Soft Exosuits for Improved Walking Efficiency and Community Based Post-Stroke Gait Rehabilitation

Abstract

For decades, exoskeleton researchers have designed complex weight bearing, multi-joint systems in attempt to reduce the metabolic cost of walking for healthy individuals and aid in gait rehabilitation of stroke survivors with little success. However, in the last decade researchers have shifted focus and begun designing simple, lightweight, and nonrestrictive exoskeletons which target single joint motions, demonstrating preliminary success in aiding both healthy and impaired gait. One type of these new breed of exoskeletons are known as soft exosuits. Soft exosuit use structured functional textiles in combination with a flexible cable-drive actuation scheme to apply forces in parallel to the underlying musculature in order to assist with natural biological motions. This dissertation details the design, control, characterization, lumped parameter modeling, and experimental evaluation of gait assistive soft exosuits for both healthy individuals and stroke survivors.

For healthy individuals, this work explores the key role power and efficiency play in the design and effectiveness of soft exosuits for gait assistance. Initially, the relationship between delivered exosuit power and metabolic power is experimentally explored. One study with a multi-joint (ankle plantarflexion and hip flexion) exosuit demonstrated that delivering higher exosuit power results in higher metabolic power reductions, including up to a 23% metabolic reduction relative to walking without assistance, the highest reduction at the time of its publication. Building on this, another study demonstrates that two entirely different controls approaches with similar delivered exosuit power results in very similar metabolic reductions. A third study isolates the effects of hip flexion assistance, demonstrating the ratio of delivered exosuit power to metabolic power reduction is more efficient for hip flexion than the more commonly assisted joint motions of ankle plantarflexion and hip extension. Combining these results with other published manuscripts from the broader exoskeleton field, this work refines a simple lumped parameter model to predict the metabolic benefit of a device. Originally introduced by Mooney, Rouse, and Herr in 2014, this model, known as the Augmentation Factor, can estimate the metabolic benefit of an exoskeleton or exosuit based solely on the exosuit power delivered to the wearer and its mass distribution across the body.

Expanding this power analysis, this dissertation explores the static and dynamic characterization of the suit-human interface, a key component of the soft exosuit system that is often overlooked by many researchers. This interface characterization highlights the key role interfaces play in delaying power delivery from the actuator to the human, due to interface energy absorption and return dynamics. Combining both the experimental results and system characterization work, a complete lumped parameter model of the flow of power throughout the entire soft exosuit system is introduced.

Leveraging the knowledge gleaned from the experimental, characterization, and modeling work of healthy exosuits, the final part of this dissertation introduces the design, control, and characterization of a lightweight, autonomous exosuit for community based post-stroke gait rehabilitation. This device is the first wearable robot designed for community based post-stroke gait rehabilitation and its unique antagonistic architecture enabled an actuation scheme with considerably higher torque density than similar wearable devices. A torque controller was developed that could accurately track a desired net exoskeleton torque with RMSE of .4 Nm using this antagonistic architecture. Preliminary evaluation of the immediate and rehabilitative effects of this autonomous ankle exosuit on chronic stroke survivors demonstrates its ability to enable high-intensity overground gait training. The foundational work presented herein demonstrates the potential of exosuit technology to both reduce the metabolic cost of walking for healthy individuals and improve community based post-stroke gait rehabilitation.