Design, Control, and Perception of Bionic Legs and Exoskeletons
Ph.D. Dissertation Defense
Mechanical Engineering, University of Michigan
11 September 2020
Dissertation Committee Chair: Elliott Rouse
Bionic systems—wearable robots designed to replace, augment, or interact with the human body—have the potential to meaningfully impact quality of life; in particular, lower-limb prostheses and exoskeletons can help people walk faster, better, and safer. From a technical standpoint, there is a high barrier-to-entry to conduct research with bionic systems, limiting the quantity of research done; additionally, the constraints introduced by bionic systems often prohibit accurate measurement of the robot’s output dynamics, limiting the quality of research done. From a scientific standpoint, we have begun to understand how people regulate lower-limb joint impedance (stiffness and damping), but not how they sense and perceive changes in joint impedance. To address these issues, I first present an open-source bionic leg prosthesis; I describe the design and testing process, and demonstrate patients meeting clinical ambulation goals in a rehabilitation hospital. Second, I develop tools to characterize open-loop impedance control systems and show how to achieve accurate impedance control without a torque feedback signal; additionally, I evaluate the efficiency of multiple bionic systems. Finally, I investigate how well people can perceive changes in the damping properties of a robot, similar to an exoskeleton. With this dissertation, I provide technical and scientific advances aimed at accelerating the field of bionics, with the ultimate goal of enabling meaningful impact with bionic systems.