Sound source identification based on functional generalized inverse beamforming

Abstract

Generalized inverse beamforming is a kind of highly effective and widely used sound source identification technology based on the microphone array. However, the performance of traditional generalized inverse beamforming is still limited by its poor side-lobe suppression capability and low dynamic range when identifying complex sound sources. It is difficult to achieve high accuracy of sound source identification. To improve the dynamic level of the generalized inverse beamforming, a new identification method based on functional beamforming is presented, which is named functional generalized inverse beamforming. The basic theory of functional generalized inverse beamforming is introduced by discussing the conventional generalized inverse beamforming and matrix functions. The influences of the source type and measurement error on the performance of sound source maps have been investigated. The range of typical order is determined by the analysis results. The results of numerical simulations and experiments show that the proposed algorithm can ensure the accurate identification of sound source intensity and location, and it also remarkably suppresses beamforming sidelobes by increasing the value of order. Therefore, the proposed algorithm can lead to higher spatial resolution and more accurate sound source localization compared with the conventional generalized inverse beamforming.