Coherence aspects of GMTI using MSAR

Previous work is briefly reviewed. It is shown that when a ground moving target is highly coherent an optimum steering vector for linearly combining images of slowly moving targets in sets of images produced by a Multi-channel Synthetic Aperture Radar (MSAR) is simply a discrete Fourier transform. That work is then extended in this paper to show that the assumption of perfect target coherence over the timescale of the MSAR image set is robust to the non-optimum processing resulting from target phase noise caused, for example, by motion on a rough surface.