Tensor Methods for DoD and DoA Estimation for Bistatic MIMO RADAR in VANET

Accurate localization is essential in Vehicle-to-Everything (V2X) systems to ensure reliable autonomous navigation and effective traffic coordination. Conventional terrestrial localization techniques often suffer from reduced accuracy due to multipath propagation, interference, and positioning uncertainties inherent in urban environments. This work proposes a bistatic Multiple-Input Multiple-Output (MIMO) V2X communication architecture designed to address these 1 challenges. In the proposed architecture, Roadside Units (RSUs) transmit local-ization signals that interact with autonomous vehicles in a bistatic sensing configuration. A pre-processing approach is introduced to improve the estimation of the direction of departure (DoD) and direction of arrival (DoA), as well as a low-complexity tensor-based approach. The proposed approaches efficiently explore the multidimensional tensor structure of the received signals, providing superior target localization accuracy. Simulation results validate the effectiveness of the proposed approach, demonstrating notable improvements in localiza-tion performance and robustness in noisy environments with low computational overhead.