A subspace method for space time adaptive processing

The problem of space-time adaptive processing (STAP) using a nonlinear array is considered. A key part of STAP is the estimation of the space-time covariance matrix of the received data. The conventional method of doing this causes significant performance degradation at short ranges because of the nonstationarity of the data. We present an alternative algorithm which circumvents this problem by projecting the data on the subspace orthogonal to the clutter and jammer subspaces. The clutter subspace is computed from the known array manifold, while the jammer subspace is estimated from clutter-free measurements. Numerical examples illustrate the performance improvement achieved at short ranges.