Remote Asynchronous PhaseDetection of Thoracoabdominal Breathing: A Novel Human Backscattering Model With FMCW Radar

The detection of asynchronous thoracoabdominal movement is of great significance for the diagnosis of obstructive sleep apnea (OSA) and for determining the optimal time for weaning patients from mechanical ventilation in intensive care units. However, noncontact detection of this asynchronous phase requires radar systems with both high spatial resolution and high temporal fidelity to distinguish between the subtle, independent motions of the thorax and abdomen. This article proposes a novel detection approach based on frequency-modulated continuous wave (FMCW) radar utilizing a specific radar configuration. We first introduce an innovative human 4-D electromagnetic (EM) backscattering model that, grounded in physiological respiratory characteristics, accurately models the radar echoes generated by the distinct EM scattering from the thorax and abdomen. Building on this model, a robust FMCW baseband signal processing method is developed to enable the precise detection of the asynchronous phase of thoracoabdominal movements. Simulations and experimental results demonstrate that the proposed model aligns well with real-world measurements, achieving a baseband signal waveform correlation exceeding 0.7, thus providing a solid theoretical foundation for related radar detection applications. Furthermore, clinical sleep studies involving 10 participants have validated that our proposed FMCW detection method successfully achieves noncontact detection of asynchronous thoracoabdominal movement, with the correlation between the estimated thoracoabdominal phase difference and the phase difference measured by the polysomnography (PSG) achieved a mean correlation of r =0.82, showcasing its potential for medical application. © 2026 IEEE.