A De-Noising 2-D-DOA Estimation Method for Uniform Rectangle Array

This letter tackles the problem of 2-D direction-of-arrival (2-D-DOA) estimation for uniform rectangle array (URA) in the presence of unknown nonuniform noise. By exploiting the diagonal characteristic of the noise covariance matrix, a reduced covariance tensor is formulated, which links the 2-D-DOA estimation problem to the quadrilinear decomposition model, and an alternating least squares strategy is developed to obtain the 2-D-DOA estimates. Furthermore, the stochastic Cramer-Rao bound of 2-D-DOA estimation for URA is derived. The proposed method does not require eigendecomposition and can achieve automatically paired 2-D-DOA estimates. Numerical experiments demonstrate the effectiveness of the proposed method.