Cuffless blood pressure estimation by error-correcting output coding method based on an aggregation of AdaBoost with...

Area of Science:

• Biomedical Engineering
• Medical Devices
• Signal Processing

Background:

• Traditional blood pressure (BP) monitoring relies on cuff-based devices, which can be cumbersome and invasive.
• There is a growing need for convenient, non-invasive methods for continuous BP estimation.
• Photoplethysmography (PPG) signals contain rich information about cardiovascular dynamics.

Purpose of the Study:

• To develop and validate a novel cuffless and non-invasive technique for BP estimation.
• To utilize pattern recognition with PPG signals for accurate Systolic Blood Pressure (SBP) prediction.
• To assess the feasibility of a smaller, more portable BP monitoring device.

Main Methods:

• A pattern recognition approach was employed using data from a Photoplethysmograph (PPG) sensor.
• Error-Correcting Output Coding (ECOC) was used as a multi-classifier machine, with AdaBoost as the binary classifier.
• Data from 368 volunteers were analyzed to estimate Systolic Blood Pressure (SBP) from Pulse Wave (PW) features.

Main Results:

• The developed technique achieved an Mean Difference (MD) of -1.2 mmHg and a Standard Deviation (SD) of 11.7 mmHg when comparing estimated SBP to measured SBP.
• The method demonstrated the potential for accurate BP estimation without traditional cuffs.
• The system utilizes individual information and Pulse Wave (PW) features for SBP calculation.

Conclusions:

• The proposed cuffless, non-invasive BP estimation technique shows promise for advancing continuous monitoring systems.
• The use of PPG sensors and machine learning (ECOC, AdaBoost) offers a viable alternative to conventional BP measurement devices.
• The developed method results in a smaller, more convenient device for BP monitoring.