Higher-order unitary propagator method for 2D-DOA estimation of non-circular sources via uniform rectangular array

Abstract

In this paper, we propose a higher-order unitary propagator method (HO-UPM) to estimate two dimensional (2D) direction-of-arrival (DOA) of Non-Circular (NC) sources via a uniform rectangular array (URA) of antennas. The proposed method benefits from the jointly augmented multidimensional structure of measurement data and enhances the estimation accuracy by exploiting the NC property with an additional advantage of inexpensive computations as compared to the higher-order singular value decomposition (HOSVD). Though the joint mode augmentation brings redundancy in the resulting NC manifold tensor, it still doubles the antenna count virtually and yields a larger URA that acts as a 2D separable spatial sampling grid retaining the centro-symmetry and translational invariance. Therefore, the proposed approach not only improves the model identifiability but also avoids individual augmentation and processing of each mode of the measurement tensor. Numerical simulations displaying accuracy of subspace estimation, computational cost and root mean square error (RMSE) performance along-with the deterministic NC Cramer-Rao bound (CRB) are also included to verify and compare the effectiveness of proposed method with the existing matrix and tensor-based approaches.