Single-Beat Cuffless Blood Pressure Estimation Using Ear-PPG and ECG with a Lightweight Hybrid Learning Framework

Summary

Continuous cuffless blood pressure estimation is undermined in practice by motion artifacts, physiological variability, and the fragility of conventional pulse transit time models under dynamic conditions. Most prior approaches lean on multi-second windows to stabilize their estimates — an assumption routinely violated in real-world monitoring, where signal corruption is intermittent and unpredictable. This paper shows that discriminative BP-related information is in fact preserved at the single-beat level, and builds a lightweight multimodal wearable framework around that observation.