The Purpose of this study was to identify the potential of inertial sensor information to estimate whole-body mechanical power (WBP) in running. We recorded three-dimensional (3D) whole-body kinematic and kinetic data of eleven male subjects by means of optoelectronic motion capturing and an instrumented treadmill at speeds between 2.0 and 3.5 m/s. We simulated 3D acceleration and gyroscope signals for 15 segments of the whole body from marker trajectory data. We calculated one statistical model for each subject to estimate WBP from a set of 279 predictor variables derived from simulated sensor signals. Overall, WBP was estimated with root mean square errors between 4% and 20%. This highlights the potential of inertial sensor signals to estimate WBP. Nonetheless, in its current form, the method requires too many sensors for practical applications.
© Copyright 2019 ISBS Proceedings Archive (Michigan). Northern Michigan University. Published by International Society of Biomechanics in Sports. All rights reserved.