ANM-Based Source Localization Under Mixed Field

Abstract

This paper proposes a mixed field source localization algorithm based on atomic norm minimization (ANM) under low snapshots signal. It presents the first application of atomic norm theory to source localization algorithms. A general mixed-field steering vector is then proposed. The covariance matrix based on vectorization is analyzed. A matrix-based atomic set is formulated to represent the set of possible sources, which allows the transformation of the source localization problem into a two-dimensional ANM problem. This ANM problem is subsequently converted into a semidefinite programming (SDP) problem. The optimal solution determined from the SDP problem serves as the process variables. Through pairing and parameter transformation techniques, the algorithm determines the direction of arrival (DOA) and range parameters. The feasibility of the proposed algorithm is finally demonstrated through numerical simulations.