Extraction of Signal in Noncontact Human Vital Signs Signal Using Machine Learning Approaches

Assessing the static condition of the human organic entity relies on a comprehensive understanding of crucial parameters, namely heart and respiratory rates (HRs and RRs). These fundamental vital signs find application across various domains, including survivor localization in disaster-stricken environments, deception detection, and emotion recognition. The burgeoning popularity of individual pmhysiological signal monitoring devices can be attributed to advances in embedded hardware and the miniaturization of sensors dedicated to tracking vital signs. However, the integration of wearable physiological monitoring technologies often disrupts individuals’ daily routines. To mitigate these disruptions, this paper explores the potential of noncontact measurement tools, specifically the Doppler radar system, as a pragmatic and discreet alternative to conventional methods like video monitoring and infrared sensing. Beyond enhancing privacy, the Doppler radar system demonstrates the capability to discern signal variations induced by bodily movements during continuous physiological monitoring. This research leverages machine learning approaches to extract meaningful insights from noncontact human vital signs signals, encompassing HRs and RRs, while also capturing data on body movements. The amalgamation of machine learning and noncontact signal extraction promises to advance our comprehension of human physiological states, offering a promising avenue for improving health assessments and monitoring in various applications.