Adaptive tuning of fuzzy set-point weighting for PID controllers

In this paper a new adaptive procedure for the tuning of Proportional-Integral-Derivative (PID) controllers is proposed. It is based on the concept of fuzzy set-point weighting, i.e. the set-point weight value is continuously updated depending on the current system error and its time derivative, according to a fuzzy inference system. In this way, both the load disturbance rejection and the set-point following specifications can be properly addressed. The parameters of the PID gains are fixed according to the Ziegler-Nichols method, while the coefficients of the fuzzy mechanism are adaptively modified based on the step responses. Simulation results show how the overall methodology is very effective in improving the performances in the step response, without requiring the acquisition of specific knowledge from the operator and without requiring a significant extra-effort from a computational point of view. Copyright (C) 2000 IFAC.