Human stepping response to perturbations during quiet standing:experiments and predictions from metabolic energy optimization

Autor: Lehtinen, Kevin M.
Jazyk: angličtina
Rok vydání: 2016
Předmět:
Druh dokumentu: Text
Popis: The mechanical complexity of humans allows for many possible ways of stabilizing quiet standing. Here, we test the hypothesis that humans, when pushed or pulled during quiet standing, recover back to quiet standing using an energy optimal trajectory. We consider a simple biped model with point-mass body and legs capable of length change and ankle torques. For this model, for different forward pull perturbations from quiet standing, we computed the energy optimal trajectories back to quiet standing. For small perturbations, the optimal strategy involved no stepping and for large perturbations, the optimal strategy involved stepping. To test these predictions quantitatively, we collected experimental data on humans, subjected to forward and backwards perturbations. As predicted and as seen in prior studies, the subject did not use stepping below a perturbation threshold and used entirely stepping above another independent threshold, while in between these bounds, a mixture of the two strategies were utilized. Our simulations also discovered a range in which recovering without the need of a step is possible, but where taking a step is the more energetically optimal option. To study this effect in experimentation, additional testing was conducted in which subjects were instructed to avoid stepping, known here as the “constrained” testing. As mentioned, the “unconstrained” testing revealed a range where mixed stepping and non-stepping recoveries were utilized with equal frequency (50% stepping, 50% non-stepping). The constrained results in this same range showed a significant decrease in the frequency of stepping recoveries (20% stepping, 80% non-stepping). This decrease suggests that subjects also naturally choose stepping recoveries when non-stepping alternatives are available. However, quantitatively, the stepping versus non-stepping ranges identified in experiments and simulations were not identical. In addition, the large variation in experimentation suggests other variables exist which determine the stepping response that were not considered in this test. Other potential governing principles include minimization of forces, maximization of stability, and fear of injury (which may result in much more complex behavior). Testing of these hypotheses could help better characterize human stepping during balance. In addition, improvement of the model to include finite rotational inertia at the torso and more accurate foot geometry may improve quantitative results in simulation. However, despite the potential for improvement and further work, the qualitative consistency between simulation and experimentation suggest that metabolic optimality is a reasonable predictor of the human stepping response.
Databáze: Networked Digital Library of Theses & Dissertations