Models of Energetically Optimal Locomotion in Cursorial Mammals

Date
2020-02-14
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
It is widely held that cursorial mammals use gaits that minimize the energetic cost of locomotion, but it is hard to compare alternative strategies in real organisms. In this thesis, I show how simple models can inform and predict the gaits used by cursorial mammals from an energetics perspective. A simple work optimization model predicts how human subjects reduce their takeoff velocity and "bounciness" while running in reduced gravity. This work-based perspective is extended to quadrupeds using trajectory optimization, where it is shown that an additional term– the so-called force-rate penalty– is necessary to explain some features of canid locomotion. The shape of ground reaction forces, leg sequence at slow to moderate speeds, changes in duty factor at moderate to fast speeds, and the walk-trot transition are all predicted by this planar model. Next I use this model to show how changing pitch moment of inertia affects energy-optimal gait choice, matching gait preferences between dogs, horses, giraffes and elephants. Finally, I compare various modelling approaches used for four-legged mammals, and show how center-of-mass considerations alone do not explain the typical, four-beat walking gait used by most cursorial quadrupeds– despite the success of the center-of-mass approach in humans, as demonstrated in this very thesis. These results show that energetic optimization can be remarkably predictive of gait choice in mammalian cursors, even with a small number of modelling components. Where the predictions are deficient, they point to missing levels of complexity that could be added in future models.
Description
Keywords
optimization, comparative biomechanics, locomotion, models, energetics
Citation
Polet, D. T. (2020). Models of Energetically Optimal Locomotion in Cursorial Mammals (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.