ABSTRACT

This Faculty Early Career Development Program (CAREER) award supports research that will advance understanding of human walking and create a new type of lower-body exoskeleton for mitigation of disability and augmentation of human performance. When people walk normally, two factors largely determine their balance and forward progress: the location of each footstep, and how hard each leg in stance pushes on the ground. In general, when an exoskeleton produces forces that are slightly different from what the person wearing it expects or wants, it affects how the person takes their next step, which in turn affects the future exoskeleton response. The exoskeletons studied in this project supplement the natural muscular strength of their wearers by adding to the support forces during walking. Systematically varying the exoskeleton force and measuring the response will reveal a map of the dynamics of balance and movement during walking, including how these change on uphill and downhill slopes. This new fundamental understanding of walking can then be used by the exoskeleton to adapt to individual characteristics of any user, and from there to stimulate desired walking patterns while decreasing the likelihood of a fall. This research will be complemented by educational outreach, including informing the broader public about robotics and human-robot interaction via science videos on social media, and engaging pre-college students with workshops on biomechanics and exoskeleton control.

This project is centered on a new type of exoskeleton that uses a single prismatic actuator on each leg to apply force from the ground next to a planted foot, directly to the wearer's center of mass. The actuator can provide up to 80 percent of the unassisted ground reaction force during walking. The exoskeleton hip is unactuated and free to rotate, which allows the wearer to choose their footstep locations as though the exoskeleton was absent. Human subject experiments will study how the human kinematics, muscle activity, metabolic cost, foot placement, and stability are affected by different forces on the person's center of mass under various conditions including varying speeds and slopes -- essentially accomplishing system identification of the coupled human-robot system. These system identification results, together with the divergent component of motion (DCM) framework and machine learning, will define an exoskeleton controller to provide optimized walking assist forces based on the wearer's current state and expected future footstep locations. The project will subsequently investigate the bi-directional adaptation between the human and exoskeleton: how humans adapt to the exoskeleton forces, how their subsequent footstep locations are affected by these forces, and how the exoskeleton can best adapt to an individual human and their personal locomotion patterns and footstep placements.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Please report errors in award information by writing to: awardsearch@nsf.gov.