3D multiple input single output near field automotive synthetic aperture radar

Abstract

Radar sensors play an important role in the technology field of autonomous driving. In contrast to optical cameras, they work reliable by night, mist, snow and fog. Recent radar sensors have problems in near-field scenarios. In this paper, we investigate how the technique of synthetic aperture radar (SAR) can be used to reconstruct the 3D near-field environment up to 30m with several physical radar sensors using the global backprojection algorithm. The main challenge applying SAR to an automotive 77 GHz radar sensor with 2 GHz bandwidth and a short sweep duration of 10 µs is to avoid azimuth aliasing, which means avoiding ghost targets. In the case of monostatic SAR, the sensors have to be placed 1mm next to each other, which is unrealistic in practice. By using multiple input single output (MISO) sensors arranged in a circular manner we increase this distance to 4 cm, which is now feasible in practice. This sensor arrangement enables applying SAR to near-field automotive applications.