Three-dimensional variable center of mass height biped walking using a new model and nonlinear model predictive control

This paper presents a trajectory generation algorithm for a three-dimensional (3D) biped robot that can adjust the center of mass (CoM) according to the environment. We adopt a new abstract model that supports vertical motion and rotation. Differing from traditional abstract models, the proposed full centroid dynamics inverted pendulum model fully considers the robot's movement and rotation. Unlike the zero moment point (ZMP), which only ensures the feet do not flip over, we also propose a new additional stability criterion, named zero frictional moment point (ZFMP), guaranteeing no yaw rotation while walking. Next, a nonlinear model predictive control is designed to generate the CoM trajectory, torso rotational angle, and adaptive footholds to induce various biped gaits. A full -dynamics 3D humanoid robot is simulated to test the proposed method while steering, walking underneath a low door, and walking with disturbances.