ioloizou/g1_locomotion

This is a low-traction research/personal project applying established convex MPC techniques (standardized by the MIT Cheetah and subsequent papers) to the Unitree G1 hardware. With only 12 stars and no recent activity, it lacks a community moat. Frontier labs and hardware manufacturers (like Unitree itself) are rapidly moving toward end-to-end Reinforcement Learning (RL) for humanoid locomotion, which typically outperforms these manual control implementations.