ESPRIT-Based Coherent Source Localization With Forward and Backward Vectors

Novel methods are presented which estimate the direction of arrival (DOA) of coherent signals by efficiently exploiting forward and backward vectors. Some correlation vectors or the eigenvector corresponding to the largest eigenvalue of the covariance matrix is used as the forward vector, and the backward vector is then obtained by reversing the order of the complex conjugated elements of the forward one. The use of both forward and backward vectors allows us to find a signal subspace of larger dimension than in the existing method which obtains that from a Toeplitz matrix. The proposed methods with an increased effective array aperture resolve more signals and provide better DOA estimates than the Toeplitz based method.