Towards Understanding the Effects of Exoskeletons on Sensory Integration and Step Regulation

Abstract

Robotic exoskeletons have, for a few decades, proven their potential to augment human performance and assist those with mobility impairments. Extensive research has demonstrated that robotic assistance can be used to reduce the energetic cost of walking, improve the dynamic stability of the user, increase preferred walking speed, or augment load-carrying capacity. In the rehabilitation world, exoskeletons have shown that they can assist patients, aid in interventions and improve outcomes. So, why are exoskeletons not yet part of our day-to-day lives? Unlike purely legged robots, exoskeletons involve a human and all the complexities of their sensorimotor system in the control loop. The general idea behind assistance at the lower limbs involves augmenting the joint forces while not interfering with the user's intent. However, the human body is highly redundant, locomotive tasks can be performed through a large range of joint coordination and the way in which locomotion is achieved is closely tied to the objective prioritization of the user. Our exoskeleton devices have been successful in assisting specific tasks but they do not generalize well to everyday activities and often result in highly variable outcomes across populations. In order to design effective controllers that generalize to a real-world setting we must understand the effects of the exoskeleton on the sensorimotor process and how objectives may vary when assisted by an exoskeleton. In this work, I present a novel algorithm for controlling a treadmill that adapts to the user’s walking speed. This method relies on kinematic parameters and significantly improves accuracy compared to previous work, resulting in a controller that does not need position feedback from the user. Using our self-paced treadmill algorithm, we study the effects of assistance at the ankle on the step regulation process in order to identify possible changes in movement objective when walking with an exoskeleton. Some of the variable outcomes observed across studies may arise from users prioritizing objectives other than energetics. Furthermore, I present a first study evaluating how the user integrates the novel sensory information of the exoskeleton during a balancing task and how under certain sensory conditions assistance can be detrimental.