Optimal Gait Synthesis of a Planar Biped

Abstract This paper deals with dynamic optimization of biped locomotion, and is focused on the unipodal phase planning in the sagittal plane. The optimal motion is generated by minimizing the joint actuating torques. It must obey hard state constraints to prevent counter-flexing of the swing leg, as well as collision of the foot with the ground or an obstacle. The minimization technique used to solve this problem is the Pontryagin Maximum Principle. Numerical simulations presented are applied to a biped robot whose mechanical characteristics are closely related to those of the human biped. The patterns of the gait generated are similar to human gait in the sagittal plane.