Combined adaptive and fuzzy control using multiple models

Abstract

This paper presents a control strategy to deal with processes in which both system parameters and orders are unknown and undergo abrupt structural variations. To handle such complicated control problems, a combined adaptive and fuzzy control scheme is developed. A number of Generalized Minimum Variance (GMV) controllers are designed according to all the possible process structures. A higher level model selection mechanism will decide which controller candidate is the best. A fuzzy modification algorithm is introduced to improve the system responses in transient period by detuning control weights of GMV controllers. To cope with possible instability caused by non-minimum phase dynamics, a fuzzy PID backup is introduced as well. Some adaptation is added to the fuzzy PID backup to obtain better system performance.