Ankle Actuation for Limit Cycle Walkers

Limit Cycle Walkers are bipeds that exhibit a stable cyclic gait without requiring local controllability at all times during gait. Well-known example are McGeer's “Passive Dynamic Walkers”, but the concept expands to actuated bipeds as involved in this study. Current state-of-the-art Limit Cycle Walkers excel in being very energy efficient, but their ability to handle disturbances (i.e. disturbance rejection) is still limited. A way to improve this ability while maintaining low energy consumption is the use of ankle actuation, which has so far seen few applications in this type of walker. In this paper we study the effect of (1) applying (passive) stiffness in the ankle joint, (2) applying control in the stance ankle based only on local sensor information and (3) modulating ankle push-off. For all three strategies the paper shows how they influence energy use and disturbance rejection of a simple point mass walking model, a more realistic model and a physical prototype. We find that applying a passive ankle spring that results in premature heel rise is energetically optimal and gives an actuation pattern that largely resembles that of humans. Local stance ankle control and ankle push-off modulation can improve the disturbance rejection of a Limit Cycle Walker by at least 60%, without increasing its energy use. These findings are substantiated by showing that our prototype is able to handle large disturbances such as a step-down of 5% of its leg length, while walking efficiently at a mechanical cost of transport of 0.09.