A broadband maximum likelihood imager for a class of extended space-time separable sources

An algorithm is presented that reconstructs the shape of an extended incohe rent source using only the broadband signals radiated from the source to a sparse array. The source is modeled with a small set of parameters. For the specific class of space-time separable sources, which have similar spatial structure at multiple frequencies, the information radiated at each freque ncy is recombined into a broadband likelihood function in order to improve the estimation of the source parameters. Expressions for the Cramer-Rao bou nds (CRB's) are provided, and the performance of the algorithm is demonstra ted for a "Gaussian lump" shaped source.