Identification of Multiple Dipole Sound Sources: Maximum Likelihood, Iterative Beamforming, and Source Number Estimation

The present paper investigates the spatial array imaging of multiple dipole sound sources. In this paper, a method for estimating the positions, angles, and energies of dipole sound sources based on the maximum likelihood framework is proposed. To address the high parameter dimensionality issue in the optimization problem for the traditional maximum likelihood methods, this paper introduces latent variables representing the contributions of the individual sources and iteratively estimates model parameters via the Expectation-Maximization (EM) algorithm. In this paper, we conducted array imaging experiments using multiple dipole signal generators, demonstrating that the proposed algorithm has advantages over existing methods such as conventional beamforming, Capon, MUSIC, dipole beamforming including the super-resolution, the wider working frequency range, and the ability to estimate the number of sound sources.