AI Football Robots: The World’s First Autonomous Tournament
July 07, 2025
3 mins
Share
Beijing hosts the world’s first autonomous football tournament | Credit: Booster Robotics
Autonomous humanoid robots powered by AI algorithms are competing without human intervention in China’s first AI robot football tournament in Beijing
In recent years, advancements in AI have propelled robotics, enabling machines to learn, adapt and perform increasingly complex tasks – including mastering the game of football.
Four teams of autonomous humanoid robots have now competed in China’s first AI-powered football tournament in Beijing’s Yizhuang Development Zone.
The event, part of the Robo League robot football tournament, is a unique milestone for real-world AI competition in China.
Booster Robotics, the company that supplied the robot players, provided humanoid units that operated without human control.
Each team fielded three active robots plus a substitute, playing two ten-minute halves with a five-minute break.
The robots in action
The robots utilised AI algorithms and sensor technology to navigate the pitch independently.
Computer vision systems enabled the machines to detect the ball from 65 feet away with more than 90% accuracy.
The robots also recognised teammates, field markings, goals and opponents through optical sensors and cameras.
In the future, we may arrange for robots to play football with humans.
Cheng Hao, Founder and CEO of Booster Robotics
“This is the first fully autonomous AI robot football match in China,” says Dou Jing, Executive Director of the Organising Committee, according to Fox News.
“It represents a combination of technological innovation and industrial application.”
The tournament featured zero human intervention, contrasting with remote-controlled robot competitions.
The humanoid units ran, walked, kicked and made real-time decisions using onboard processing systems.
Despite displaying an awkward gait, the robots operated independently without external commands.
How do the robots work?
University teams adapted the Booster Robotics hardware with their own algorithms.
Tsinghua University’s THU Robotics team defeated the Mountain Sea team from China Agricultural University 5-3 in the final match.
The competition allowed collisions provided they were not malicious.
One co-organiser compared the skill level to that of five- or six-year-old children but expressed optimism that performance would improve with enhanced hardware and AI training.
Cheng Hao, Founder and CEO of Booster Robotics
“In the future, we may arrange for robots to play football with humans. That means we must ensure the robots are completely safe,” says Cheng Hao, Founder and CEO of Booster Robotics, according to the Guardian.
Cheng added that the match represented a step toward integrating intelligent machines into everyday scenarios.
How AI is accelerating the robot’s development
Organisers acknowledged that robots still face challenges, particularly dynamic obstacle avoidance.
However, progress is occurring rapidly in autonomous navigation and decision-making systems.
The competition involved university teams that developed custom algorithms for the standardised robot hardware.
Teams programmed their units with specific tactical approaches and movement patterns.
The robots also demonstrated capabilities in real-time processing, spatial awareness and coordination.
Machine learning (ML) algorithms enabled the units to adapt their behaviour based on game situations and opponent movements.
The controlled environment allows engineers to evaluate performance metrics and identify areas for improvement.
Beijing preparing for 2025 World Humanoid Robot Sports Games
The tournament serves as preparation for the 2025 World Humanoid Robot Sports Games, scheduled for Beijing from 15 to 17 August.
The games will include eleven events modelled after traditional sports, including track and field, gymnastics, football and synchronised dancing.
Beijing previously hosted a humanoid marathon event.
The city’s second humanoid sports event will feature the fastest robot completing a 13-mile course in two hours and 40 minutes, compared to the fastest human time of just over one hour.
Leave a Reply