Rule Interpolation in Fuzzy Rule Bases with Defects

Abstract

This research report treats sparse fuzzy rule bases with defects. The authors propose a method to interpolate a rule at an arbitrary point in such a rule base by using bi-splines. Also, they investigate the method to make the size of the rule base smaller with keeping accuracy, by applying a kind of smoothing splines. Some simple numerical simulations of a two-inputs one-output control system show the effectiveness of the proposed methods.