Performance of the space-time AR filter in non-homogeneous clutter [airborne radar STAP]

The space-time autoregressive (STAR) technique has been proposed for space-time adaptive processing (STAP). This algorithm utilizes the structure inherent in airborne STAP applications (i.e., uniformly spaced pulses) to improve convergence and reduce computation. This paper addresses the performance of the STAR algorithm in the presence of non-homogeneous clutter, particularly in a littoral region. It is shown that lite STAR filter produced a narrower (and thus a lower minimum detectable velocity) than either the optimized pre-Doppler or post-Doppler algorithms. This is due to file STAR algorithm estimating a joint space-time filter for clutter suppression as well as having a small sample support region. It is also shown that the pre-Doppler algorithm does exhibit some robustness when the radar system is limited by its channel match.