부산대학교 도서관

This paper considers the problem of human vital sign detection in a non-line-of-sight scenario by exploiting a millimeter-wave radar system. Specifically, first, the multipaths induced by wall reflections and corner diffraction are separated in the range-angle domain. Then, a path association method is proposed to match different multipaths with different targets to obtain the locations of the targets. After estimating the targets’ range cells, the time series signs with respect to multipath are formed and used to estimate the respiration and heartbeat information of the human targets by exploiting the cross-power spectrum. Finally, the proposed method is verified to get high-precision location information and vital sign values by both simulations and experiments.