Cuff-less Continuous Blood Pressure Estimation via Multimodal Integration of Millimeter-Wave Radar and PPG Signals

Area of Science:

• Biomedical Engineering
• Cardiovascular Monitoring
• Signal Processing

Background:

• Blood pressure is a critical indicator of cardiovascular health.
• Traditional blood pressure measurement methods are often uncomfortable and not suitable for continuous monitoring.
• There is a need for accurate, non-invasive, and continuous blood pressure monitoring solutions.

Purpose of the Study:

• To develop and validate a cuffless blood pressure measurement technique using millimeter-wave radar.
• To assess the accuracy of a hybrid machine learning model for blood pressure estimation.
• To provide a comfortable and continuous blood pressure monitoring alternative.

Main Methods:

• Utilized a 24 GHz millimeter-wave radar system to acquire photoplethysmographic (PPG) signals from a finger clip.
• Employed variational mode decomposition (VMD) and finite impulse response (FIR) filters to reconstruct heartbeat waveforms from radar echoes.
• Developed a hybrid machine learning model combining Long Short-Term Memory (LSTM) and Support Vector Regression (SVR) for blood pressure estimation, synchronized with PPG signals.

Main Results:

• The millimeter-wave radar and finger clip combination demonstrated precise blood pressure measurements.
• Reconstructed heartbeat waveforms accurately reflected physiological changes.
• The hybrid LSTM-SVR model showed acceptable accuracy in estimating blood pressure compared to standard methods.

Conclusions:

• The developed millimeter-wave radar-based system offers a promising approach for cuffless and continuous blood pressure monitoring.
• This hybrid method provides an accurate and comfortable alternative to traditional blood pressure measurement devices.
• The findings highlight the clinical relevance of radar technology in non-invasive cardiovascular monitoring.