Enhancing Voluntary Motion in Broad Patient Populations With Modular Powered Orthoses Renewal

The overall goal of this project is to develop modular, lower-limb, powered orthoses that fit to user-specific joints and control torque in a manner that enhances voluntary motion and mitigates musculoskeletal pain. Commercial exoskeletons typically use actuation and control methods that force the human user to follow specific, rigid gait patterns. This has prevented emerging wearable robotics from effectively addressing the weakness and pain associated with mild to moderate impairments, such as knee osteoarthritis (OA). These populations require partial, task-agnostic assistance that works harmoniously with their voluntary motion rather than constraining it. To bridge this gap, this project utilizes a quasi-direct drive actuation paradigm, a high-torque motor combined with a low-ratio transmission, integrated into conventional knee stabilizer and unloader braces. This hardware is uniquely capable of producing large output torques without causing perceptible resistance when backdriven by the human joint. To control the device across various activities of daily living (ADLs) without requiring pre-programmed trajectories, the investigators are developing a neural network-based formulation of "energy shaping" (effectively combining virtual springs, dampers, and gravity/inertia compensation) trained on multi-activity human data. The specific objectives of this study period are to: 1. Integrate the modular quasi-direct drive actuator and miniaturized electronics into both a modified knee stabilizer brace and a modified unloader knee brace. 2. Implement and validate the neural network-based energy shaping controller to provide task-agnostic, biomimetic torque assistance with able-bodied subjects. 3. Establish the clinical feasibility of this technology in individuals with multi-compartment knee OA. The investigators hypothesize that the biomimetic torque assistance provided by these motorized braces will significantly reduce quadriceps effort, knee joint moment loads, and subjective pain across ADLs.