The anaerobic threshold (LT) serves as a pivotal marker in cycling training but its regular monitoring is hindered by cost and invasiveness. This study explores a modelling approach for LT estimation using heart rate (HR) and power data collected from wearable technology. Twenty-four cyclists underwent incremental tests while wearing various commercial sensors. A discrete-time transfer function method was employed for modelling, with time-variant parameter (TVP) models showing promising accuracy (average error: 4%) in LT estimation. The adaptability of TVP models to capture HR dynamics contributed to their efficacy. This modelling technique offers a potential alternative for routine LT monitoring, leveraging widely used wearable sensors in cycling. Further validation and adaptation to field data are warranted.
© Copyright 2024 Journal of Science and Cycling. Cycling Research Center. All rights reserved.