A parametric approach to hot clutter cancellation

Many reduced dimension STAP algorithms have been developed for airborne radar applications which rely on a stationary Doppler component of the interference in order to maintain acceptable performance. Two cases where this assumption is violated are when the ground clutter contains intrinsic clutter motion (ICM) and when hot clutter is present. In addition to the non-stationary Doppler component, hot clutter contains non-zero correlations in fast-time (across range bins) as well. This paper presents an algorithm designed to mitigate both ground clutter and hot clutter in the same step using a two dimensional vector autoregressive model to whiten the data in space, fast-time, and slow-time. This is an extension of the space-time autoregressive (STAR) filter that we have previously proposed. Using a simulated data set for circular array STAP augmented with synthetic hot clutter, we demonstrate that the extensions we present do result in a significant performance increase over the standard STAR filter. In addition we also show that the STAR filters have a narrower clutter notch than the optimized pre-Doppler filter when a finite sample support is used to train the filters