The lower limb is the most frequently injured body region of occupants during automotive frontal crashes. In particular, trauma to the ankle and subtalar joints is among the most common and severe of lower limb injuries. While the injury mechanisms of the ankle and subtalar joints during vehicle accidents have been previously investigated, existing experimental studies have been limited due to the variability of post-mortem human specimen (PMHS), the simplified loading conditions, and the relatively high experiment costs.

The primary objective of this dissertation is to develop a detailed and realistic numerical model of the human hindfoot capable of predicting the ligamentous and bony injuries observed in the ankle and subtalar joints. A finite element (FE) model of the foot and ankle was developed based on the reconstructed geometry of a male volunteer close to the anthropometry of a 50 th percentile male and a commercial anatomical database. While the forefoot bones were defined as rigid bodies connected by ligament models, the ankle and subtalar joints together with their surrounding bones and the leg were modeled as deformable bodies. The material and structural properties were selected based on a synthesis on current knowledge of the constitutive models for each tissue. Failure of the tissues was assessed using local strain measures in the bone and ligament. The FE model was validated against data recorded during tests with human volunteers and PMHS in various loading modes (flexion, xversion, axial rotation, and axial loading). Good overall agreements were observed in the target response comparisons. The predicted injury characteristics corresponds to acceptable injury risks (19 ~ 37 %) based on current injury risk functions. While physical dummies and PMHS are frequently used in the crash tests, the interpretation of their response and injury tolerance of the foot and ankle are challenging. Thus, a secondary objective of this study is to construct an injury hyper- surface of ankle and subtalar joints using the validated FE model, which may help in analyzing the experimental data. Complex loading conditions involving combinations of simultaneous loading modes were evaluated and the tissue failure was assessed. The strain distributions of the model were consistent with injury patterns observed in the experiments. The motion and loading design space of the ankle and subtalar joints was discretized and the injury subspace was determined from the FE simulations. Response surface methodology was applied to determine injury boundaries (hyper-surface). The capability of an injury surface to predict an injury relative to the direct prediction from the tissue strain measures in the FE model was evaluated under vehicle crash conditions. Non-injury condition was correctly identified as the maximum tibial force was far from the injury boundary (less than 50 % of the fracture force), while bony fracture was predicted when the maximum tibia force was close to the injury boundary (92 % of the failure level estimated by the boundary). According to the computational results, it is proved that the injury surface can be a tool to decide injury or non-injury in certain loading combinations. Used in conjunction with either dummy test data or multi-body models obtained from frontal crash tests, the injury hyper-surfaces and hindfoot FE model enable the development of injury countermeasures that minimize the risk of ankle and subtalar injuries.