Space-Time Adaptive Processing by Employing Structure-Aware Two-Level Block Sparsity

Abstract

Traditional radar space-time adaptive processing (STAP) cannot efficiently suppress heterogeneous clutter because of a small number of independent and identically distributed training snapshots. In the article, we propose a new STAP approach exploiting structure-aware two-level block sparsity (STBS) of radar echoes, namely STBS-STAP. It enhances the performance on clutter suppression and target detection with limited training snapshots. The clutter angle-Doppler profile always appears in a continuous diagonal clustering structure and the radar echoes at the adjacent range cells commonly share the same sparse pattern. STBS-STAP employs STBS, i.e., both the diagonal clustering structure and the common sparsity property, to acquire a precise clutter covariance matrix estimation. Thus, the new STBS-STAP achieves better performance on clutter suppression compared with existing STAP methods with a small number of training samples. Besides, STBS-STAP achieves superior target detection performance due to the precise estimation of the statistical properties of the clutter. The superiority of STBS-STAP is verified by experiments on both simulated data and measured Mountain-Top data.