Author(s): CHEN, C., MIAU, J.J., CHEN, Y.R., TSENG, Y.H., TASI, W.C. , Institution: NATIONAL TAIPEI UNIVERSITY OF TECHNOLOGY, Country: TAIWAN, Abstract-ID: 1954

INTRODUCTION:
Accurate cycling power estimation is essential for performance monitoring, yet existing models often lack validation across varying speeds, gradients, and aerodynamic conditions. This study evaluated the accuracy of a cyclist-specific power estimation model against power meter measurements under different speeds and gradients.
METHODS:
The model incorporated gravity, rolling resistance, acceleration, and aerodynamic drag, with Cda quantified from cyclist trunk dimensions and wind speed measured by an anemometer. Field tests were conducted at 15–40 km/h and 0–15° gradients. Agreement was assessed using Bland–Altman analysis and intraclass correlation coefficients (ICC).
RESULTS:
Mean differences between estimated and measured power were minimal, and Bland–Altman plots showed no systematic bias. ICC was 0.946 for speed conditions and 0.996 for gradient conditions, indicating high agreement.
CONCLUSION:
The model accurately predicts power across varying speeds, gradients, and wind conditions, demonstrating strong agreement with measured power and potential for application in outdoor cycling performance monitoring and training decisions.