Robust adaptive beamforming with interference-plus-noise covariance matrix reconstruction for FDA-MIMO radar

Abstract

Frequency-diverse array multiple-input-multiple-output (FDA-MIMO) antenna offers promising potential applications such as joint range-angle estimation, secure communication, and dual radar-communication systems. However, robust adaptive beamforming (RAB) for FDA-MIMO plays an important role, but it has not been well explored. In this paper, we identify that both steering vector and interference-plus-noise covariance (INC) matrix in FDA-MIMO antenna are time-variant, which may cause significant performance degradation. To address this issue for practical applications, we develop a RAB beamformer for FDA-MIMO by employing a two-dimensional decoupled atomic norm minimization (2D-DANM) approach for the INC matrix reconstruction. Unlike traditional methods that rely on multiple data snapshots, the proposed approach requires only a single snapshot, which can efficiently reconstruct the INC matrix to mitigate the time-variance. The steering vector is corrected through the reconstructed INC matrix by solving a quadratically constrained quadratic programming (QCQP) problem. The superiority is verified with simulation results, particularly in the term of output signal-to-interference-plus-noise ratio (SINR).