AI-Enhanced Smart Textile Microwave Sensor for Real-time On-body Lactate Monitoring in Sports Applications

Abstract

The growing demand for noninvasive, real-time physiological monitoring in sports and healthcare has propelled the development of advanced wearable biosensors. This study presents an artificial intelligence (AI)-powered smart textile microwave sensor designed for continuous on-body lactate monitoring in sweat, addressing the limitations of conventional electrochemical and microwave-based systems, particularly their limited linear detection ranges. This limitation hinders the ability to accurately capture the full spectrum of typical lactate levels in sweat. The proposed sensor employs a flexible, textile-based substrate integrated with a dual-port microwave configuration, enabling simultaneous monitoring of multiple microwave parameters. Analyzing these parameters using AI models enhances accuracy and sensitivity, and enables broad-range lactate monitoring. Experimental validation was conducted using a broad lactate solution concentration range (0 to 125 mM) and on-body trials during exercise sessions. The AI-driven analysis significantly improved predictive performance, reducing the root mean square error (RMSE) by 63.34% and the mean absolute error (MAE) by 54.84% compared to traditional single-parameter approaches. Furthermore, real-time sweat lactate monitoring on a volunteer demonstrated a strong correlation with a commercially available blood lactate analyzer, confirming the sensor’s practical reliability. This innovative textile-based biosensor, combining microwave sensing with AI analytics, offers a robust, noninvasive solution for continuous lactate monitoring, highlighting its potential for applications in sports performance optimization, diagnostics, and personalized health monitoring.