Super-Resolution Localization of Low-Frequency Sound Sources by Beamforming with Singular Value Decomposition

Abstract

The spatial resolution of beamforming is constrained by the wavelength-order diffraction limit, rendering it less effective for pinpointing low-frequency sound sources. To achieve subwavelength focusing, the focusing field of beamforming is initially simplified to a product of the column vector of the received signals and the transfer matrix between the measurement and focusing planes. By applying singular value decomposition to this transfer matrix, the focusing field can incorporate more evanescent waves by adjusting the weights of its singular values. Both numerical simulations and experimental results are performed to validate the proposed method. Compared to conventional beamforming, the proposed method significantly enhances spatial resolution at low frequencies and yields impressive results even in half-space environments with fewer microphones.