This study evaluates the performance of a 32-marker motion capture (MoCap) system in estimating respiratory frequency (f) and tidal volume (V) during cycling exercise. Fourteen well-trained cyclists performed an incremental step test on a cycle ergometer, while simultaneously recording a raw flow signal with a reference metabolic cart (COSMED) and respiratory-induced torso movements with twelve optoelectronic cameras registering the position of 32 markers affixed to the torso. f and V were calculated from both systems on a breath-by-breath basis. The MoCap system showed a strong correlation with the COSMED system when measuring f and V (r2 = 0.99, r2 = 0.87, respectively) during exercise. For f, the mean absolute error (MAE) and mean absolute percentage error (MAPE) were 0.79 breaths/min and 2.1%, respectively. For ????, MoCap consistently underestimated values compared to COSMED, showing a bias (MOD ± LOA) of -0.11 ± 0.42 L and MAPE values of 8%. These findings highlight the system`s capabilities for real-time respiratory monitoring in athletic environments.
© Copyright 2025 Sensors. All rights reserved.