The goal of my research is to develop ankle-foot prosthesis controllers that reduce balance-related effort during walking. Although great progress has been made in ankle foot prostheses, individuals with below knee amputation still report difficulty with balance. While the effective balancing method of foot placement is unavailable in ankle-foot prosthesis, the balance restoring resource of ankle actuation holds potential for amputee walking. I explored the possibilities for prosthetic foot designs to improve balance through simulation studies, hardware development, and human subject experiments. I demonstrate that ankle actuation control can be very important in balance maintenance, and present two new approaches to reduce balance-related effort for people with lower limb amputations.

Through a simulation of three-dimensional limit cycle walking of amputee gait, I demonstrate that ankle actuation can be equally effective as foot placement, especially in once-per-step modulation of ankle push-off work. I implemented the ankle push-off work controller in an ankle-foot prosthesis emulator and tested the controller on human subjects. I found that with this push-off work controller, both able-bodied subjects with simulated amputation and individuals with below knee amputation reduced balance-related effort. One possible explanation of amputee’s reduced metabolic rate could be their reduced intact limb control effort during stance phase. In addition, more training seemed to help amputee participants realize the benefits of the controller.

Simulation results also suggest that inversion/eversion control could improve balance. To test control ideas, I developed a two degree-of-freedom ankle-foot prosthesis with plantarflexion and ankle inversion/eversion. Using this device, I investigated the balancing effect of passive ankle inversion/eversion stiffness and active once-per-step modulation of inversion/eversion torque. The inversion stiffness strongly affected amputee’s balance-related effort. Active inversion controller lowered metabolic rate, a balance-related effort indicator. While these step-to-step variations in ankle/inversion torque reduced balance-related efforts, these effects were not as effective as those of the ankle push-off work control.

The results from these simulation studies and human experiments suggest that step-to-step alteration in ankle actuation can reduce balance-related effort. This finding will help inform future design of prosthetic devices, which could reduce balance-related effort, increase balance confidence, and improve overall quality of life.