A low-complexity RARE-based 2-D DOA estimation algorithm for a mixture of circular and strictly noncircular sources

Abstract

A new rank reduction (RARE)-based two-dimensional (2-D) direction of arrival (DOA) estimation algorithmis proposed considering a mixture of circular and strictly noncircular sources. To enhance array aperture, a geometry ofthree uniform linear arrays is considered and then treated as displaced arrays from a virtual array using a simple lineartransformation. The received data and the conjugated counterpart are combined together, exploiting the noncircularproperty. Both sources can be estimated separately by designing and exploiting the distinctive nature of circular andnoncircular steering vectors. However, a 2-D spectrum search would lead to a high computational complexity burden.To reduce this high computational complexity burden, a novel RARE-based method is proposed, which plays a vital roleby decomposing 2-D observation space into two successive 1-D peak search functions. The proposed method has somedistinctive advantages: it can enhance the array aperture utilization, it can provide better estimation accuracy whenmixed sources are greater than the number of sensors, it can estimate a larger number of mixed sources than the numberof sensors, and finally it can automatically pair 2-D DOAs without any complicated pairing formulation. Extensivesimulation results are provided to demonstrate the effectiveness of the proposed method.