Tuning of Fuzzy Controller by Variable Clustered Fuzzy Rules and Its Application to Overhead Crane

This paper proposed efficient rule extraction and rule reduction methods for the self-tuning fuzzy controller. The Fuzzy Clustering Method (FCM) and similarity approach are applied to extract and reduce the fuzzy gain rules. The proposed rule extraction scheme is investigated on a self-tuning fuzzy proportional plus derivative controller (STFLPDC), having 49 fuzzy rules and 49 fuzzy gain rules. The utility of the scheme is validated with various clustering validity indices. The effectiveness of the self-tuning fuzzy controller with the different reduced number of extracted fuzzy gain rules generated from clustering data is found to be quite satisfactory in comparison with the initial (49) fuzzy gain rules. The effect of gain rule variations on STFLPDC is tested to control the position and swing of an overhead crane and a comparison is made with other fuzzy and conventional controllers.