International Conference on Biomedicine and Bioinformatics Engineering · Apr 17, 2026

Research on PPG-Based Blood Pressure Detection Method

By Yan J, Zhang C, Gong N et al.

Cardiovascular diseases remain a leading global cause of morbidity and mortality, with blood pressure (BP) monitoring serving as a cornerstone for their early diagnosis and clinical intervention. Traditional cuff-based BP measurement methods, while clinically validated, are limited by poor user comfort, inability to support continuous monitoring, and high operational costs, making them unsuitable for long-term daily health management. To address these limitations, this study presents an intelligent non-invasive BP detection system based on photoplethysmographic (PPG) signals, combining machine learning with embedded technologies. First, 78 multi-domain features are extracted from raw PPG signals, and a 9-dimensional optimal feature subset is identified via feature selection. A BP prediction model is then constructed using Support Vector Regression (SVR). For embedded deployment, features are extracted on the STM32 microcontroller using Fast Fourier Transform (FFT) and Adaptive Multiscale Peak Detection (AMPD) algorithms, with the SVR decision function ported to realize on-board BP prediction. Experimental validation demonstrates that the embedded system achieves high consistency with MATLAB-based reference predictions. The proposed system enables offline, continuous BP monitoring, enriching the methodologies for non-invasive BP estimation and providing a feasible technical solution for the development of portable wearable monitoring devices for home healthcare.

doi.org/10.1145/3822183.3822219Open accessBiomarkers

AgingResearch.News