Locomotion Control on Human-Centaur System with Spherical Joint Interaction

This paper presents a locomotion controller for a novel human-augmented legged robot, the Centaur robot, which is primarily developed to extend the human's ability to carry load. For such a human-robot walking system, there are requirements for the robot to maintain a balanced posture, provide a proper range of motion for human, and walk in omnidirection with human. In this paper, we propose a locomotion planning and control framework that attempts to fulfill these requirements. The reduced-order dynamics model-based controller is utilized to regulate the robot's posture and height. A torso state planner combined with the spherical joint interaction is proposed to allow human's trunk motion to be independent of the robot's torso. To further enable the Centaur robot to accommodate human's walking, an omni-directional walking strategy, driven by human's real-time movement, is presented. Experiments are conducted to verify the methods. Results show the robot can walk with human subject while retaining the mean absolute error of posture angles (roll, pitch, and yaw) at 1.50 pm 0.23°, 4.20 pm 2.39°, and 5.79 pm 2.13°. The methods provide the human with moderate degrees of freedom in rotational and vertical motion when interacting with the robot. The walking planning approach enables the wearable legged robot to walk forward, backward, laterally, and turn around with the wearer.