Comparison of defuzzification methods for cabin noise prediction of passenger cars

Abstract

These days passenger cars have to fullfil far-reaching expectations. Among the most prevelant ones is to provide a high level of travelling comfort. That issue contains acoustic well-being which includes cabin noise as well. In this paper, the results of acoustic measurement are presented and used for build up a fuzzy inference system. Five types of defuzzification techniques were compared: cetroid, bisector, MOM, LOM and SOM methods. It was revealed that LOM provided the best fitting and the lowest range of errors. The concept was verified by further confirmation measurements.