An Improved Two-Stage Mixed Source Localization Algorithm in the Presence of Unknown Mutual Coupling

An algorithm for mixed far-field (FF) and near-field (NF) source localization is proposed with uniform linear arrays (ULAs) under unknown mutual coupling effects in this paper. The directions-of-arrival (DOAs) of FF sources are first estimated with an improved rank reduction (RARE) method based on the minimum eigenvalue, followed by the calculation and compensation of mutual coupling coefficients. The FF source covariance matrix is then reconstructed and subtracted. Subsequently, the DOAs of NF sources are estimated via the rotational invariance technique, with their corresponding ranges determined through one-dimensional spectral search while maintaining automatic parameter pairing. The proposed approach enhances estimation accuracy by improved FF DOA estimation method and ensures efficient aperture utilization by avoiding subspace dimension expansion during source classification. Simulation results demonstrate the algorithm's superior performance in both source classification and localization.