A grid-less direction of arrival (DOA) estimation algorithm via covariance matrix recovery is developed to improve the DOA estimation performance under low signal-to-noise ratio (SNR) for coherent and non-coherent sources. Firstly, we propose to recover the covariance matrix by utilizing both the Hermitian Toeplitz structure of the covariance matrix and the noise statistical characters in the multiple measurement vectors (MMVs). An efficient grid-less DOA estimation algorithm based on the covariance matrix’s first column is also presented to reduce grid mismatch effects. The simulation results show that the proposed DOA estimation algorithm can better recover the covariance matrix than the covariance fitting algorithms with non-coherent and coherent sources. Besides, simulations also verify that the proposed algorithm has a smaller root mean square error (RMSE) and higher estimation probability than the state-of-the-art algorithms in low SNR and small angle separation.

开发了一种通过协方差矩阵恢复的无网格到达方向（DOA）估计算法，以提高相干和非相干源在低信噪比（SNR）下的DOA估计性能。首先，我们建议通过利用协方差矩阵的Hermitian Toeplitz结构和多个测量向量（MMV）中的噪声统计特征来恢复协方差矩阵。还提出了一种基于协方差矩阵第一列的高效无网格DOA估计算法，以减少网格不匹配的影响。仿真结果表明，与非相干和相干源的协方差拟合算法相比，提出的DOA估计算法能够更好地恢复协方差矩阵。除了，