Off-grid DOA Estimation for Temporally Correlated Source via Robust Block-SBL in Mutual Coupling

In the field of array signal parameter estimation, the research on direction-of-arrival estimation (DOA) by using the sparse Bayesian learning (SBL) technique has always been a very important aspect. Most of the existing SBL-based methods are not consider the temporal correlation between snapshots, which is not consistent with the real environment. Hence, the block sparse Bayesian learning (Block-SBL) has recently received a lot of attention. To enhance the robustness of the traditional Block-SBL based method to mutual coupling (MC), a robust Block-SBL method is proposed to achieve off-grid DOA estimation for temporally correlated source in this paper. Firstly, in the proposed method, a linear transformation is conducted based on the banded complex symmetric Toeplitz structure of the mutual coupling matrix (MCM) to eliminate the influence of MC between the array elements. Then, in order to improve the estimation performance, the signal subspace fitting (SSF) technique is introduced to construct an equivalent signal model, which is realized by eigenvalue decomposition of the received signal covariance matrix. Based on the constructed equivalent signal model, the SBL process is finally utilized to estimate parameters. In the process of the estimation of parameters, the equivalent signal variance is estimation by expectation maximization (EM) method and the off-grid error is reduced by the dynamic updating of spatial discrete grid points, which is realized by finding the roots of a polynomial. The simulation results demonstrate that the proposed method is more robust to the off-grid error and array mutual coupling, and can provide better DOA estimation performance than the traditional SBL and Block-SBL method.