SUMMARY

The field of human augmentation has been studied for decades to expand the limitations of human mobility. Specifically, hip exoskeletons have started to gain attention from different research groups, mainly due to the nature of the hip joint generating large amounts of mechanical power during locomotion. Additionally, hip exoskeleton application can be expanded to clinical populations such as stroke patients who tend to overtax the hip joint to compensate for lack of coordination and weakness in the distal muscles of the leg. Recent hip exoskeleton studies show promising results of benefiting the user both energetically and biomechanically. However, most of these studies are extremely tailored for specific applications and users, which illustrates a huge gap in translating the control framework to a real-world scenario. Additionally, existing exoskeleton systems have limited high-level understanding of user state information, such as walking speed and ambulation mode, precluding exoskeleton assistance to adapt to the changes in user’s biomechanical demand during various locomotor tasks. Therefore, current exoskeleton assistance strategies are nonoptimal and need to be investigated further to develop an adaptive controller that can accommodate the dynamic changes of the user’s state as well as variant subject-dependent information. This work focuses on three key research objectives: 1) Explore the optimal hip exoskeleton design approach for maximal human exoskeleton performance during wide ranges of locomotor tasks, 2) Understand the contributions of sensor fusion-based user state estimation in improving the hip exoskeleton controls over a simulated community terrain, and 3) Quantify the biomechanical and clinical effects of the hip exoskeleton in improving the stroke patient’s community ambulation capability. The study findings provide valuable information for the future exoskeleton designer in developing a more efficient exoskeleton system.