A Robust Dynamic Heart-Rate Detection Algorithm Framework During Intense Physical Activities Using...

Area of Science:

• Biomedical Engineering
• Signal Processing
• Wearable Technology

Background:

• Accurate heart-rate (HR) estimation from photoplethysmogram (PPG) signals during physical activity is challenging due to motion artifacts (MAs).
• Reconstructing clean PPG signals and extracting reliable HR from contaminated data requires robust algorithms.

Purpose of the Study:

• To propose a robust algorithm framework for accurate HR estimation using single-channel PPG and tri-axis acceleration data.
• To reconstruct corrupted PPG signals and calculate HR by analyzing PPG features and spectral characteristics.

Main Methods:

• A novel algorithm framework combining PPG and acceleration data for MA detection and removal.
• Utilizing spectral peak filtering, a novel NFEEMD (repeated single-notch filter and ensemble empirical mode decomposition) algorithm, and HR calibration.
• Signal processing methods adapted based on the presence and severity of MAs.

Main Results:

• The NFEEMD algorithm achieved low average estimation errors: 1.12 BPM (training), 2.63 BPM (testing), and 1.87 BPM (all datasets).
• A high Pearson correlation of 0.992 was observed, indicating strong agreement.
• The algorithm demonstrated suitability for both continuous and intense physical activities, including running and arm exercises.

Conclusions:

• The proposed algorithm framework provides a robust solution for accurate HR estimation from PPG signals contaminated by MAs.
• The method is effective across a range of physical activities, offering potential for improved wearable health monitoring.
• The integration of acceleration data significantly enhances the reliability of HR tracking during dynamic movements.