Reduced-rank space-time adaptive processing using a modified projection approximation subspace tracking deflation approach

The projection approximation subspace tracking deflation (PASTd) algorithm is investigated in the context of the reduced-rank space-time adaptive processing (RR-STAP), where a recursive approach is employed to estimate the clutter subspace. Instead of a direct use of the classic PASTd algorithm in the RR-STAP, which suffers from a slow convergence rate because of the sequential tracking of multiple clutter eigenvectors, a modified PASTd adapted to the STAP is presented. In comparison with the conventional eigenvalue decomposition approach, the presented algorithm is computationally much more efficient and is also able to achieve comparable convergence effectiveness. The presented methodology is validated by the Monte Carlo simulation.