The Effects of Gait Variability and Encouraged Exploration on Adaptation to Assistive Ankle Exoskeletons

Abstract

Exoskeletons have potential applications in the rehabilitation and enhancement of human gait. In the past decade, numerous exoskeletons have reduced the energetic cost of walking, potentially allowing users to walk further and longer before fatiguing. However, achieving these benefits can require hours of experience, during which the user must adapt to the novel and complex dynamics created by the device. This is a nontrivial feat, as gross gait characteristics, joint level kinematics and kinetics, and individual muscle activities may all need to adapt to optimally move in concert with the device. Recent work suggests that high gait variability, achieved through various controller designs, may facilitate this adaptation process by providing experience with different-cost gaits. However, it remains unclear if self-directed gait variability, achieved through exploration, can expedite adaptation to exoskeletons. My purpose is to test how gait variability, whether naturally occuring or encouraged through exploration, effects a users� adaptation to ankle exoskeletons. A secondary purpose is to perform a preliminary investigation as to how risk-taking propensity can influence variability and ultimately adaptation. Many non-human animal studies have demonstrated a boldness-exploration behavioural relationship, where animals who are more risk-tolerant are also more likely to display greater locomotor exploration of a novel environmental context. However, this is untested in humans. To test these hypotheses, I used bilateral ankle exoskeletons to apply assistive torques using a custom real-time angle-based controller. During a baseline trial the exoskeletons were unpowered, allowing habituation and qualification of natural gait variability. Next, the exoskeletons were turned on and I assessed the level of adaptation both before and after an exploration trial where I incited high gait variability by encouraging participants to explore different ways of walking using verbal prompts. I then compared their adaptation magnitude pre- and post-exploration. Risk-taking propensity was measured using a questionnaire. Overall, I achieved 11.5% cost savings with the exoskeletons, though my results suggest that inducing variability through this exploration protocol had no effect. Interestingly, we found that a higher risk-taking propensity was correlated with greater implicit gait variability. These findings provide insight into the role of variability in adaptation to exoskeletons.