A NEW SELF-TUNING BLENDING MECHANISM FOR FUZZY PI AND PD CONTROLLERS

In this study, a new mechanism that blends the "PI-type " and "PD-type " output portions of a fuzzy logic controller (FLC) is presented. The FLC used in this study has a new structure and a newly devised input (s) named as “normalized acceleration”. An empirical relation and a rule-base are presented for blending PI and PD portions of the FLC in an on-line and self-tuning fashion. Both the empirical relation and the rule-base use the same inputs of the FLC as intelligent procedures. The robustness and effectiveness of the blending mechanism are illustrated through simulations done on a second-order system with varying parameters. I