Closed-form eigenstructure-based direction finding using arbitrary but identical subarrays on a sparse uniform rectangular array grid

The use of sparsely and uniformly spaced rectangular arrays in arrival angle estimation leads to cyclic ambiguity in the Cartesian direction cosine estimates due to the spatial Nyquist sampling theorem. Several recent papers disambiguate this cyclic ambiguity by populating the sparse regular array grid with special subarrays that capture, from the impinging wavefield, information complementary to that contained in the cyclically ambiguous direction cosine estimates. Examples of such subarrays include the six-component electromagnetic vector sensor, the underwater acoustic particle velocity hydrophone, or some distinctively spaced cluster of identical scalar sensors. This paper proposes an improved disambiguation algorithm for the aforementioned array setup for enhanced disambiguation accuracy, array sparsity, subarray conflgurational flexibility, and computational efficiency. This new disambiguation algorithm involves a computationally efficient ESPRITbased step narrowing the continuous support range of the unknown parameters to a small finite set of candidate estimates from which the more accurate estimation methods of MUSIC or MODE is used to identify the best estimation candidate. This proposed scheme is applicable to any general subarray configuration wherein the overall array manifold is one-to-one related to the incident source's Cartesian direction cosines.