Fuzzy and controller with fuzzy supervision

摘要

The best-known industrial process controllers are proporitonal-integral-derivative (PID) controllers because of their simple structure and robust performance in a wide range of operating conditions. Fuzzy controllers can be viewed as a natural similarity of conventional PID controllers, which can perform the same tasks into the control system. This paper describes the development of a fuzzy PID controller with fuzzy supervision of its parameters. Fuzzy rules and reasoning process as a part of the fuzzy supervisor are utilised on-line to determine the values of the controller parameters based on the ultimate gaina nd the ultimate period of oscillations. Learning algorithm for fuzzy parameters in fuzzy neural implementation of the controller is additionally applied. Simulation results demonstrate that in compariosn with the approaches using Ziegler-Nichols tuning for conventional PID controllers a better system performance can be achieved.