Abstract
Characterizing exercise-induced fatigue is challenging due to difficulties in monitoring biochemical and electrophysiological signals consistently. While wearable sensors able to monitor both signals are available, they struggle to maintain stable signal acquisition across exercise-rest scenarios because of significant sweat variation. Here we describe the development of a system, FatigueVisual, combining an adaptive sweat-regulating hybrid electronic textile (e-ASRHT) patch with artificial intelligence (AI) analytics to improve fatigue management. The key innovation lies in the precise regulation of the breakthrough pressure of micropores in fabric electrodes and sweat-excreting areas, allowing for accurate sweat management. This design enables efficient enrichment at an ultralow sweat rate (5.0×10-4 ml cm-2 min-1) while rapidly transporting excess sweat at a rate 2533 times the human physiological limit, ensuring stable chemical-electrophysiological sensing from exercise to rest. A cloud-connected AI model, trained on over 200,000 time-resolved observations collected from 10 participants, recognizes six fatigue-related states and provides personalized exercise management via a mobile app. Validation in three participants showed 87.3% agreement with clinical assessments. The system reduced exercise fatigue by 51.8% and accelerated recovery by 48 h compared with the control group. In practice, FatigueVisual may enable intelligent sports health management.