Motion-Robust Contactless Heartbeat Sensing Using 4-D Imaging Radar

In ambient-assisted living, in-cabin sensing, search and rescue, and public security, radar-based wireless technology has been proven effective in sensing respiration and heartbeat signals. However, existing research mainly focuses on stationary human subjects while body movements are inevitable in real applications. To robustly sense the heartbeat signal during body movement, we propose a motion-robust heartbeat sensing method based on joint blind source separation (JBSS) exploiting a 4-D imaging radar. This method combines physical model and data knowledge by exploring multiview measurements of asymmetric movements in the 4-D radar image sequence. Various kinds of body movements including swaying and random body movements (RBMs) are verified in real-scene experiments. The experimental evaluation demonstrated that the heartbeat signal can be extracted with a heart rate (HR) accuracy of 92.2%.