Psychoacoustic impacts estimation in manufacturing based on accelerometer measurement using artificial neural networks

In recent years, psychoacoustic impacts of a product has won increasing importance in various design and manufacturing sectors. However, conventional measurement setups based on microphones are expensive and noise-sensitive. This paper proposes a novel method to estimate psychoacoustic parameters from accelerometer measurement by using artificial neural networks. The proposed method has been successfully applied on automotive vehicle interior components which produce nonstationary sounds when operated. In order to develop and tune the proposed method, the operation sounds are first measured by a microphone and an accelerometer simultaneously. Then, static and dynamic psychoacoustic parameters are calculated from the microphone signals according to the auditory model. Finally, the relationship between the psychoacoustic parameters and the accelerometer signals is approximated by feedforward multilayer neural networks. As a result, the performance of the proposed method using artificial neural networks is successfully validated on the existing database.