Wearable ExoNETs for Personalized Multijoint Actuation

Machine-assisted therapy has shown to be reasonable for enhancing physical outcomes for stroke patients, while costs continue to rise, funding has been “capped”, and research supports lengthy, task-specific therapy. Past robotic rehabilitation research has focused on active robotic devices where users are physically guided by the robot to accomplish a specific movement and clinicians serve as simple bystanders. Although devices have been shown to motivate patients to practice and to have a positive therapeutic advantage, subjects do not typically improve if the technology dominates, mainly because active guidance leads to laziness or a lack of motivation. Here we provide a simplified approach to robotics development, removing the dependency on electronics, sensors, computers, and power supply. The ExoNET (Exoskeletal Network of Elastic Torque) solution provides a soft, biomimetic, and elastic alternative to robotics that embodies intelligence within mechanical design using springs with custom-tuned parameters via optimization. A simple reconfigurable system that can not only assist performance, it also makes therapeutic training easier, faster, and more complete. This thesis first demonstrates and validates a theoretical framework for personalized ExoNETs for several rehabilitation applications. It then designs and successfully shows reduced activity and unhindered motion in a wearable anti-gravity ExoNET prototype on 10 unimpaired individuals. Finally, it presents safety, feasibility, and efficacy results of the ExoNET in an early clinical trial that delivers ExoNET gravity compensation during therapy on 20 post-stroke individuals. The ExoNET has the potential to substantially impact the current state of post-stroke rehabilitative care by marrying simplicity in design with customization in application, aligning closely with the real-world requirements of both clinicians and patients.