Robust Method for Detecting Target-like Signals in Training Samples in STAP

Abstract

Heterogeneous environment has a severely deteriorative impact on the performance of STAP, among which target-like signals in training samples (TTS) always tend to be the most serious. Traditional detectors for TSS such as generalized inner product (GIP) and adaptive power residue (APR) employ sample covariance matrix to estimate the interference covariance matrix in the cell under test (CUT), which may suffer large performance degradation in heterogeneous environments due to estimation error. This paper proposes two modified detectors based on subaperture smoothing technique which only exploits the sample in CUT to estimate the interference covariance matrix. The proposed method can estimate the interference covariance matrix more accurately and identify TSS more effectively. Simulation and measured results confirm the effectiveness and robustness of the proposed method.