Dynamic Locomotion of a Quadruped Robot with Active Spine via Model Predictive Control (ICRA 2023)

This paper presents a general model predictive control approach for high-dynamic torque-controlled spined quadruped robots, considering the effects of spine joint movements and variable inertia. The proposed control method supports the spined quadruped robot in arbitrarily changing its spinal posture during high-speed motion. We have developed Yat-sen Lion, a quadruped robot with a 3-DOF active spine, and validated the proposed control method through experiments on both simulation and physical platforms.