Novel Continuous Respiratory Rate Monitoring Using an Armband Wearable Sensor

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Wearable Technology

Background:

• Wearable devices commonly use photoplethysmography (PPG) and accelerometer (ACC) for pulse and activity tracking.
• Accurate, continuous respiration rate (RR) monitoring via wearables remains a significant challenge.
• Existing methods struggle with accuracy and are often affected by factors like skin tone.

Purpose of the Study:

• To develop and validate a novel algorithm for estimating respiration rate (RR) from upper-arm wearable sensors.
• To improve the accuracy and robustness of continuous RR monitoring using PPG and ACC data.
• To assess the algorithm's performance across different skin tones and compare it with existing methods.

Main Methods:

• A new algorithm was developed to derive respiratory surrogate signals from PPG and ACC data.
• The algorithm was evaluated retrospectively on a clinical dataset of 38 subjects.
• Data included simultaneous recordings from an upper-arm wearable device and a capnography monitor (RR reference).

Main Results:

• The proposed RR algorithm demonstrated high accuracy, with a bias of -1.3 brpm and a mean absolute error (MAE) of 2.7±1.6 brpm.
• The algorithm showed robustness across a wide RR range (4-59 brpm) with minimal outage (0.3±1.2%).
• Unlike the standard PPG Smart Fusion method, the proposed algorithm showed no significant performance degradation in subjects with darker skin tones (p=0.63).

Conclusions:

• A highly accurate and robust algorithm for unobtrusive, continuous respiration rate monitoring using an armband wearable device has been demonstrated.
• The novel method overcomes limitations of existing techniques, particularly concerning variations in skin pigmentation.
• This advancement holds significant potential for remote and daily life physiological monitoring.