Design and Evaluation Methods for a Lightweight Ankle Exoskeleton for Community-Based Post-Stroke Walking

Abstract

Designing ankle exoskeletons for community-based post-stroke walking requires addressing a wider range of factors than are typically relevant in controlled laboratory studies. To be useful in everyday settings, these systems must deliver meaningful assistance while remaining lightweight, responsive, wearable, and robust to variability across users and walking conditions. This dissertation presents design and evaluation methods for lightweight ankle exoskeletons intended for community use after stroke, with particular focus on actuator performance, iterative design and evaluation cycles, and actuator sizing for heterogeneous users. First, this work develops a systematic method for quantifying power transmission through a wearable ankle exoskeleton. Using on-body walking experiments and benchtop replay of walking trajectories, power flow was measured through the motor, transmission, and human-device interface. This analysis showed that substantial losses occurred at both the motor and the interface, identifying compliance and sensing limitations as key barriers to effective torque delivery. Second, this work presents iterative redesign of ankle exoskeleton hardware to improve on-body assistance. A second-generation actuator improved structural stiffness and sensing fidelity, reducing sensing noise, torque error, and dynamic compliance while enabling adaptive and personalized control studies. A third-generation system then introduced a faster, more compact actuator integrated with a custom orthotist-fabricated ankle-foot orthosis (AFO). In treadmill testing with individuals post-stroke, this redesigned system improved torque tracking and increased peak positive exoskeleton power delivery across walking speeds, while also improving wearability and independent donning potential. Finally, this dissertation develops a data-driven framework for actuator optimization and sizing based on walking biomechanics. This framework shows that co-optimization of motor, transmission, and battery parameters can reduce wearable system mass while preserving required torque-speed capability, and that sizing actuators to user subgroups can provide additional weight savings. Together, these contributions establish a design framework that integrates modeling, benchtop characterization, and on-body evaluation to guide the development of lightweight ankle exoskeletons for community-based post-stroke walking assistance.