PPG-Based Electronic Sensing System for Cuff-Less Calibration-Free Blood Pressure Estimation Using Machine Learning Algorithms

Continuous blood pressure monitoring is an important aid for preventing, diagnosing, and timely treating different pathologies, especially in the field of cardiovascular diseases. Photoplethysmography (PPG) is an advanced and non-invasive technology that is increasingly used in vital signs monitoring. One of the main advantages of photoplethysmography is its ability to continuously and accurately measure a biological signal, relying on the sensitivity to blood volume changes in blood vessels, which allows the real-time recording of each heartbeat and the assessment of the presence of any abnormalities. The aim of this paper is to present an innovative electronic system for cuff-less blood pressure prediction, relying on the only photoplethysmo-graphic signal. After accurate signal pre-processing which aims to remove motion artifacts and slow fluctuations, this system uses an algorithm based on Feed-Forward Artificial Neural Network (ANN) to estimate both systolic and diastolic blood pressure, using temporal and spectral features. This algorithm has been tested using a brand-new dataset, which includes multi-wavelength photopletysmographic (PPG) signals and blood pressure values derived from a reference device. The results are promising, with a Mean Absolute Error on the Test Set of 5.08 ± 8.83 mmHg and 4.37 ± 7.08 mmHg for systolic and diastolic blood pressure respectively.