With the FIFA Club World Cup in full swing, we spoke to Steven van Roon from the Robotics Centre. What makes football so difficult for robots? And what does that tell us about the future of robotics, AI and autonomous systems?
Why is football so difficult for robots
According to Steven, robot football combines two major challenges. “A robot has to be physically capable of playing the game, understand what is happening and respond accordingly,” he says. “It has to move, perceive, decide, anticipate and cooperate.”
That starts with movement itself. A robot needs to be fast, stable and dynamic. Humanoid robots in particular face significant challenges. Walking, turning and balancing already demand a lot from the system. Even kicking a ball turns out to be surprisingly complicated. While standing on one leg, the robot has to control its body and apply exactly the right amount of force.
At the same time, a robot has to continuously understand what is happening around it. It needs to know where the ball is, where it is positioned itself and how teammates and opponents are moving. It then has to determine the best course of action. A decision that humans make in a split second requires a complex interplay of sensors, software and computing power in a robot.
More than just seeing
With cameras and sensors, robots can gather information. But recognising the ball is not the same as understanding the game. Humans unconsciously filter information at a remarkable speed. We automatically ignore the crowd, the noise and the lines on the pitch, allowing us to focus on what really matters.
According to Steven, that is far less obvious for robots. “In principle, a robot sees everything, but it still has to learn what is important. Humans do this automatically.” Robots are normally programmed to avoid obstacles. On a football pitch, however, they need to actively move towards the ball. Sometimes they are allowed to pressure an opponent, but not too aggressively. “So the question is not only what a robot sees,” Steven says. “It is mainly about how it interprets that information and which action should follow.”
The role of AI
Artificial intelligence helps robots recognise patterns and analyse situations more quickly. Through simulations, robotic systems can evaluate thousands of scenarios. Where are the players positioned? Which options are available? Which strategy is most likely to succeed?
Football video games such as EA Sports FC, formerly known as FIFA, are also interesting in this respect. To make such a game realistic, you need to understand how football works: how players move, where space emerges and which decisions make sense in a particular situation. These digital environments can provide valuable insights into robot football. Before robots cooperate on a real pitch, many scenarios can already be tested in simulation.
Still, AI does not solve everything. Steven stresses that a smart decision made in software still has to be executed by a physical robot. It has to react in time, maintain its balance and move safely. According to him, the biggest challenge often still lies in the hardware. As robots become physically more capable, the challenge increasingly shifts towards understanding the game and anticipating what is about to happen.
Could robots ever beat humans?
For years, the international robot football competition RoboCup has pursued an ambitious goal. By around 2050, a team of fully autonomous robots should be able to defeat the human world champion. Steven does not rule out that robots will eventually outperform humans. “I don’t know when, but I do know that it is possible.”
Yet if robots become better, another question arises: when is the comparison still fair? A robot might run faster, shoot harder or react more quickly than a human. Should there be limits? And how do you compare a machine with players who become tired, make mistakes or suffer injuries? According to Steven, these very human elements are what make football what it is.
“Emotion, fatigue, unexpected decisions and mistakes are all part of the game. Players can become frustrated when everything goes wrong, or they can suddenly gain confidence when an impossible shot curls perfectly into the top corner. That is what makes football so exciting to watch. Robots do not have those emotions or shifts in momentum - at least not yet.”
If robots outperform humans in every aspect, it may no longer be a contest between humans and machines. Instead, the real competition would shift to the engineers. Who can build the best robot?
What robot football teaches us
Robot football reveals just how complex everyday actions really are. The skills required extend far beyond the football pitch. A robot in a hospital, factory, greenhouse, warehouse or unknown environment does not need to score goals. But it does need to move safely, make independent decisions and deal with unexpected situations.
For Steven, the greatest promise of robotics lies not in replacing people, but in supporting them. “A robot is and remains a tool,” he says. “If robots can take over dangerous, heavy or repetitive tasks, they can make work safer and better.”
Perhaps that is the most important lesson robot football teaches us. The real question is not whether robots will one day become world champions. More interesting is how much intelligence is required for something humans consider completely natural. Giving a pass, accelerating or reacting to an unexpected movement seems simple. Yet every action depends on perception, movement and cooperation. It is precisely this interplay that makes robotics such a fascinating and challenging field.
