Airborne radar space time adaptive processing based on atomic norm minimization

Abstract

Existing sparse recovery based space-time adaptive processing (SR-STAP) methods discretize the angle-Doppler plane to generate the space-time steering dictionary, which will cause the off-grid problem, resulting in performance loss. This paper proposes an alternative processing model established in the continuous domain. Based on the positive semidefinite (PSD), block-Toeplitz, and low rank properties of the clutter covariance matrix (CCM), the subspace of clutter can be estimated by solving an atomic norm minimization problem. Then, the CCM is directly calculated by the Eigen-decomposition based process. If multiple independently and identically distributed (IID) training samples are available, the proposed method can be easily extended to the multiple measurement vector (MMV) model. With the joint sparsity and same subspace assumptions of different samples, MMV based atomic norm minimization STAP (ANM-STAP) method can further increase the estimation accuracy of the clutter subspace, and thus improve the clutter suppression performance. Simulation results demonstrate that, in comparison with typical SR-STAP methods, the proposed method can avoid the off-grid problem, achieve more accurate CCM estimation, and enjoy better clutter suppression performance with fewer training samples than the statistic STAP method.