MobiVital: A Dataset for Contactless Respiration Monitoring Using UWB Radar

Respiration is an important vital sign that reflects essential physiological functions of human bodies. Continuous respiration monitoring can assist in diagnosing health issues and enhancing rehabilitation training. However, while numerous learning-based methods have been proposed to extract respiration, public datasets for wireless respiration sensing are scarcely available. To support fast prototyping and encourage reproducible research for wireless vital signal monitoring systems, we introduce MobiVital, a 12-person, 24-hour dataset collected using an ultra-wideband (UWB) radar to monitor the subjects' vital signals, with time-synchronized ground truth obtained from wearable sensors.