Towards Adaptive System Behavior and Learning Processes for Active Exoskeletons

Abstract

Industrial workers still face work-related musculoskeletal disorders daily and therefore physical support systems like exoskeletons are being developed. Making these wearable robots adaptable to different tasks and users in terms of its support characteristics is expected to generate greater performance and broader acceptance. By analyzing relevant elements of joint tasks in groups of humans and the environment exoskeletons are typically being used in, this paper derives the need for a framework allowing for adaption of the exoskeleton to the task, but also predictability for the user of the exoskeleton. A situation aware gain-scheduling controller with internal state feedback to the user is proposed as a means for adaption and predictability.