Humanoid robot learns human-like soccer skills with a unified system
Playing Soccer Like Humans Do: Learning Versatile and Agile Humanoid Soccer via Latent Motion Priors
Robotics
Summary
Playing soccer like a human is hard for robots because it needs smooth movement, ball control, and quick contact actions. The authors developed a system called soccerbot that learns these skills from real human soccer motion data. This lets the robot perform many soccer moves naturally and switch between them smoothly. They tested this system on a real humanoid robot that could play soccer skills well without extra training.
What this means in practice
- •For robot developers: Build humanoid robots that execute multiple soccer skills with natural, smooth motion using learned human motion data.
- •For robotics simulation engineers: Design robust simulations that support transferring complex motion policies from simulation to real humanoid robots without retraining.
Authors
Haoyu Shao, Yunshen Wang, Jingze Zhang, Shaohang Zhu, Yong-Lu Li, Tianzhu Zhang, Baoxiong Jia, Siyuan Huang
Abstract
Humanoid soccer requires a robot to seamlessly combine agile locomotion, ball manipulation, and impact-rich contact dynamics across a continuous, dynamic behavioral spectrum. Yet existing approaches either cover only a limited set of skills, or forego motion priors entirely and produce unnatural behaviors, leaving the combination of a unified skill set and human-like motion quality unaddressed. We present soccerbot, a hierarchical reinforcement learning framework that resolves both limitations through a structured latent motion prior distilled from large-scale human soccer motion capture data. By constraining policy learning within this dynamically plausible motion manifold, a single skill-conditioned policy naturally learns to perform moving, dribbling, trapping, and kicking in a unified formulation, while inherently producing human-like motion and fluid transitions during commanded skill switching. Augmented by robustness-oriented simulation design, soccerbot transfers zero-shot to the Unitree G1 humanoid robot, demonstrating agile, natural, and robust execution of the full soccer skill repertoire under real-world conditions.