Effect of Foot Shape on Walking Performance of a Biped Robot Controlled by State Machine

Abstract

Bipedal robot is a multi-degree-of-freedom, high-dimensional, naturally unstable system. The control method based on kinematics and dynamics is complex in theory and implementation, and the control algorithm usually involves many parameters, which is difficult to design. In this paper, a control framework based on a state machine is designed to achieve stable walking of a 3D bipedal robot, which only involves 6 parameters to be designed. In terms of the structural design of biped robots, researcher’s interests are mostly focused on legs, knees, and ankles, and there are few studies on the shape of the robot foot. In this paper, we build a three-dimensional biped robot model in Webots and use random searching method to find the control parameters that lead to stable walking. For the stable walking gaits, we compare the performance of five foot shapes in terms of the walking style, control efficiency, and stability. We found that the yaw angle is a key factor affecting the diversity of the robot’s gait. In addition, it is found that the overall performance of the flat foot is most satisfying. The research in this paper can be helpful for the bipedal robot walking algorithm and the design of the foot shape.