This paper presents a super-resolution framework for joint direction-of-arrival (DOA) and polarization estimation in polarization-sensitive sparse arrays, tailored to multi-frequency (MF) scenarios. To address the challenge of inconsistent delay scales across frequencies, a delay-domain alignment strategy is proposed to construct a unified virtual covariance matrix by aggregating frequency-specific spatial correlations onto a common grid. A decoupled atomic norm minimization (DANM) model is then formulated to reduce the computational burden of two-dimensional (2D) semidefinite programming (SDP) by exploiting the separability of the underlying 2D structure. The DOA parameters are recovered via Vandermonde decomposition, and a closed-form polarization recovery strategy is subsequently applied based on beamformed projections and eigenstructure analysis. The method constructs the estimated DOAs, the method constructs a polarization-dominant beamspace covariance matrix through spatial conditioning via spatial projection, followed by vector fitting and regularized inversion to robustly estimate the polarization orientation and phase. Numerical simulations validate the effectiveness of the proposed method under various signal-to-noise ratio (SNR) and snapshot conditions, demonstrating improved accuracy and robustness with reduced computational complexity.
Hu et al. (Tue,) studied this question.