JCMF: A Novel Wideband DOA Estimator

Covariance matrix fitting is a classic problem in direction-of-arrival (DOA) estimation, but there has been little discussion about wideband signals. In this paper, a novel wideband DOA estimation method based on covariance matrix fitting is developed. First of all, a novel wideband signal model is proposed and the wideband joint covariance matrix is formulated. The fitting criterion is deduced by the statistical characteristics of this model. Then the optimization problem of the criterion is transformed into a block sparsity problem using Holder's inequality. The problem can be divided into two subproblems and the closed-form solution of each part is obtained by dual method. The result of our wideband estimator can be solved by alternating iteration. Finally, the Cramer-Rao bound (CRB) is derived for wideband signals. Compared with other algorithms, the presented estimator has higher accuracy. Simulation results show the effectiveness and feasibility of the proposed algorithm.