Joint 2D-DOD, 2D-DOA, and Polarization Angles Estimation for Bistatic EMVS-MIMO Radar Via PARAFAC Analysis.
Fangqing Wen,Junpeng Shi,Zijing Zhang
DOI: https://doi.org/10.1109/tvt.2019.2957511
IF: 6.8
2019-01-01
IEEE Transactions on Vehicular Technology
Abstract:Multiple-input multiple-output (MIMO) is a technical hotspot in physical layer with numerous applications in wireless communications, radars, sonars, and well beyond. In this paper, we focus on the multi-dimensional angle estimation problem in a bistatic electromagnetic vector sensors (EMVS) MIMO system. Namely, we need to simultaneously estimate two-dimensional (2D) direction-of-arrival (DOA), 2D direction-of-departure (DOD), 2D receive polarization angle (RPA) and 2D transmit polarization angle (TPA). To tackle this issue, a parallel factor (PARAFAC) analysis-based estimator is proposed. Firstly, a third-order PARAFAC analysis data model is established, which can efficiently exploit the tensor structure of the array measurement. After performing PARAFAC decomposition on the tensor measurement, the factor matrices are achieved. By combining the estimation method of signal parameters via rotational invariance technique (ESPRIT) with the vector cross-product method, joint estimates of 2D-DOD, 2D-DOA, 2D-TPA and 2D-RPA are obtained without further pairing calculation. Compared with the state-of-the-art ESPRIT-Like approach, the proposed method can achieve better performance by enforcing the third-order structure information, and it is suitable for arbitrary array manifolds. Theoretical analyses are given and numerical results corroborate our analysis.