An Improve mmW Automotive Radar DOA Estimation Approach: Joint Matrix Completion and Spectrum Extrapolation

Millimeter wave (mmW) automotive radar with limited array aperture resulting in low resolution of direction of arrival (DOA) estimation. The sparse linear array (SLA) can effectively improve the resolution of DOA estimation, which can also reduce automotive radar system complexity than uniform linear array (ULA). However, compared with ULA, there are data missing in SLA causing the high sidelobe or grid-lobe interference in DOA estimation. Therefore, we joint matrix completion and spectrum extrapolation to improve DOA estimation in automotive radar. Firstly, we construct an incomplete low-rank Hankel matrix by single-snapshot radar echo data in SLA. Then, to reduce high sidelobe or grid-lobe problem, we use the low rank prior to recover the missing data to reconstruct a complete Hankel matrix. Finally, to further improve the resolution, we use the completed Hankel matrix to expand into a larger virtual array by spectrum extrapolation. The simulation results show that the proposed method not only can effectively suppress sidelobe or grid-lobe of SLA, but also can further improve the resolution of DOA estimation.