AEM Seminar:
Developing Generalizable Model-Based Control Methods for Lower-Limb Prostheses
Abstract:Lower-limb powered prostheses hold potential to improve quality of life for 600,000 people with lower-limb amputations by enabling mobility with less human energy than traditional prostheses. However, current prosthesis control strategies require many hours of heuristic tuning for every user and for every behavior. To develop a control approach that applies across users, we construct model-based prosthesis control methods that rely solely on local information by translating formal nonlinear bipedal walking control methods to prostheses. On hardware, we realized the first model-dependent prosthesis controller that utilizes real-time force sensing to complete the model, achieving improved tracking performance across terrains and subjects without additional tuning. Going forward, we will examine how to incorporate biomechanics principles into these model-based formulations to recover natural behavior and improve user
outcomes.
Bio:Dr. Rachel Gehlhar Humann is an Assistant Professor in Mechanical Engineering at the University of Minnesota. Here she leads the Humann Bionics Lab, which seeks to improve generalizability of lower-limb powered prosthesis control methods through model-based control strategies that incorporate human sensor feedback and biomechanics principles. Prior to joining UMN in 2024, she conducted her postdoctoral research at UCLA, exploring the effects of ankle stiffness on walking gait. She completed her M.S. (2018) and Ph.D. (2022) in Mechanical Engineering at the California Institute of Technology where she investigated nonlinear control methods for powered prostheses. Before Caltech, she graduated with a B.S. in mechanical engineering from the University of St. Thomas in St. Paul, MN in 2016.