Non-invasive estimation of ground and joint kinetics through deep learning

Bill Johnson

Research output: ThesisDoctoral Thesis

70 Downloads (Pure)

Abstract

Captive instrumentation binds biomechanical analysis to the laboratory. This research aimed to break this hold by training deep learning models in the relationship between motion capture inputs and kinetic outputs (ground reaction forces/moments, and knee joint moments). Creating models driven by optical marker-based systems was the first step towards the goal of using non- invasive inertial measurement units (or 'wearable sensors'). This work is a precedent for large-scale treatment and modeling of spatio-temporal biomechanics data. The deployment of deep learning models in lieu of laboratory-based instrumentation is a potential disruptor for the discipline of biomechanics.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • The University of Western Australia
Thesis sponsors
Award date14 Oct 2019
DOIs
Publication statusUnpublished - 2019

Fingerprint Dive into the research topics of 'Non-invasive estimation of ground and joint kinetics through deep learning'. Together they form a unique fingerprint.

  • Cite this