Estimation of Lower-body Kinetics from Loading Profile and Kinematics Alone, Without Measured Ground Reaction Forces



Journal Title

Journal ISSN

Volume Title


48th International Conference on Environmental Systems


Biomechanical models of human motion can estimate kinetic outcomes, such as joint moments, joint forces, and muscle forces. Typically, one performs an inverse dynamics (ID) analysis to compute joint moments from the kinematics and measured external forces. Sometimes it is impractical to measure ground reaction forces and moments (GRF&M). We devised an empirical method for performing ID analysis of resistance exercises without measured GRF&M. The method solves the multibody dynamics equations of motion with four key assumptions about the GRF&M that reduce the number of unknowns. The assumptions are 1) negligible ground reaction moments, 2) fixed lateral/medial location of the center of pressure (COP), 3) equal fore/aft location of the COP between the feet, and 4) constant angle of the GRF vector relative to vertical in the frontal plane. We used evaluation trials from a prototype exercise countermeasure device to test this approach. Four participants performed squat and deadlift exercises at various loads. We compared results from traditional ID analysis to results without measured GRF&M using our method. We found that joint moment trajectories in the sagittal plane were similarly shaped between the two methods, and the amount of root mean squared error (RMSE), measured by difference in joint moment impulse, was typically under 15% (except for deadlift at the knee, <50%). Non-sagittal joint moment trajectories, which are much lower in overall magnitude, were not always similarly shaped between the two methods. Non-sagittal moments displayed much higher RMSE, with values ranging from 50-1000%. These findings were further supported by validation metrics (Sprague and Geers’ P and M metrics, Pearson’s r correlation coefficient). Based on these findings, we concluded that useful kinetic results are obtained from ID analysis of squat and deadlift exercises, even without measuring GRF&M, as long as the outcomes of interest lie in the sagittal plane.


William Thompson, NASA
R. Kenneth Huffman, Metecs, Inc.
Christopher Gallo, NASA
John Dewitt, KBRWyle
B.T. Humphreys, Zin Technologies
Aaron Godfrey, ZIN Technologies
David Frenkel, CACI NSS
Beth Lewandowski,NASA
ICES513: Computational Modeling for Human Health and Performance Analysis
The 48th International Conference on Environmental Systems was held in Albuquerque, New Mexico, USA on 08 July 2018 through 12 July 2018.


biomechanics, biomechanical modeling, inverse dynamics