Scholarlink

<p>Air quality is a decisive factor in endurance sports performance, but many existing training systems do not account for the fluctuating environmental conditions and unique sport-specific physiological demands and requirements. This work presents a real time, air quality aware framework that combines electrochemical air quality sensing with physiological sensing for adaptive training decisions. The architecture includes the following modules: AirQ-FeatNet is a deep neural network model for preprocessing and feature selection, DARC-Net is a deep learning model to perform the dynamic air-risk classification, and AErO-Net is designed to estimate the training load and accordingly suggest the best training load. It uses concentrations of pollutants (CO, NO₂, O₃, PM₂. ₅, PM₁₀) and temperature and humidity to evaluate their effects on athlete’s performance. The experimental results show superior performance with RMSE of 0.150, classification accuracy and prediction accuracy of 96% and 98%, respectively which further increases to 99.2% after applying the optimization. The overall system allows for safe, adaptive, personalized training modifications within a changing environment. The results demonstrate that environmentally informed physiological monitoring significantly improves safe training, decision accuracy, and performance guidance in real-life endurance sports.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Smart endurance training through electrochemical air quality sensing and adaptive data analytics

  • 期刊论文

  • 发表日期：

    2026年5月30日

查看全文
  • Chen Ju

摘要

Air quality is a decisive factor in endurance sports performance, but many existing training systems do not account for the fluctuating environmental conditions and unique sport-specific physiological demands and requirements. This work presents a real time, air quality aware framework that combines electrochemical air quality sensing with physiological sensing for adaptive training decisions. The architecture includes the following modules: AirQ-FeatNet is a deep neural network model for preprocessing and feature selection, DARC-Net is a deep learning model to perform the dynamic air-risk classification, and AErO-Net is designed to estimate the training load and accordingly suggest the best training load. It uses concentrations of pollutants (CO, NO₂, O₃, PM₂. ₅, PM₁₀) and temperature and humidity to evaluate their effects on athlete’s performance. The experimental results show superior performance with RMSE of 0.150, classification accuracy and prediction accuracy of 96% and 98%, respectively which further increases to 99.2% after applying the optimization. The overall system allows for safe, adaptive, personalized training modifications within a changing environment. The results demonstrate that environmentally informed physiological monitoring significantly improves safe training, decision accuracy, and performance guidance in real-life endurance sports.