Non-contact Beat-to-Beat Pulse Interval Estimation from Radial Artery Micro-Movements Using mmWave Radar

Abstract

This study presents a novel contactless method for beat-to-beat pulse interval estimation using 24 GHz millimeter-wave radar detection of radial artery surface motion. The proposed system employs Doppler effect principles to detect physiological micro-movements, incorporating comprehensive signal processing including band-pass filtering (0.7-3 Hz), baseline wander removal, and advanced nonlinear denoising algorithms. An adaptive peak-finding algorithm extracts inter-beat intervals (IBI) from processed radar signals. Validation was performed against simultaneous electrocardiogram (ECG) recordings in 15 healthy subjects during 60-second measurement sessions. Results demonstrated excellent agreement with reference ECG measurements, achieving a high Pearson correlation coefficient of r = 0.991, mean absolute error of 11.8 ms, and root mean square error of 16.93 ms. Bland-Altman analysis revealed 95% limits of agreement within ±33.15 ms. The mmWave radar approach provides a highly accurate, contactless alternative for cardiovascular monitoring, offering enhanced patient comfort and suitability for long-term health assessment applications while maintaining clinical accuracy comparable to traditional methods.