Robust Beat-to-Beat Interval from Wearable PPG using RLS and SSA

Area of Science:

• Biomedical Engineering
• Signal Processing
• Wearable Technology

Background:

• Ambulatory Photoplethysmogram (PPG) offers a user-friendly alternative to Electrocardiogram (ECG) for continuous cardiac monitoring.
• Wearable PPG devices are susceptible to motion artifacts, compromising data accuracy.
• Accurate beat-to-beat interval extraction is crucial for reliable cardiac assessment.

Purpose of the Study:

• To develop a novel pipeline for motion-resistant beat-to-beat interval extraction from noisy PPG signals.
• To enhance the accuracy and reliability of cardiac monitoring using wearable PPG devices.
• To address the challenge of motion artifacts in PPG signal analysis.

Main Methods:

• Adaptive Recursive-Least-Square (RLS) Filtering and Singular Spectrum Analysis (SSA) were employed to minimize motion artifact effects.
• A peak identification and location correction method using weighted local interpolation was implemented.
• Outlier peak-to-peak intervals were identified and marked as incorrigible.

Main Results:

• The proposed method achieved a 1.68% mean peak detection error rate.
• The mean absolute error for detected beat-to-beat intervals was 11.32 milliseconds.
• Performance metrics significantly outperformed state-of-the-art techniques, by at least 12.58 and 5.74 times for different metrics.

Conclusions:

• The developed pipeline effectively extracts reliable beat-to-beat intervals from motion-corrupted PPG signals.
• This method offers a robust solution for continuous cardiac monitoring with wearable PPG devices.
• The findings demonstrate a substantial improvement over existing techniques for PPG signal analysis.