Range dependent clutter rejection using range-recursive space-time adaptive processing (STAP) algorithms

We here address the issue of ground clutter rejection for the detection of slowly moving targets in a non-side looking (NSL) array configuration airborne radar. The optimum space-time adaptive processing (STAP) filter needs the knowledge of the inverse of the space-time covariance matrix. In practice, it is unknown and has to be estimated. The most popular approximated method is the sample matrix inversion (SMI) method which consists in inverting the covariance matrix estimated by an average of the sample matrix over the secondary range cells. This estimator is unbiased in case of i.i.d. data. In an NSL configuration, the clutter power spectrum is range dependent and the data are consequently not i.i.d. We here present a solution to mitigate this range dependency of the data: the range recursive subspace-based algorithms. They are used in two architectures: a fully and a partially adaptive ones. Then a new range-recursive algorithm using Taylor series expansion is investigated. The performance of these algorithms are compared with that of the conventional STAP algorithms in term of SINR loss.