Fourth-order cumulant structure forcing: application to blind array processing

Abstract

In blind array processing, the array manifold is unknown but, under the signal independence assumption, the signal parameters can be estimated by recourse to higher-order information. The author considers the 4th-order cumulant tensor and shows that sample cumulant enhancement based on rank and symmetry properties yield cumulant estimates with exact theoretical structure. Any identification procedure based on enhancement cumulants is then equivalent to cumulant matching, bypassing the need for initialization and optimization.<>