Breathing Rate Estimation Using Kalman Smoother With Electrocardiogram and Photoplethysmogram

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Signal Processing

Background:

• Direct breathing rate measurement requires uncomfortable sensors.
• Electrocardiogram (ECG) and photoplethysmogram (PPG) signals offer alternative, non-invasive sources for breathing rate (BR) estimation.
• Extracting accurate BR from ECG/PPG is challenging due to signal noise and individual patient variability.

Purpose of the Study:

• To develop accurate breathing rate (BR) estimation methods using only ECG or PPG signals.
• To introduce Respiratory Quality Indices (RQIs) for quantifying the quality of derived respiration waveforms.
• To present two novel methods for BR estimation: automatic signal selection and Kalman smoother-based signal fusion.

Main Methods:

• Derivation of multiple respiration waveforms from ECG and PPG signals based on amplitude, frequency, and baseline wander modulations.
• Quantification of modulation waveform quality using Respiratory Quality Indices (RQIs).
• Two fusion methods: automatic selection of the highest RQI signal and Kalman smoother-based fusion of high-RQI signals.

Main Results:

• The proposed methods were evaluated on two independent datasets: a benchmark database and recordings from patients during daily activities.
• Both proposed methods demonstrated superior performance compared to existing techniques in BR estimation.
• The combination of RQIs and a fusion algorithm significantly enhanced the accuracy of BR estimations.

Conclusions:

• Respiratory Quality Indices (RQIs) coupled with a fusion algorithm improve the accuracy of breathing rate estimation from derived modulation signals.
• A robust Kalman Smoother method, applicable in diverse clinical settings, enhances breathing rate estimation through data fusion.