Phase Error Estimation for Automotive SAR

Phase error estimation and correction plays an essential role in high quality synthetic aperture radar (SAR) image formation. Especially in automotive applications, because of the highly non-linear driving paths, it is required to alleviate uncompensated motion errors. In this work, we use the general version of phase gradient autofocus (PGA) which is compatible with time-domain image formation algorithms. The adapted method overcomes the typical problems of conventional approaches and shows remarkable robustness against a large range of simulated errors. Furthermore, the proposed approach is evaluated on real radar data acquired by mounting a 77-GHz radar system on a bumper of a car. It is demonstrated that the implemented algorithm removes phase errors and improves the quality of automotive SAR image formation.