Towards evolving fuzzy reference controller

In this paper an attempt of designing the direct self- learning/evolving fuzzy controller is presented. The controller is comprised of several local-controllers each valid in a certain part of input-output space. The input-output space is partitioned by on-line clustering and evolving mechanisms taken from eFuMo method. The controller gains are adapted using the fuzzy model reference adaptive control approach. The controller was tested on a heat exchanger simulation model, linear model and pH process model.