Heart Rate Estimation using PPG signal during Treadmill Exercise.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Wearable Technology

Background:

• Wearable photoplethysmographic (PPG) sensors are widely used for heart rate monitoring.
• Estimating instantaneous heart rate from PPG signals is challenging due to non-stationary motion artifacts, especially during exercise.
• Current methods often rely on averaged heart rates, missing crucial beat-to-beat fluctuations during physical activity.

Purpose of the Study:

• To develop and validate a novel technique for accurate instantaneous heart rate estimation from PPG signals.
• To address the challenge of motion artifacts in wearable PPG sensors during dynamic activities like running.
• To improve upon existing state-of-the-art methods for real-time heart rate tracking.

Main Methods:

• Proposed a three-part method involving time-frequency spectrum estimation of PPG and accelerometer signals.
• Implemented motion artifact removal by subtracting accelerometer time-frequency spectra from PPG signals.
• Utilized postprocessing with cubic spline interpolation to refine heart rate estimates and reject artifacts.

Main Results:

• Achieved an average absolute error of 2.93 beats per minute and an average relative error of 2.31 beats per minute.
• Demonstrated significant improvements of 121% and 119% in average absolute and relative errors, respectively, compared to a state-of-the-art algorithm.
• Validated the method on datasets from the IEEE Signal Processing Cup Challenge and a custom dataset with treadmill running.

Conclusions:

• The novel technique provides accurate instantaneous heart rate estimation from wearable PPG sensors, even with significant motion artifacts.
• This method offers a substantial improvement over existing algorithms for real-time, beat-to-beat heart rate monitoring during exercise.
• The findings have implications for enhanced physiological monitoring in wearable health and fitness devices.