End-to-end Impedance Modeling Towards Task-Invariant Control of Powered Prosthetic Legs

Millions of Americans live with lower-limb loss, and powered prosthetic legs can return joint functionality via effectively designed control paradigms. While individual control approaches have shown effectiveness for multiple primary activities of daily living (ADLs), such as walking on flat ground, ramps, or stairs, the integration of these approaches into a single unified controller has remained a research challenge. State-of-the-art approaches use classification strategies to discretely change between various task-specific controllers, but the susceptibility of this approach to misclassification errors, its finite ability to span ADLs, and its unnaturally abrupt nature render this approach undesirable. No lower-limb prosthesis controller currently exists that can enable the primary ADLs continuously and without classification. As such, the objective of this project is two-fold: to discover explicit models for the relationship between lower-limb kinematics and desired prosthetic joint impedance parameters (stiffness, damping, and equilibrium angle), and to use these models to develop a novel end-to-end impedance control paradigm that can allow individuals with transfemoral amputation to seamlessly, accurately, and reliably span all primary ADLs. Encouraged by tangential relationships in the literature and the preliminary data presented in this proposal, I hypothesize that at least one model can be discovered for each impedance parameter of the prosthetic knee and ankle to implement such a control scheme. Specifically, Aim 1 of this proposal will focus on the model discovery using a publicly available dataset of human locomotion across a wide range of ADLs. Then, in Aim 2, the discovered models and related findings will be used to inform the experimental control of a powered knee-ankle prosthesis under a variety of conditions. Successful completion of these aims will contribute to enhanced understanding of human limb coordination and biological joint impedance, as well as the development and validation of a novel prosthetic control paradigm that can greatly enhance the performance of powered prostheses in clinical and real-world environments. These outcomes have the potential to improve the mobility, independence, and quality of life of individuals with transfemoral amputation and can lead to complimentary contributions for transtibial prostheses and lower-limb exoskeletons. Additionally, the proposed technical and professional training will prepare the PI for a successful career as a tenure-track professor at a research-intensive university, with focus in the fields of lower-limb prosthetic and exoskeleton advancement. Project Number: 5F32HD116414-02 | Fiscal Year: 2026 | NIH Institute/Center: Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) | Principal Investigator: Ryan Posh | Institution: UNIVERSITY OF MICHIGAN AT ANN ARBOR, ANN ARBOR, MI | Award Amount: $32,130 | Activity Code: F32 | Study Section: Special Emphasis Panel[ZRG1-F10B-Q(20)L] View on NIH RePORTER: https://reporter.nih.gov/project-details/5F32HD11641402