Abstract
Radar is of vital importance in many fields, such as autonomous driving,safety and surveillance applications. However, it suffers from stringentconstraints on its design parametrization leading to multiple trade-offs. Forexample, the bandwidth in FMCW radars is inversely proportional with both themaximum unambiguous range and range resolution. In this work, we introduce anew method for circumventing radar design trade-offs. We propose the use ofrecent advances in computer vision, more specifically generative adversarialnetworks (GANs), to enhance low-resolution radar acquisitions into higherresolution counterparts while maintaining the advantages of the low-resolutionparametrization. The capability of the proposed method was evaluated on thevelocity resolution and range-azimuth trade-offs in micro-Doppler signaturesand FMCW uniform linear array (ULA) radars, respectively.