Linear constraints in pre-Doppler STAP processing

This paper addresses the issue of employing space-time adaptive processing (STAP) prior to Doppler filtering in radar systems. When STAP beamformer processing is applied to spatial/temporal samples that include successive radar pulses, the adaptive weights can cause modulation (spreading) of the desired target Doppler. In this paper, a linearly constrained adaptive beamformer is proposed that ensures pre-Doppler adaptive processing will not degrade desired signal coherence. A formulation of the processor is presented and its properties described. This formulation is first given for the case of slow-time STAP and is then extended to include processors which use both fast and slow-time samples. Examples showing the application of the proposed structure to recorded data are used to illustrate its performance. Comparisons are made with adaptive systems that do not employ constraints to illustrate the advantages of the proposed system. An extension to the full STAP system which employs time taps in both range and pulse number is described.