DOA estimation for large array with nonuniform spacing based on sparse representation

Abstract

The problem of grating lobes false alarm is easy to occur in Direction of Arrival (DoA) estimation of large array with nonuniform spacing. The sparse spatial signal reconstruction can be used to suppress the grating lobes. This paper applied the Sparse Bayesian Learning (SBL) algorithm in DOA estimation for large spacing array. To solve the steering vector mismatch problem, the concept of peak confidence is proposed. Utilizing the confidence function to screen the peaks, the false alarms of grating lobes are avoided effectively. The effect of different input k value on the grating lobes suppression performance is analyzed, and the threshold selection in practical application is given. Simulation and sea trial data results confirm the grating lobes suppression performance of the proposed algorithm.