Wideband DOA estimation in spherical harmonic domain using sparse Bayesian learning

This paper focuses on the direction-of-arrival (DOA) estimation of wideband signals using a spherical array which has many advantages in 3D space. We propose two DOA estimation methods in spherical harmonic domain. First of all, spherical Fourier transform (SFT) is used to construct the array signal model. The first method exploits the joint sparsity of all the frequency bins and the second method transforms multiple frequency information into only one measure by high order singular value decomposition (HOSVD). Sparse Bayesian learning (SBL) is then utilized to estimate the DOAs. Simulation results verify the effectiveness of the proposed methods.