Ankle-foot orthoses are a category of biomechanical assistive devices intended to address a variety of gait impairments by providing support and stability about the ankle joint. Passive-dynamic AFOs (PD-AFOs) are broadly designed with a foot plate, a cuff that wraps around the shank, and a strut connecting the foot plate and cuff. A PD-AFO’s strut is the primary contributor to its stiffness properties, which can be customized and tuned to match an individual’s level of need. However, the vast majority of PD-AFOs are manufactured with a single, unchanging strut stiffness. Recent work examining the stiffness properties of the ankle joint during walking have shown that the ankle does not exhibit a single stiffness property, but changes throughout the gait cycle and changes with walking speed. Specifically, ankle joint stiffness during is low at the beginning of the loading phase of stance and progressively increases throughout the loading phase. Thus, a mismatch exists between current PD-AFO designs with a single stiffness and the adaptive ankle joint stiffness properties exhibited by healthy walking. This dissertation seeks to characterize these adaptive ankle joint stiffness properties in healthy individuals (Aim 1), design a type of PD-AFO – called a bi-linear stiffness AFO (BL-AFO) – with stiffness properties that closer approximate healthy ankle joint properties (Aim 2), and observe healthy individuals walking with a BL-AFO to understand the fundamental interactions that occur between the BL-AFO and a healthy musculoskeletal system (Aim 3).

Aim 1 results show that modelling healthy ankle joint stiffness as having two stiffnesses during the stance phase of gait – called bi-linear natural ankle quasi-stiffness (BL-NAS) – fits measured ankle joint net moment vs. ankle joint angle data much better than historical models of NAS. Aim 2 results show that a BL-AFO could successfully be designed such that a low stiffness is exhibited at the beginning of deformation, then a higher stiffness is exhibited as the BL-AFO is deformed more. Aim 3 results indicate that healthy individuals who walk with this BL-AFO can off-load some of their physiological function onto the BL-AFO while maintaining a walking speed. Generally, subjects maintained their peak net ankle joint moment (device + physiological ankle), decreased peak ankle joint power during push-off, and increased net ankle joint stiffness during early and late loading. The results of this dissertation indicate that healthy individuals clearly exhibit complex ankle joint stiffness properties during walking, passive devices can be designed that match these complex stiffness properties, and these devices can provide assistance to healthy individuals. Future work includes using these devices with patient populations.