SUMMARY

Exoskeleton technology holds a large potential to help improve human mobility and physical capability. Most of the previous exoskeleton studies focused on targeting the hip or the ankle joint due to their large contributions to power generation during level walking. However, the knee joint still plays an important role during slope walking in terms of mechanical power, and there are still gaps in the field in understanding how to assist the user more effectively using a powered knee exoskeleton during slope walking. Translating exoskeleton technology from a constrained laboratory setting to real-world locomotion environments, in which variations in the walking condition are introduced, remains challenging in the research field. Patients walking in genu recurvatum gait or crouch gait are another potential population that can benefit from a powered knee exoskeleton as the assistance can help the user achieve an improved walking pattern. In addition, whether repetitive gait training using a knee exoskeleton would yield a therapeutic effect on patients' walking is unknown. The main objective of this thesis work is to create intelligent controllers for the knee exoskeleton and evaluate the effectiveness of powered assistance for enhancing human locomotion for both able-bodied adults and the patient population. The thesis study focuses on three key objectives: 1) Develop and validate a robotic knee exoskeleton and evaluate the biomechanical effectiveness of assistance strategies for able-bodied adults during incline walking, 2) Create and evaluate a user-independent, speed-invariant, slope-adjustable controller for the knee exoskeleton using a machine learning approach, 3) Explore the use of a robotic knee exoskeleton for improving gait quality for patients walking in abnormal walking patterns centered around the knee joint.