Integrating Visual Biofeedback in Exoskeleton-Assisted Training of Propulsion

• Sophia May, Biomedical Engineering, University of Delaware

• Sophia May, Biomedical Engineering, University of Delaware

Propulsion is a critical component of walking that is affected by stroke and targeted in post-stroke rehabilitation. One of the methods for training propulsion is modifying the posture of the limb during push-off. This effect can be achieved implicitly using exoskeletons, or explicitly using visual biofeedback. While these paradigms have been used separately, the combination of these two methods has not been adequately studied. This project aims to use visual feedback during exoskeleton-assisted walking to cue participants to modify their maximum hip extension angle. Using MATLAB, Simulink and Quanser, we developed a visualization of a vertical bar representing the user’s maximum hip extension angle for the right hip and two horizontal lines displaying the target range. The bar is green when the user is within the range and red when outside. Five young adult participants were asked to complete four sessions of walking wearing a hip exoskeleton to measure their hip extension angle. No assistance from the exoskeleton was given. Session one was a two minute familiarization session to find the participant’s baseline. During sessions two and four, participants were asked to increase their maximum hip extension angle by five degrees relative to the baseline without feedback. In session three, participants were given feedback to help modify how close their maximum hip extension angle was to the target. Two-three minutes rest was given between each session. A matched-pairs t-test compares the average maximum hip extension angle using feedback and target (n=5, 95% CI), p = 0.703 > 0.05. Therefore, there is no statistical significance, indicating feedback is successful in training propulsion. Comparing pre/post feedback to target, p = 0.1413 (pre) & p = 0.6449 (post), difference = 0.5036, suggesting participants improved after using feedback. This visualization can be adapted for further studies in our lab including measuring hip flexion.