Understanding the human contribution to the human-exoskeleton system
Abstract/Contents
- Abstract
- Exoskeletons have the potential to enhance mobility. Current state of the art exoskeletons have reduced the energy cost of walking and running in several different contexts and have improved functional gait metrics in people with movement disorders. A persistent challenge in designing assistance strategies is an incomplete understanding of the human operator. Unlike prosthetic devices which replace anatomical structures, exoskeletons act in parallel with the person they are assisting, so the efficacy of the device is completely determined by how the user interacts with it. This interaction likely varies on an individual basis and depends on how familiar the user is with the device. Therefore, in order to evaluate how to best assist people, we need to first improve our understanding of how people interact with exoskeletons. This thesis describes a series of studies probing the human contribution to human-exoskeleton performance. The first is an extensive experiment to assess different factors involved in reducing the energy cost of walking with ankle exoskeletons. An existing algorithm to optimize device parameters was used in a training protocol to test the relative contributions of human motor adaptation and parameter customization to improve energy economy. We discovered that traditional training protocols for ankle exoskeletons should be much longer, moderate variation speeds adaptation while too much variation inhibits learning, and customization can improve outcomes. For the second study, we characterized the biomechanical response to exoskeletons in adapted users. Ankle plantarflexion assistance led to reduced plantarflexor activity and increased total ankle work. While other exoskeletons resulted in increased tibialis anterior activity, co-contraction at the ankle in fully trained users returned to unassisted walking levels, thus eliminating one mechanism for increased metabolic cost. We were unable to wholly assign the metabolic cost reductions to group-level changes in joint mechanics and muscle activity in the lower limbs, although qualitative changes to proximal joint mechanics and large intersubject variability indicated a need for further analysis into individual differences in assisted gait. The final study was a robotic application to understand dual tasking with periodic movements. A bipedal robot was controlled to simultaneously balance and stably bounce a ball. Two low-level controllers, one optimizing for ground contact forces and one adjusting leg length to control for height, achieved stable juggling in simulation, and the latter produced juggling in experiment. This study illustrates a method for characterizing how people balance when periodically interacting with a device, by controlling the forces at the distal joint to prioritize balance or by modulating leg length to react to external devices. These results deepen our understanding of how people interact with assistive devices and illustrate the need for more research on the human component of the human-exoskeleton system. The difficulties exoskeleton researchers have faced in improving mobility may have been the result of inadequate training rather than a failure in device design or control. Future implementations of assistive devices for walking would benefit from increased training time or revised training protocols based on the characterized response of adapted users. With sufficient training and optimized assistance, exoskeletons can deliver very large improvements to locomotor performance.
Description
| Type of resource | text |
|---|---|
| Form | electronic resource; remote; computer; online resource |
| Extent | 1 online resource. |
| Place | California |
| Place | [Stanford, California] |
| Publisher | [Stanford University] |
| Copyright date | 2021; ©2021 |
| Publication date | 2021; 2021 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Author | Poggensee, Katherine Lin |
|---|---|
| Degree supervisor | Collins, Steve (Steven Hartley) |
| Degree supervisor | Sreenath, Koushil |
| Thesis advisor | Collins, Steve (Steven Hartley) |
| Thesis advisor | Sreenath, Koushil |
| Thesis advisor | Brunskill, Emma |
| Degree committee member | Brunskill, Emma |
| Associated with | Stanford University, Department of Mechanical Engineering |
Subjects
| Genre | Theses |
|---|---|
| Genre | Text |
Bibliographic information
| Statement of responsibility | Katherine Lin Poggensee. |
|---|---|
| Note | Submitted to the Department of Mechanical Engineering. |
| Thesis | Thesis Ph.D. Stanford University 2021. |
| Location | https://purl.stanford.edu/fy671rd8583 |
Access conditions
- Copyright
- © 2021 by Katherine Lin Poggensee
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...