Unitary-ESPRIT for 2D-DOD and 2D-DOA Estimation with EMV Sensors in MIMO Radar

This paper introduces a unitary ESPRIT algorithm for jointly estimating the two-dimensional directions of departure (2D-DOD) and 2D directions of arrival (2D-DOA) of coherent or uncorrelated targets in MIMO radar systems, using uniform rectangular arrays of electromagnetic vector (EMV) sensors. The method employs real-valued operations for computing the covariance matrix and its eigenvalue decomposition, making it computationally inexpensive. The resulting covariance matrix inherently exhibits polarization smoothing and detects up to 72 coherent targets without the need for preprocessing techniques. Performance is evaluated through simulations and verified with the Cramér-Rao bound, as well as compared with existing ESPRIT-like and Propagator-based methods.