Non-contact Blood Pressure Estimation Method Based on Blood Pressure Category Classification

Area of Science:

• Biomedical Engineering
• Health Monitoring Technologies
• Signal Processing for Healthcare

Background:

• Non-contact blood pressure (BP) measurement is crucial for daily health monitoring.
• Doppler radar can detect pulse waves for BP estimation, but accuracy decreases with multiple subjects.
• Existing methods struggle with inter-subject variability in BP estimation models.

Purpose of the Study:

• To develop an accurate Doppler radar-based BP estimation method using BP category classification.
• To address the limitations of multi-subject BP modeling by incorporating classification.
• To enhance systolic blood pressure (SBP) estimation accuracy in a non-contact setting.

Main Methods:

• Pulse waves were extracted using Doppler radar.
• A k-Nearest Neighbor (kNN) algorithm classified pulse waves into 'Low BP', 'Normal BP', and 'High BP' categories.
• Subject-specific SBP estimation models were trained for each BP category.

Main Results:

• The proposed method demonstrated improved SBP estimation accuracy compared to models without BP category classification.
• BP category classification effectively mitigated accuracy degradation in multi-subject scenarios.
• The kNN-based classification successfully grouped pulse wave features relevant to BP levels.

Conclusions:

• Doppler radar combined with BP category classification offers a promising approach for accurate non-contact BP monitoring.
• This method enhances the reliability of BP estimation across diverse individuals.
• The study highlights the potential of categorized modeling for improving physiological parameter estimation.