Individualizing a Powered Prosthetic Leg Over Continuously-Varying Tasks
MetadataShow full item record
On a day-to-day basis, able-bodied people traverse a number of different terrains, moving at a self-selected speed and walking up and down inclines in reaction to the environment around them. The transition from one task to another is seamless and our body adapts its locomotion to meet each task without much thought. Lower-limb amputees do not have that luxury. Current technology often does not allow for amputees to change the motion, or kinematic output, of the leg to adapt to the environment around them, adding a level of difficulty to amputees’ lives and limiting their mobility. Clinical individualization of a prosthesis allows for amputees to maximize their walking performance and achieve a comfortable gait closer to the able-bodied norm. In passive prostheses, this entails a prosthetist manually adjusting the hardware to change the motion of the leg throughout the gait cycle. Powered, robotic prostheses have potential to emulate the adaptability of an intact leg, but require an engineer to change the settings with the help of a prosthetist, which is a time-consuming and arduous task. This thesis presents research allowing for easy and intuitive clinical individualization of powered prostheses that adapt to continuously varying walking tasks. First, the development of an intuitive clinical control interface, which allows a prosthetist to quickly tune a powered prosthesis for level-ground walking without the need of an engineer. Next, a tuning system was developed which individualizes a kinematic model of continuously varying walking tasks given the tuning of one level-ground task. Lastly, a 10-subject dataset was collected to expand the continuously varying model to include new tasks, such as walk to run transitions, running over level-ground, stair ascent and descent, and sit-to-stand.