Effect of foot shape on locomotion of active biped robots

This paper investigates the effect of foot shape on biped locomotion. In particular, we consider planar biped robots whose feet are composed of curved surfaces at toe and heel and a flat section between them. We developed an algorithm that can optimize the gait pattern for a set of foot shape, walk speed and step length. The optimization is formulated based on the rigid-body and collision dynamics of the robot model and tries to minimize the ankle torque. We also divide a step into two phases at collision events and optimize each phase separately with appropriate boundary conditions. Numerical experiments using walk parameters from human motion capture data suggest that having a curved toe and heel would be a way to realize locomotion at speeds comparable to human.