This paper investigates whether millimeter-wave (mmWave) FMCW radar can monitor vital signs from the angular artery/vein region near the nose, a sensing location that is relevant to future smart-glasses integration. Using an AWR1843BOOST radar with DCA1000 raw-data capture, we collected left-side and right-side recordings over the near-nose vascular region together with background measurements. Using a Polar H10 chest strap, we collected ground truth heartrate labels for the dataset. Using an Apple Watch, we collected ground truth respiration rate labels for the dataset. We initially evaluated TX1, TX2, and TX3 and intentionally centered the final study on TX1 because it produced the clearest and most stable vital-sign signatures for this anatomical target. A TX1 dataset containing 82 samples was formed, and a no-leakage feature matrix with 168 engineered features was constructed from displacement, spectral, range-profile, and range-time-map representations. Separate regression models were trained for respiration rate and heart rate using repeated 5-fold cross-validation over 20 repetitions. The best respiration model, Extra Trees, achieved a mean absolute error (MAE) of 0.470 BPM, root mean squared error (RMSE) of 0.688 BPM, of 0.871, and Pearson correlation of 0.934. The best heart-rate model, Random Forest, achieved an MAE of 1.467 BPM, RMSE of 2.244 BPM, of 0.751, and Pearson correlation of 0.869. These results demonstrate the feasibility of near-nose angular artery/vein mmWave sensing and motivate future smart-glasses-mounted physiological monitoring systems.
Onta Shahriar (Wed,) studied this question.