AI-enhanced superior bionic skin with dual-layer nanofiber membranes for sweat-based multiple biomarker detection

Wearable smart sweat sensors hold promise for non-invasive health monitoring, but existing devices often face issues with user discomfort and inefficient sweat collection. Herein, we present a novel sandwich-structured bionic skin with dual-layer nanofibers integrating multifunctional detection of glucose, lactate, uric acid, Na + , K + , NH₄ + , skin impedance, and temperature. Fabricated via thermal transfer printing and zone-specific localized compression, the device reduces production time and cost. The inner polyacrylonitrile (PAN/F127) layer enhances sweat absorption, achieving over 8× faster absorption than microchannels, while the outer polydimethylsiloxane (PVB/PDMS) layer provides waterproofing, radiative cooling, and antifouling, ensuring comfort and enzymatic stability. The sensor exhibits high sensitivity, selectivity, and retains over 84% performance after nine days. Integration with a redesigned Deep Learning Continuous Analyte Monitoring with Real-time Engagement (D-CARE) system, featuring two flexible circuits and a deep learning attenuation correction algorithm, enables precise real-time monitoring and health anomaly alerting. In situ exercise tests confirm continuous, accurate physiological monitoring, advancing personalized healthcare with a reliable, user-friendly wearable platform.