Long‐Term Used Continuous Blood Pressure Monitoring via Personalized Object Transfer Learning Architecture Enhanced Triboelectric Pulse Sensors

Abstract Cardiovascular diseases remain the leading cause of global mortality, underscoring the importance of continuous blood pressure (BP) monitoring for early diagnosis and management. However, due to individual diversity and temporal variability of patients physiological characteristics, traditional BP models suffer from inaccuracies over time, hampering the practical applications of continuous BP monitoring devices. Herein, a personalized object transfer learning architecture (POTLA) enhanced triboelectric pulse sensor is developed for accurate long‐term used continuous BP monitoring. Utilizing inversed molding‐replica technique, a nanopillar triboelectric pulse sensor (NTPS) is fabricated, which possesses desirable linear response, high sensitivity, and remarkable durability. The NTPS can detect the detailed features of pulse signals for BP derivation. Further, the POTLA is proposed to enhanced the BP models, which effectively addresses the individual diversity and temporal variability, thereby minimizing errors in BP monitoring for long‐term use of the sensors. A 6‐month follow‐up test involving five participants demonstrates that the POTLA significantly optimizes the BP derivation, with accuracy improvements of >27%, comparing to the universal and personalized BP models without the POTLA. This POTLA‐coupled sensor strategy holds promise for enhancing precision medicine and advancing healthcare applications.