Multiple range bin selection for radar-based monitoring yielded lower mean absolute errors and higher correlations with polysomnography reference for both breathing and heart rate in acute stroke patients, but provided lower temporal coverage compared to single range bin selection.
Multiple range bin selection produced lower mean absolute errors for breathing rate compared to single range bin selection in acute stroke patients.
Multiple range bin selection produced substantially lower mean absolute errors for heart rate compared to single range bin selection.
Multiple range bin selection showed higher correlation with polysomnography reference for both breathing and heart rate.
Single range bin selection produced radar-derived rates for a larger proportion of recording time than multiple range bin selection.
The study evaluated radar-based contactless sleep monitoring as a tool for monitoring acute ischemic stroke patients to reduce secondary stroke risk and prevent worse functional outcomes.
The authors conclude that multiple range bin selection provides more accurate estimates of breathing and heart rate but has lower temporal coverage compared to single range bin selection.
This research suggests that radar-based, contactless monitoring systems can accurately measure breathing rate and heart rate in stroke patients while they sleep in a hospital setting. The study tested two different methods for processing radar signals — one that analyzes a single region (range bin) of the radar data and one that analyzes multiple regions simultaneously — in 49 patients who had recently experienced an acute ischemic stroke. A gold-standard sleep monitoring test called polysomnography was used as a comparison to evaluate how accurate each radar method was. The study found a clear accuracy-versus-coverage trade-off between the two methods. The multiple range bin method was considerably more accurate: its average error for breathing rate was about 0.39 breaths per minute compared to 0.87 for the single bin method, and its average error for heart rate was 0.84 beats per minute compared to 3.99 for the single bin method. However, the single bin method provided measurements for a much larger portion of the recording time, particularly for heart rate (about 82% of the time versus only about 20% of the time for the multiple bin method). This means the more accurate method misses more data, while the less accurate method keeps more data. This research matters because sleep disturbances in stroke patients are associated with a higher risk of having another stroke and poorer recovery. A contactless, easy-to-use monitoring system could make it more practical to track patients' vital signs overnight in a clinical environment without requiring invasive sensors attached to the body. The findings suggest that further development is needed to improve the temporal coverage of the more accurate multiple bin method before it could be widely used in clinical care.
Szmola B, Hornig L, Vox J, Liman T, Radeloff A, Kollmeier B, et al.. (2026). Radar Multiple Bin Selection for Breathing and Heart Rate Monitoring in Acute Stroke Patients in a Clinical Setting.. Sensors (Basel, Switzerland). https://doi.org/10.3390/s26010251