Iterative pre-filtering over non-uniform sampling grids

The author presents an extension of the Steiglitz-McBride technique of iterative prefiltering to the situation in which the input signal is sampled over a nonuniform grid. The extension amounts to incorporating an interpolation step within the iteration to estimate values over a larger uniform grid that contains the nonuniform grid as a subset. Then a least-squares filtering problem can be solved for the pole-zero model needed to generate the next iteration. This technique has two applications for array processing: the first is the case of a nonuniform sensor array, and the second is that of missing or inoperative sensors within a uniform array