Robot Football in China: AI-Driven Teams Make History on the Pitch
- Jun 29
- 2 min read
On June 28, 2025, Beijing witnessed a watershed moment in humanoid robot sports: four AI-powered, fully autonomous robots competed in a 3-on-3 football match at the inaugural RoboLeague—an official test for the upcoming World Humanoid Robot Games 2025 . Știrinoi.com reports on the event's major highlights, technological implications, and expert perspectives.
1. Event Summary
Held in Yizhuang, the match featured two 10-minute halves with a 5-minute break . The robots, mounted on T1 hardware from Booster Robotics, used optical cameras and sensors to achieve ball detection accuracy above 90%, even at distances up to 20 meters . Each university team—Tsinghua, China Agricultural, and Beijing InfoSciTech—coded its own AI strategies.
In the finals, THU Robotics (Tsinghua University) triumphed over China Agricultural University's Mountain Sea with a 5-3 win . The tournament implemented lenient rules on collisions, minimizing disruptions .
2. Interesting Facts
First fully autonomous 3v3 match in China: milestone ahead of World Humanoid Robot Games .
T1 humanoid robots: stature of 1.2–1.5 m, packed with advanced perception tech .
High-accuracy ball detection: AI-powered deep reinforcement learning delivers over 90% precision .
Self-righting capability: robots can get up after falling—key for game flow .
Real-world application: performance on endurance, coordination, and tactical execution showcase robotic autonomy .
3. Expert Opinions
Cheng Hao (Booster Robotics CEO): “Sports provide an ideal testbed. Ensuring safety in future human–robot games is essential” .
Dou Jing (Organizer): “It’s a window into real-world application and public engagement with robotics” .
Audience reactions: Viewers described the match as more thrilling than China's men’s football games .
4. Tech Perspectives
Events like this integrate robust hardware (T1) with high-level AI (deep RL, computer vision), shaping innovations in humanoid autonomy and human-interactive robots . University teams are embracing multi-agent RL frameworks to coordinate robot teams, even enabling quadrupedal robot soccer . China’s enriched robotics ecosystem, showcased in competitions like RoboMaster, further accelerates progress .
Comments