Parameterizing Running Kinematics

Studenteropgave: Speciale (inkl. HD afgangsprojekt)

  • Kristoffer Iversen
  • Brian Vendelbo Kloster
4. semester, Idrætsteknologi (cand.scient.techn), Kandidat (Kandidatuddannelse)
The aim of the present study was to develop a parametric running technique model with a minimum set of parameters. Full-body kinematic trajectories of 90 running trials was used to drive musculoskeletal models in the AnyBody Modeling System for simulation of a variety of running motions. A mixed terms Fourier series (FS) representation of the anatomical joint angles along with the translational and rotational components of the pelvis was performed. The coefficients of these FS served as the primary parameters of the running simulation model. All FS coefficients in company with the subject anthropometry were exposed to a Principal Components Analysis (PCA) which revealed that 90% of the variance could be explained by the first 12 principal components (PCs). Transforming only these 12 PCs back to its original domain allowed for recreation of the running trials. New running techniques were generated by multiplying the standard deviation of different PCs with different factors, adding them to the mean values of the 12 first principal components, and importing them back into AMS for a pseudo generated population of different running techniques.
Udgivelsesdato6 jun. 2017
Antal sider25
Ekstern samarbejdspartnerPeter Birk Andersen
Peter Birk Andersen
ID: 259245165