Online Non-linear Centroidal MPC for Humanoid Robot Locomotion with Step Adjustment

The project is a specialized academic implementation of a control strategy for humanoid robots. While the underlying math is highly complex, the project has zero stars and no community momentum after four years, functioning primarily as a reference for a published paper. Frontier labs have largely moved toward end-to-end Reinforcement Learning (RL) for locomotion, making classical MPC a niche domain for hardware-specific research.