Buzz, squeak, and rattle noise classification by using acoustic-fingerprinting technology

The buzz, squeak, and rattle (BSR) noises are three representative types of the automotive interior or exterior noise. Some of BSR noises have very short duration in time, and hence, it is difficult to detect various BSR noises in low SNR situation. However, each BSR noise signal has a unique time-frequency characteristic, depending on the various contacting materials as well as the excitation forces. Therefore, it is necessary to utilize the time-frequency characteristic of the BSR noise to specify the origin of a noise source. In this paper, we propose a novel method and system for identifying BSR noises. For accurate classification of noise sources, a noise-fingerprinting and matching technique based on the pattern classification is devised. The identification test with the real BSR noise data shows that the proposed method can accurately classify the noise source even in the low SNR condition.