This dissertation informs the design of assistive technology and rehabilitative techniques for improving walking function in clinical populations. Ankle-based exoskeletons have the potential to increase impaired limb function, reduce energetic requirements, and improve ambulation and quality of life. However, previously researched assistive ankle-based robotics have not impacted metabolic cost consistently, perhaps because walking speed was restricted below the speed where assistance was advantageous. We investigated impact of a novel speed-adaptive myoelectric exoskeleton applying assistance to the paretic ankle of six persons post-stroke walking at increasing walking speeds. Our exoskeleton controller successfully increased assistance with walking speed, and we observed increased paretic ankle joint power as well as total limb power. Interestingly, integrated paretic ground reaction forces decreased and there were no observable metabolic benefits. Reductions in paretic limb trailing limb angle were reduced with exoskeleton assistance suggesting that suboptimal limb poster may have limited joint level benefits from propagating to whole body walking improvements.

Secondly, we investigated the independent impacts of reducing lower limb joint motion on resulting gait deviations and metabolic consequences. Ankle and knee motion can be reduced by injury or disease-induced impairments and restrictions in motion at each joint are associated with metabolically expensive compensatory strategies. It is difficult to isolate the independent impacts of reducing ankle or knee motion because lower limb joint motion is neuromechanically coupled. Ankle and knee bracing were used to unilaterally reduce ankle motion (restricted-ank), knee motion (restricted-knee), and ankle and knee motion simultaneously (restricted-a+k) while 15 unimpaired participants walked on an instrumented treadmill. Restricting ankle motion resulted in decreased peak propulsion relative to the braced condition, and ipsilateral hip hiking when knee motion was restricted increased ipsilateral circumduction relative to the restricted-ank condition. Ankle restriction increased energy requirements compared to the braced condition, and simultaneous restriction of the ankle and knee was more expensive than restricting the knee. We reproduced gait deviations similar to clinical populations and results indicate that ankle-based rehabilitation has potential to improve metabolic outcomes.

Next, we investigated the interaction between walking asymmetry and metabolic cost by unilaterally (asymmetric) and bilaterally (symmetric) restricting the ankle, knee, and simultaneous ankle & knee joint motion in nine unimpaired participants. Asymmetry was not inherently metabolically expensive since metabolic cost increased with symmetric restrictions compared to asymmetric restrictions. The number of restricted joints or degrees of freedom correlated significantly with metabolic rate for 7 of the 9 participants, and this correlation accounted for between 63 and 96% of the variability in metabolic data. Thus, rehabilitation focusing on improving impaired limb function rather than symmetry metrics may have more promise to reduce energetic requirements.

Finally, we investigated impact of reduced joint motion and walking asymmetry on joint reaction forces associated with increased incidence of comorbidities including osteoarthritis and joint pain. Personalized musculoskeletal models were used with the computed muscle control algorithm to determine the joint reaction forces at the ankle, knee, and hip, with recorded EMG from six muscles used to constrained simulated muscle activities. Knee restriction resulted in increased limb loading quantified by ground reaction force peak and loading rate ipsilaterally but decreased peak ground reaction forces contralaterally in comparison to the braced condition. Ground reaction force peak and loading rate increased with symmetric when compared to asymmetric restriction. We did not observe increases in joint reaction forces associated with increased limb loading;​ instead, a reduction in muscle forces during loading response counteracted changes in limb loading such that joint reaction forces were relatively unchanged. Our work demonstrates reduced muscle contributions can offset increased limb loading such that joint reaction forces are unaffected