Weak target detection by multiscale space-time adaptive processing

STAP is a recent popular research field in which a multichannel and multipulse radar system is utilized to enhance the performance of radar detection under severe clutter and jammer conditions. Real-time STAP requires an intelligent lower dimensional processor realizable from a computational point of view. Several "reduced-dimension (or partially adaptive) STAP" techniques have been introduced to represent the noise sub-space and reduce the dimension of the weight vector. Beside these approaches, we introduce a new technique which is based on modeling the radar clutter as a multiscale stochastic process. Our approach is shown to be capable of outperforming the popular partially adaptive STAP algorithms although the computational complexity is significantly reduced.