Gridless Angle and Range Estimation for FDA-MIMO Radar Based on Decoupled Atomic Norm Minimization

Abstract

The frequency diverse array (FDA), which exploits a small frequency increment across the transmitting antenna elements, has been recently introduced to multiple-input- multiple-output (MIMO) radar to offer the capability of range-angle-dependent beampattern and target localization. Traditionally, the two-dimensional (2D) subspace-based methods and grid-based sparse reconstruction algorithms are used to obtain angle and range in FDA-MIMO radar. In this paper, with a single measurement, we propose a gridless angle and range estimation approach for FDA-MIMO radar based on decoupled atomic norm minimization (DANM). First, we convert the 2D-ANM based angle and range estimation into a decoupled semi-definite programming problem with two one-dimensional (1D) Toeplitz Hermitian matrices. Then we apply the matrix enhancement and matrix pencil method and construct a permutation matrix to obtain joint angle and range estimation and parameter pairing. Numerical results verify that the proposed approach overcomes the grid-mismatch effect of the sparse reconstruction-based OMP algorithm and outperforms the subspace-based MUSIC method.