Localization and identification of three-dimensional sound source with beamforming based acoustic tomography

Abstract

Beamforming based commercial planar microphone array could only localize and identify the sound source when the distance between source and array is known. This paper presents a beamforming based acoustic tomography (BBAT) method to locate and identify the source in 3D space, say, the BBAT method can not only locate the source in X-Y plane that is parallel to the array, but also the depth Z of the source. In this method, the sound field is reconstructed on the virtual planes at different distances along depth direction (Z direction). The maximum response of sound field on every virtual reconstruction plane is tracked, where the largest value among those maximum responses appears at Z direction is the depth of the source. The location of source at X and Y directions can then be easily identified based on beamforming principle, which is utilized by commercial planar array. The BBAT method is evaluated theoretically by simulation of monopole source, the experimental evaluation is done as well in anechoic chamber. The results from both simulation and experiment indicate that this method is capable to locate and identify sound source in 3D space. However, it cannot recognize the sound source located in front of the planar array or behind because of the genetic limitation of 2D planar array in identification of source depth.