Fundamental study on sound source localization inside a structure using a deep neural network and computer-aided engineering

• DOI: 10.1016/j.jsv.2021.116400
• Publication Date: 2021-08-18T13:13:20Z
• Authors: Shunsuke Kita; Yoshinobu Kajikawa
• Keywords: Position (finance)

This study proposes a sound source localization (SSL) method applicable to sources inside structures such as mechanical equipment or buildings. Presently, an SSL system employing microphone array based on the time difference of arrival estimation can be used localize in same acoustic space microphone. However, conventional methods cannot adopted when is located structure. Achieving more difficult case indirect than that direct sound, because correlation between observed signals becomes stronger owing effect coupling acoustics and To solve this problem, deep neural network computer-aided engineering, which structure’s interiors, proposed. The proposed method’s effectiveness feasibility are examined via numerical actual experimentation. estimate position structure spectrum measured by accelerometer surface results experiment indicate test accuracy 93.20%, whereas yielded 61.53%. learning validation curves show lower occurrence overlearning, from small amount data applied. overcome issue, augmentation was used; consequently, improved 99.82%.