A generalized minimax-concave penalty based compressive beamforming method for acoustic source identification

Abstract

Acoustic source identification using compressive beamforming has been extensively investigated and applied in numerous fields. In this paper, a generalized minimax-concave (GMC) penalty based compressive beamforming method is proposed. The method solves acoustic inverse problems using the GMC penalty, by which the capacity of sound field reconstruction in low-frequency and low-signal-to-noise ratio environments can be enhanced. In addition, the problem of systematic underestimation of source strength can be relieved with the benefit of penalty homogeneity, and a balance can be reached between the sparsity-induced capacity and computing complexity. Moreover, the convergence of the proposed method and parameter influences were studied through numerical simulations. The results show that the method is robust within a relatively wide range of parameters, and the computational efficiency is within the limits of acceptability. Furthermore, the performance of the proposed method was compared with two other methods using simulated and experimental data. The results indicate that the proposed method performs better than existing algorithms under low-frequency and strong-noise conditions, and the spatial resolution in acoustic maps and the capacity of source strength estimation can be improved.