Multistatic radar change detection using a sparse imaging approach

This paper describes a sparse imaging approach for estimating change images from a constellation of multistatic radar. In our setup, radar antennas are arranged around the perimeter of a surveillance region. This provides large angular diversity but a very small angular sampling. Conventional backprojection imaging techniques applied to this data produce sidelobes which severely limit the utility of the imagery. We describe an innovative change imaging method which enforces sparseness on the estimated image. The method is illustrated with collected multistatic radar data, showing that the sparseness model produces excellent images with very limited sampling of the aperture.