A variable-structure fuzzy logic controller with run-time adaptation

Abstract

The design of a variable structure fuzzy logic controller for minimum-time set-point control is presented. The controller architecture is extended to accommodate runtime adaptation. This innovative approach combines the quick response of bang-bang like control with the stable convergence properties of more conservative linear control to produce a robust high performance controller. The controller is based on a Takagi-Sugeno-Kang format but the controller is warped during each step response to change its structure. The warping is done through scaling of the controller inputs and outputs. An online scheme for finding the best values for the scaling is presented.<>