Iteratively Reweighted Spherical Equivalent Source Method for Acoustic Source Identification

• Published in: IEEE access Vol. 7; pp. 51513 - 51521
• Main Authors: Ping, Guoli, Chu, Zhigang, Yang, Yang, Chen, Xu
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2019; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Acoustic arrays; Acoustic measurements; Acoustic source identification; Acoustics; Algorithms; Convex functions; Equivalence; Frequency measurement; Microphone arrays; near-field acoustical holography; Performance enhancement; rigid spherical microphone arrays; Sound sources; spherical equivalent source method
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2019.2911857

Spherical equivalent source method (S-ESM) using rigid spherical microphone arrays can simultaneously identify sound sources in all directions. In this paper, based on the reweighting and sparse representation frameworks, the sparsity-promoting iteratively reweighted least squares (IRLS) and reweighted ℓ 1 -norm minimization (referred as w-ℓ 1 -norm) are exploited to improve the performance of acoustic source identification for S-ESM. The numerical and experimental results indicate accurate acoustic source identification for the two iteratively reweighted algorithms. IRLS can provide good acoustic source identification over the wide frequency and measurement distance ranges, improving the performance of the established S-ESM. In addition, w-ℓ 1 -norm is also an alternative solution strategy for S-ESM, although at the expense of low computational efficiency and given prior information.