Doppler Resolution Enhancement Algorithm Based on Extrapolation for FMCW Radar

Abstract

This paper proposes a Doppler resolution enhancement algorithm based on signal extrapolation for frequency-modulated continuous wave (FMCW) radar systems. In conventional FMCW radar, Doppler estimation is performed by repeatedly transmitting multiple chirp signals and analyzing the phase variations of the received signals via the fast Fourier transform (FFT). However, when multiple targets have similar velocities, their Doppler shifts become indistinguishable, often resulting in the detection of multiple targets as a single peak in the Doppler spectrum. A common approach to resolving this issue is to increase the observation time, which improves resolution but at the cost of additional resources. To overcome this limitation, the proposed method enhances Doppler resolution without extending the observation window by extrapolating the received signal through simple signal manipulations. Simulation results verify that the proposed algorithm achieves improved Doppler resolution without incurring additional resource overhead. Compared to the conventional FFT approach, it achieves up to 33% reduction in RMSE and up to 68% reduction in missing rate under moderate SNR and closely spaced targets, while maintaining a similarly low computational cost.