People Education Research Industrial Agenda  
Current Research Archive Publications PhD theses Software      

next up previous contents

Fuzzy control of multivariable processes

Project members: R. Babuška, S. Mollov

Fuzzy control provides effective solutions for nonlinear and partially unknown processes, mainly because of its ability to combine information form different sources, such as available mathematical models, experience of operators, process measurements, etc. Extensive research has been devoted to single-input single-output fuzzy control systems, including modeling and control design aspects, analysis of stability and robustness, adaptive control. Multivariable fuzzy control, however, have received considerably less attention, despite strong practical needs for multivariable control solutions, indicated among other fields from process industry, (waste)water treatment, or aerospace engineering. Yet, theoretical foundations and methodological aspects of multivariable control are not well developed.

This research project focuses on the use of fuzzy logic in model-based control of multiple-input, multiple-output (MIMO) systems. Recent developments include effective optimization techniques and robust stability constraints for nonlinear model predictive control. The developed predictive control methods have been applied to the design of an Engine Management System for the gasoline direct injection engine benchmark, developed as a case study within the European research project FAMIMO (see Figure 8). An extension of the Relative Gain Array approach has been proposed that facilitates the analysis of interactions in MIMO TS fuzzy models. The results of this research are reported in the Ph.D. dissertation by S. Mollov.

Figure 8: Fuzzy predictive control of a gasoline direct injection engine.

next up previous contents
Next: Affordable digital fly-by-wire flight control Up: Controller design Previous: Neuro-fuzzy modeling in model-based fault

Back to top

Last modified: 24 March 2005, 10:16 UTC
Search   Site map