A hybrid wave superposition method based on particle filter

Abstract

In the field of acoustic holography, the wave superposition method (WSM) is one of the classical and widely studied algorithms. To improve the accuracy of sound field reconstruction, a hybrid wave superposition method based on particle filter (PF-HWSM) is proposed. The PF-HWSM mainly includes three steps: (1) calculate the initial source virtual trajectory by 3D-BMUSIC with a rotating microphone array; (2) optimize the virtual trajectory and improve the localization accuracy by the particle filter; (3) configure the equivalent sources according to the optimized localization, and reconstruct the sound field by wave superposition method. Two types of simulations are used to provide a systematic parameter analysis of the traditional WSM-based methods and PF-HWSM, and the comparisons prove the superiority of the proposed method. Finally, two types of experiments in an anechoic room verify that the proposed method achieves high-precision reconstruction. The PF-HWSM provides a method for the equivalent sources configuration of WSM, which can effectively improve the accuracy of sound field reconstruction.