Fast and Accurate Range-Doppler Estimation in Multi-Target Wideband Automotive FMCW Radar

This paper studies the application of wideband frequency-modulated continuous-wave (FMCW) radar where the received signals are sampled first before processing. It is shown that sampling at a rate of the same order as the transmitted signal bandwidth and reducing the processed fast-time interval, helps resolve the effect of some unwanted artifacts accompanied with increasing the bandwidth and the target's radial velocity. A signal model is first developed to capture range and velocity parameters, and upper bounds are then defined on the bandwidth that separate the wideband scenario from the narrowband one; finally a novel two-stage multiple signal classification (MUSIC) based algorithm is proposed and simulation results are provided to demonstrate its performance.