Development of a sound source detection system for outdoor use

We have developed a high-resolution sound detection system using microphone arrays as a noise source identification tool for building acoustics, targeting relatively high frequencies. However, there is a need to identify noise sources such as equipment noise in outdoor environments, and in such cases, it is necessary to consider weather resistance. Moreover, since equipment noise is typically in the low frequency band, it is necessary to address low frequencies as well. Therefore, we developed a waterproof sound source detection system that can accurately estimate outdoor low-frequencynoise sources. As the main component, the waterproof microphone, a simple structure is proposed in which the back of the MEMS microphone is filled with potting material. This ensures proper acoustic and waterproof performance by suppressing resonances in the microphone array. A planar array with 32 microphones was constructed using a waterproof microphone array structure and a deployment mechanism was introduced to allow expansion of the microphone array which improves estimation accuracy in the low frequency range. The high-resolution MUltiple SIgnal Classification (MUSIC) method was used to estimate the direction of arrival. Performance evaluations in an anechoic chamber confirmed that the accuracy of the estimation of low-frequency noise sources in the 125-Hz band was improved. Furthermore, real-environment measurements in a factory demonstrated that the 100-Hz dominant noise source could be accurately identified.