Contactless acquisition and decoupling of respiratory and cardiac signals using a 122 GHz FMCW radar

Abstract

This paper presents a contactless method for the simultaneous monitoring of heart rate (HR) and respiratory rate (RR) using a Frequency Modulated Continuous Wave (FMCW) radar operating at 122.5 GHz. To address the challenges posed by spectral overlap and interference between cardiac and respiratory components, we introduce a signal separation and estimation framework based on Maximum Likelihood Estimation (MLE) combined with an orthogonal projection technique. The approach models the thoracic displacement captured by the radar as the superposition of periodic sources and applies spectral projection to isolate the fundamental frequencies and their harmonics. Theoretical performance is assessed through the derivation of the Cramér-Rao Lower Bound (CRLB) and extensive Monte Carlo simulations under varying SNR and sample size conditions. Experimental validation was conducted on 23 healthy volunteers, each providing two 20 second recordings, using a clinical-grade reference monitor. The method achieves high estimation accuracy, with a mean absolute error of 0.69 bpm for HR and 0.33 bpm for RR. The algorithm is capable of reconstructing the cardiac pulse waveform while preserving key morphological features such as the systolic peak and dicrotic notch. Furthermore, the use of the projection technique further enhanced estimation accuracy under nonphysiological conditions, including both elevated and reduced rates. These findings demonstrate the potential of the proposed method as a reliable tool for contactless monitoring of vital signs and support its future integration into biomedical applications.

This work was supported by the Industrial Doctorates Plan of the Department of Research and Universities of the Generalitat de Catalunya, and is part of the I+D+i project 6-SENSES (PID2022-138648OB-I00) funded by MICIU/AEI/10.13039/501100011033 and ERDF/EU. It is also supported by the Spanish MCIN project PID2020-117303GB-C21, funded by AEI/10.13039/501100011033, and by the China Scholarship Council (CSC) under Grant 202208390068.

Peer Reviewed

Postprint (published version)