Improvement of Exhaustive Search Based Maximum-Likelihood DOA Estimation

Direction-of-arrival (DOA) estimation by maximum likelihood (ML) algorithm is essentially an optimization of nonlinear function. In nonlinear optimization, since the final estimate is highly dependent on the initial estimate, an initialization is critical. Alternating Projection (AP) initialization has been proposed as a computationally efficient method for the initialization of the ML DOA cost function. In this paper, we propose a multidimensional (M-D) search scheme of uniform exhaustive search and improved exhaustive search. In the improved exhaustive search algorithm, the two-step procedure is applied to reduce the computational load of the uniform exhaustive search initialization scheme. It is shown in numerical results that the performance of improved exhaustive search is as good as that of the uniform exhaustive search and that improved exhaustive search is computationally much less intensive than uniform exhaustive search.