Alternating projection for maximum-likelihood source localization using eigendecomposition

Alternating projection allows us to make a maximum-likelihood estimation of arrival angles by one-dimensional search. In this letter, we propose an eigendecomposition-based alternating projection method that can greatly reduce computational complexity with no performance degradation as compared with the direct alternating projection. In the proposed method, initial estimation can be readily performed with MUSIC, which may lead to further reduction in the computational burden.