Abstract:

Performance of space-time adaptive processing (STAP) is always corrupted by outliers. Hence, a novel robust STAP algorithm based on the S transform (ST) is proposed. It exploits the characteristic that ST spectrum distribution of the outlier is different from the clutter in the time-frequency domain. Firstly, the fast time sequence is transferred to the time-frequency matrix by ST. Some high frequency points of the ST matrix are abandoned to eliminate outliers. Then the remainder of the ST matrix is used to estimate the clutter covariance matrix and filter weight. Monte Carlo experiments prove that the proposed algorithm is more robust than other conventional STAP algorithms in non-homogeneous clutter environments.