MULTIMODE-ORIENTED POLYNOMIAL TRANSFORMATION-BASED DEFUZZIFICATION STRATEGY AND PARAMETER LEARNING PROCEDURE

In an earlier paper [1], we proposed a set of generalized defuzzificat ion strategies which can be characterized as single-mode-oriented stra tegies. A single-mode-oriented defuzzification strategy, although usef ul in many research projects and real world applications, cannot be ap plied to a multimode situation where two or more distinct possibility peaks exist in its membership function distribution. In this paper. fo r multimode-oriented generalized defuzzification applications, a multi mode-oriented polynomial transformation based defuzzification strategy (M-PTD) is introduced. The new M-PTD strategy, which uses the Kalman filter in parameter learning procedure, offers a constraint-free and s elf-renewal defuzzification solution.