A Smart Ring for Long-term Blood Pressure Monitoring

Traditional cuff-based blood pressure (BP) monitoring methods suffer from inherent limitations in user discomfort and intermittent measurement. In contrast, photoplethysmography (PPG)-based BP estimation provides a promising alternative. However, several critical challenges remain in PPG-based BP estimation. First, prior evidence indicates that finger-worn devices can yield signals with higher feasibility and quality than wrist-worn devices. Second, long-term variations in blood pressure have not been sufficiently investigated in this field. To address these challenges, this study develops a novel smart ring platform and collects samples to train and evaluate the proposed model. A large-scale dataset comprising 116,995 samples from 96 subjects is constructed to evaluate long-term performance. The trained model achieves mean absolute errors (MAE) of 11.20 mmHg for systolic blood pressure (SBP) and 7.51 mmHg for diastolic blood pressure (DBP). These results demonstrate that a smart ring provides a comfortable and practical solution for continuous blood pressure management in personalized healthcare.