Planning and Control for Passive Dynamics Based Walking of 3D Biped Robots

Efficient walking is one of the main goals of research on biped robots. Passive Dynamics Based Walking (PDBW) has been proven to be an efficient pattern in numerous previous approaches to 2D biped walking. The goal of this study is to develop a feasible method for the application of PDBW to 3D robots. First a hybrid control method is presented, where a previously proposed two-point-foot walking pattern is employed to generate a PDBW gait in the sagittal plane and, in the frontal plane, a systematic balance control algorithm is applied including online planning of the landing point of the swing leg and feedback control of the stance foot. Then a multi-space planning structure is proposed to implement the proposed method on a 13-link 3D robot. Related kinematics and planning details of the robot are presented. Furthermore, a simulation of the 13-link biped robot verifies that stable and highly efficient walking can be achieved by the proposed control method. In addition, a number of features of the biped walking, including the transient powers and torques of the joints are explored.