Techniques to screen for moving and large stationary discretes in space-time adaptive estimation data

When the data samples used to estimate the interference covariance matrix in Space-Time Adaptive Processing (STAP) are assumed to be independent and identically distributed, an estimation loss indirectly proportional to the number of samples will occur. Measured data has shown that data samples are often nonhomogeneous and that these nonhomogeneities can cause losses beyond that from the estimation alone. Nonhomogeneities are of various types; here we have focused on moving and large stationary discretes. Numerical examples are presented that give a rationale for a screening technique. Then we present related screening techniques and performance simulations for the detection of moving and large stationary discretes in the estimation data. The screening techniques consist of pre-processing the samples by transforming them from the space-time domain to the angle-Doppler domain and exploiting the resulting interference characteristics