Abstract
Deep learning offers interesting tools for radar data processing. This paper presents a signal processing method using a neural network to increase the range resolution of a Frequency Modulated Continuous Wave (FMCW) radar. The duration of the beat signal can be increased in the time domain for a better frequency resolution during conventional spectral estimation. This technique allows the radar bandwidth limitation to be exceeded and improves the sensing performance. A methodology is proposed as well as experimental results on real radar signal.