Multi-objective user tunable interface for assistance control of a lower limb exoskeleton

Abstract

The field of assistive lower limb exoskeletons lacks controllers that allow user adjustment according to their needs and desires. This thesis develops and shows simulation evidence for allowing user-intuitive control by providing adjustment of gait based on gait performance measures the user cares about while walking. Using the NSGA-II multiobjective optimization algorithm to generate trajectories for a virtual constraint-based exoskeleton system a lookup table was generated which provides user adjustment of speed, comfort, effort proxy, and natural walking measures. The findings in this thesis demonstrate a variety of gait across these performance measures which can be used to formulate a user-adjustable controller.