Similarity-based rules reduction of a fuzzy logic controller for a PMDC motor drive system

This paper investigates the effects of reducing the number of rules in a fuzzy logic controller (FLC) for a permanent magnet direct current (PMDC) motor. The reliability of the resulting controller, with fewer rules, is also investigated. First, rule extraction is shown front typical transient responses of a PMDC motor when the reference speed changes in a step-like manner. Next, the total number of rules is reduced from 49 rules to 9 rules by eliminating similar fuzzy rules; their corresponding membership functions are then replaced by other already available membership functions. Finally, the FLC consisting of 9 rules has been chosen in this study to investigate the sensitivity of the controller. The values in the decision table are changed in the range of /spl plusmn/30%. It is found that these variations in the rule decision table do not cause any instability problem in the proposed FLC.