Developent of localizing rattle noise source methodology using machine learning algorithm

Professor No-cheol Park's in the Department of Mechanical Engineering and his research team have developed a methodology to localize the relative location of the rattle noise sources underbody of the vehicle using the signal measured using several accelerometers. Through this study, It became possible to quantitaively determine the relative location of the rattle noise source based on the measurement locations. The rattle-induced acceleration signals were clustered using machine learning algorithms with the accuracy of the BSR expert level in the vehicel underbody. The proposed method is based only on the transmission characteristics of and rattle-induced acceleration signals measured in the driving test, and It can be used widely regardless of vehicle type. In recognition of the excellent of research results, this study was published in the international academic journal Mechanical Systems and Signal Processing (Impact factor: 6.823, 2020 JCR, top 1.52%).