High-Resolution Electromagnetic Vortex Imaging Based on Sparse Bayesian Learning

The vortex electromagnetic (EM) wave carrying orbital angular momentum (OAM) has been found great potential to improve radar imaging performance. However, to achieve high-resolution image, a large number of OAM modes should be applied in the existing EM vortex imaging methods, which seriously limits the improvement of resolution in practice. The orthogonality of OAM eigenmodes enables the reconstruction of targets with limited measurements. Thus, to improve the imaging resolution and reduce the number of measurements, the sparse Bayesian learning (SBL) is introduced to reconstruct the targets for EM vortex imaging. The sparse representation model for EM vortex imaging is first derived using the stepped frequency signal. Subsequently, based on the SBL framework, the enhanced SBL (ESBL) and variational sparse Bayesian inference (VSBI) are utilized to reconstruct the target. Simulation results show that the SBL-based reconstruction algorithms can achieve much higher resolution than conventional fast Fourier transform-based methods. The imaging results considering the influences of noise and radiation pattern optimization are presented and analyzed. Moreover, the off-grid imaging model is also derived and the target is reconstructed based on a modified VSBI algorithm. Finally, a practical application processing the real-world data from a proof-of-concept experiment in an anechoic chamber is provided.