Skip to Content
Artificial intelligence

Meet the winner of robotics’ World Cup

Designers of the top adult-size robot in RoboCup’s humanoid soccer league share their work.

If you are under the impression that the recent soccer World Cup was won by a fine French team that triumphed over Croatia in a dramatic finale, then you were watching the wrong tournament.

At about the same time, on the other side of the world in Montreal, another tournament was also under way. This was the soccer World Cup for adult-size robots. And it was won by a robot called NimbRo for the second year in succession. Unsurprisingly, NimbRo is German and represents a nation with a long history of football prowess. 

Today, NimbRo’s masters outline some of the challenges they faced in designing their electronic hero. The original design was for the 2017 RoboCup in Nagoya, the first for adult-size machines. The work suggests that while roboticists have made significant progress, there is huge potential for improvement in the next few years.

Spot the robot: The 2017 World Cup’s winning team

First, some background. RoboCup began in 1997 as a way to test various emerging technologies related to robotics, machine vision, task planning, and so on.

But many readers will recall tournaments in which the players have been less than spectacular. Robot world cups are generally remembered for the clumsy stumbling gait of the players, their comical mix-ups, and the concern of their human masters, often watching through their fingers.

These robots have generally been tiny, and often not even humanoid. That’s mainly because of the difficulty powering and controlling large machines, particularly of the two-legged variety. And in any case, walking, running, and kicking with two legs are just some of many tasks that roboticists need to solve to automate football.

With the rapid advances in artificial intelligence and robotics since then, it’s easy to imagine that machines are set to gain mastery over soccer, just as they have with chess, Go, and any number of Atari video games.

Not quite. In recent years, more powerful and efficient actuators have made it feasible to build adult-size robots that can compete in World Cup–style competitions. Enter NimbRo, a two-legged robot built in less than six months for the 2017 tournament by Grzegorz Ficht and colleagues at the University of Bonn in Germany.  

Adult-size robots must be 130 to 180 centimeters (51 to 71 inches) tall. NimbRo is 135 cm tall and weighs 18 kilograms (almost 40 pounds). The competition itself involves a series of games against single opponents followed by a final in which the player that scores the most goals wins. There is also a set of technical challenges such as kicking a moving ball, jumping, and recovering from a push.

NimbRo has a number of features that make it particularly good at these tasks. In particular, it is simple, so parts can be easily maintained and replaced. “The structure of the robot has been simplified as much as possible to maintain low complexity, but retain functionality that is required when playing soccer,” say Ficht and co.

The exoskeleton consists of just a few simple parts all 3-D-printed from nylon. “By taking advantage of the versatility of 3D printing, we were able to develop an affordable, customisable, highly-capable, adult-sized humanoid robot in little time,” say the team.

Most of the parts are symmetrical so that they can be used in multiple places. This reduces the number of spare parts required.

Vision comes via a single Logitech C905 camera with a 150-degree field of view. The data is processed by an onboard Intel NUC mini PC. This performs object detection, robot localization, task planning, and actuator control.

All this gives the robot a number of soccer-playing abilities. It must be able to locate and approach the ball, avoid obstacles such as other players, and kick and dribble.

NimbRo performs well. In five games, the robot scored 46 goals while conceding only one. That was in the final against a robot called, rather unflatteringly, Sweaty. NimbRo also picked up 21 points in the technical challenges, giving it a clear victory. “At RoboCup 2017 in Nagoya, our robots played very well, winning the AdultSize soccer tournament with high scores,” say the team. Then in June, NimbRo defended its title in Montreal.

That’s interesting work that shows just how far robots have come in performing relatively complex tasks. However, there are plenty of possible improvements.

Reliable balance is one such area. Watch any of the games and you’ll notice that each robot is followed closely by a human whose job is to catch it, should the robot lose its balance. That stops unnecessary damage. But the ability to fall and get up again is a crucial skill in any football game (and in many other situations).

Neither do the robots move particularly quickly or fluently. Their gait can best be described as a hunched shuffle. And their football skills are rudimentary at best. These robots are no match for humans of any ability. (You can watch some highlights from the 2017 tournament here.)

And that’s why rapid improvement should be expected. There are certainly adult-size robots capable of balancing more effectively and walking or running more fluently, as anybody who has watch a Boston Dynamics video will know.

Robots like this would make much more formidable opponents. It’s only a matter of time before they take to the football field.

Ref: : Grown-up NimbRo Robots Winning RoboCup 2017 Humanoid AdultSize Soccer Competitions

Deep Dive

Artificial intelligence

Sam Altman says helpful agents are poised to become AI’s killer function

Open AI’s CEO says we won’t need new hardware or lots more training data to get there.

An AI startup made a hyperrealistic deepfake of me that’s so good it’s scary

Synthesia's new technology is impressive but raises big questions about a world where we increasingly can’t tell what’s real.

Taking AI to the next level in manufacturing

Reducing data, talent, and organizational barriers to achieve scale.

Is robotics about to have its own ChatGPT moment?

Researchers are using generative AI and other techniques to teach robots new skills—including tasks they could perform in homes.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at with a list of newsletters you’d like to receive.