Real-valued sparse Bayesian learning algorithm for off-grid DOA estimation in the beamspace

Abstract

In recent years, sparse Bayesian learning (SBL) algorithms in the array element space have become very popular for off-grid direction-of-arrival (DOA) estimation. However, the SBL algorithm in beamspace has not been researched. Therefore, this paper proposes a real-valued SBL algorithm in the beamspace. First, the received data in the array element space is transformed to the beamspace using the receive beamspace matrix. Subsequently, singular value decomposition (SVD) is introduced to reduce the computational burden further. Finally, the SBL algorithm is combined to estimate the off-grid DOA. Compared with the SBL algorithms in the array element space, the proposed algorithm has less computational complexity, better robustness to grid interval, and better DOA estimation performance under certain conditions. In addition, we also show the DOA estimation performance of the proposed algorithm for correlated signals and large-scale arrays. Numerical simulations verify the effectiveness of the proposed algorithm.