Design of fuzzy controller based on data mining
Peng Xia, Yan Yuan, Cao Weihua, Min Wu · 2015
The design of existing empirical fuzzy controller is based on the practical experience or expert knowledge, which has obvious subjectivity and strong uncertainty. Meanwhile, the value of large process data that imply various patterns and much useful information is ignored. Aiming at these problems, one type of design method of fuzzy controller based on data driven is proposed in this paper, which extracts fuzzy subsets and fuzzy rules from the test data directly. Especially, an improved data mining(iDM) method of rule base is presented. Using a second-order plus time delay model, a series of simulation experiments are conducted and the results show that the method is feasible and effective. As a contrast to the PID controller and the general fuzzy controller, the control effect of this proposed method is demonstrated and the superior performance is verified.