An Alternating Proximal Method for Sea Clutter Suppression in Over-the-Horizon Radar Systems

The over-the-horizon (OTH) radar systems are seriously affected by sea clutter. The precise estimation of sea clutter covariance (SCC) plays an important role in suppressing the sea clutter, especially in the case of limited samples. Toward this end, this letter proposes a robust dictionary learning (DL) based on the SCC reconstruction method to suppress sea clutter with a single observation sample. The DL procedure is formulated as a nonconvex optimization problem, which is known to be NP-hard. To address this problem and reconstruct the SCC effectively, we propose a novel alternating proximal (AltProx) method designed for solving
$\ell _{0}$
-norm-based problems. Extensive analyses demonstrate that our proposed algorithm exhibits global convergence with a sublinear convergence rate. Through simulation results, we verify the effectiveness of our proposed algorithm, demonstrating a notable performance improvement compared with conventional methods using single samples. In addition, we conduct experiments using real sea clutter datasets, further demonstrating the practical applicability of our method.