DOA Estimation with a Combination of Iterative MUSIC and SBL in Low SNR Conditions

Abstract

This paper proposes a two-stage direction-of-arrival (DOA) estimation method combining Multiple Signal Classification (MUSIC) with Sparse Bayesian Learning (SBL) algorithm. In the first stage, we take advantage of the divergence of noise power spectrum and have designed an iterative MUSIC method. Performing MUSIC iteratively, we can obtain an angle set, including all estimates of DOA. Then, we transfer the angle set to the next stage to pick the correct values out. In the second stage, the angle set acquired from the iterative MUSIC is utilized to constitute the dictionary matrix used in SBL, which can decrease the computational complexity effectively. Simulations indicate that the proposed method can achieve high estimation accuracy even when the Signal-to-Noise Ratio(SNR) is below 0dB with a DOA gap of 3 degrees, outperforming the conventional MUSIC and SBL algorithm.