Robust Heart Rate Detection via Multi-Site Photoplethysmography.

Area of Science:

• Biomedical Engineering
• Wearable Technology
• Physiological Monitoring

Background:

• Smartwatches are widely used for noninvasive physiological monitoring, particularly heart rate (HR) estimation via photoplethysmography (PPG).
• Current PPG-based HR monitoring faces accuracy challenges due to motion artifacts and inconsistent sensor placement during daily activities.
• Robust and accurate HR estimation outside clinical settings remains a significant unmet need.

Purpose of the Study:

• To develop and evaluate a novel sensor fusion method for enhanced HR estimation using multiple PPG sensors.
• To improve the robustness and accuracy of wearable-based HR monitoring in real-world, dynamic environments.
• To address the limitations of single-device PPG accuracy during physical activity and varied placement.

Main Methods:

• A novel sensor fusion algorithm was developed to integrate data from multiple PPG sensors (wrist, ankle, head, sternum).
• Signal quality assessment was performed on each sensor input to dynamically weight their contribution.
• A joint, robust PPG signal was created for HR estimation, validated against ECG.
• The method was tested on a new dataset from 14 participants during a full day of real-world activities.

Main Results:

• The proposed multi-site sensor fusion method achieved a mean HR error of 2.4 bpm.
• This represents a 46% improvement in accuracy compared to the best single-device performance (4.4 bpm).
• The method demonstrated robust HR estimation across diverse activities and participant placements.

Conclusions:

• Multi-site PPG sensor fusion offers a significant advancement in accurate, real-world HR monitoring.
• This approach overcomes key limitations of single-sensor PPG devices, enhancing reliability.
• The developed method holds promise for more dependable physiological tracking in consumer wearables.