tracking error between the reference to be followed and the plant's output. They begin at the current time instant and continue up to a given time horizon in the future. A fuzzy model of the plant is used for forecasting. A particular control strategy is specified by selecting the type of fuzzy model to be used as a plant description, the algorithm used for model parameter estimation and the assumptions required for the long-range cost minimization. The simulation results presented show that the control performance depends not only on the assumption required for multi-step optimization, but most notably it depends on the convergence rate of the parameter estimation algorithm used in the online model identification.