Generalized covariance-based ESPRIT-like solution to direction of arrival estimation for strictly non-circular signals under Alpha-stable distributed noise

Direction of arrival (DOA) estimation with high resolution and low computational complexity has been one essential research topic. Meanwhile, many solutions have been proposed to handle parameter estimation problems for strictly non-circular (NC) signals. However, second-order statistics-based solut...
Ausführliche Beschreibung

Direction of arrival (DOA) estimation with high resolution and low computational complexity has been one essential research topic. Meanwhile, many solutions have been proposed to handle parameter estimation problems for strictly non-circular (NC) signals. However, second-order statistics-based solutions usually degrade under fractional lower-order Alpha-stable distributed noise. To solve this problem, three concepts are first defined, and then related properties are proved, including a specific generalized covariance (GC), the non-circularity based on GC, and the extended GC matrix. Based on these concepts and theorems, a novel ESPRIT-like method is derived especially for strictly non-circular sources and termed NC-GC-ESPRIT. Monte-Carlo simulations under four different experimental conditions are executed, involving eleven state-of-the-art DOA approaches as comparison candidates. Thus, the superior performance of the proposed solution is validated through the comparison with these candidate algorithms. Ausführliche Beschreibung