FMCW radar based high-precision gridless non-contact vital signs monitoring for Internet of Medical Things.

Background: Vital signs monitoring is of paramount importance in healthcare, serving as a crucial component in disease prevention, diagnosis, and management. Traditional contact-based devices, including electrocardiographs and pulse oximeters, while providing vital data, face limitations in long-term use owing to patient discomfort.

Objective: This study aims to propose a non-contact monitoring system utilizing Frequency Modulated Continuous Wave (FMCW) radar for continuous, non-invasive health monitoring. The objective is to overcome the constraints of traditional methods and enhance the feasibility of remote chronic disease management.

Methods: The proposed system employs the Multiple Signal Classification (MUSIC) algorithm to estimate respiration and heart rates. To tackle challenges such as noise interference and signal overlap, an enhanced root-MUSIC algorithm is introduced. This algorithm transforms the single-channel model into a multi-channel one and optimizes signal estimation through semi-definite programming (SDP) and the Alternating Direction Method of Multipliers (ADMM). Simulations and real-world experiments were conducted to validate the system's effectiveness.

Results: The validation process demonstrated the system's efficacy, revealing that the multi-channel model significantly reduces theoretical error bounds. In experimental trials, the method achieved a respiration rate Root Mean Squared Error (RMSE) of 0.0131 Hz and a heart rate RMSE of 0.0394 Hz, with corresponding accuracies of 96.05% and 90%. Bland-Altman analysis further corroborated strong concordance with contact-based devices.