Multivariate Models Utilize Accelerometers to Estimate Peak Vertical Ground Reaction Force
Loading...
Date
2015-05
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
The Ohio State University
Abstract
Three-dimensional (3-D) motion capture is a valid method to quantify human motion, yet requires expensive equipment and a specialized laboratory not available in a clinical or athletics environment. Tri-axial accelerometers pose a portable, cost-effective solution to analyze human motion. Previous research indicates that accelerometers can predict peak vertical ground reaction forces (vGRF) during low-impact tasks, but these units have not been validated for dynamic, high-impact landings. The objective of this study was to develop a multivariate model that utilized anthropometric measures and peak accelerations to estimate peak vGRF during dynamic jumping tasks. Ten healthy subjects were recruited for the study. Activity monitors were secured bilaterally to the foot, medial tibial surface, lateral femoral epicondyle, and midpoint between the right and left anterior superior iliac spine. Subjects performed 10 drop vertical jump tasks and 10 bilateral single leg drop tasks off a 31cm tall box onto two floor embedded force plates. All tasks were performed during continuous collection of 3-D motion capture data and tri-axial accelerations. Peak vGRF was extracted from motion capture data for each DVJ and SLD trial. Peak acceleration was extracted from tri-axial acceleration data. Multivariate linear regression models that incorporated anthropometric data and peak acceleration magnitudes were separately developed to predict peak vGRF for DVJ and SLD trials. Height, weight, peak waist acceleration, and peak thigh acceleration were significant predictors of vGRF during a DVJ task. Peak waist, thigh, and shoe accelerations were significant predictors of vGRF during a SLD task. The correlations between recorded vGRF and predicted vGRF for both DVJ and SLD trials were significant (DVJ: R2 = 0.7451; SLD: R2 = 0.7266). Models that utilize anthropometric data and activity monitors to accurately predict vGRF provide a cost-effective method to collect human motion data. This study is the first to our knowledge to utilize multiple activity monitors to produce these results during dynamic, high-impact landing tasks. Future work will attempt to validate activity monitors to measure lower extremity kinetics and kinematics during multiple dynamic tasks.
Description
Keywords
biodynamics, sports medicine, motion capture, accelerometers