Human Machine Interfacing with A Variable Speed Treadmill During Sensory Perturbation

Date of Award


Degree Name

M.S. in Mechanical and Aerospace Engineering


Department of Mechanical and Aerospace Engineering


Megan Reissman


Treadmills are preferred tools in both gait analysis and gait retraining as they can allow for long term locomotion in a small area. Fixed speed treadmills have been proven to decrease joint angles and decrease stride length of the user when compared to overground walking. They also restrict subject walking to a specific speed, when overground walking speed naturally fluctuates. Variable speed treadmills (VST) are designed to mitigate these artifacts and provide a more natural walking experience by adjusting to the users walking in real time. This study implements a VST developed with collaborators at The Ohio State University that utilizes the sacrum position of the user to adjust the treadmill speed. It should be mentioned that presently little is known about the human interfacing for this specific VST or how sensory perturbations will affect walking consistency in users. This study seeks to investigate the effect VST control gain (���� = 0.40 and ���� = 0.58) has on walking kinematic and spatiotemporal metrics, as well as the effect sensory perturbations (visual, vestibular, and auditory) have on human-machine interfacing. Spatiotemporal and kinematic metrics were connected through motion capture data, sacrum position and gait speed were collected through controller data collected during iv each trial. A total of 14 participants, aged 24±7.56 years old, performed three baselines, (fixed speed, ���� = 0.40, and ���� = 0.58), and five walking trials (Visual, Dark, Dark Auditory, Auditory, and Vestibular) for each VST gain. Each trial was two and a half minutes in length. Spatiotemporal metrics were calculated from a segment-based motion capture analysis in V3D. A three-factor repeated measure ANOVA of stride length and stride time was conducted across all gains and participants. A Tukey-Kramer multiple comparison statistical analysis was applied to sacrum position data normalized about the center of the dead band and walking speed normalized to the participants self-selected walking speed. Response to sensory perturbations can be characterized as significant increases in positional variability, normalized speed, and delta speed, but no changes in speed variability. Participants tended to speed up at the start of the perturbation but then maintain the faster walking speed for the rest of the trial. This was counter to the hypothesis that sensory perturbations would lead to a decrease in walking speed or at least increased speed variability. Vestibular and multi-sensory perturbations generated the largest deviations and should be investigated further. Results demonstrated that the gain setting of a variable speed treadmill will have a significant impact on mean position, speed delta, normalized speed, and speed variability, but not position variability. This reflects the fact that both gain settings were acceptable for young healthy participants to continue walking and stay centered on the treadmill despite sensory perturbations being applied. While the magnitude of several metrics was sensitive to gain the relationships between conditions appeared to be preserved for either gain setting. Extension of the study would benefit from including healthy older adults to consider the impact of ageing on sensory integration and gait behavior. The VST approach represents a strong advantage over a standard fixed speed treadmill as it can allow for extended walking with a safety harness, in a natural way that allows speed changes.


Biomechanics, Mechanical Engineering, Variable Speed Treadmill

Rights Statement

Copyright © 2021, author