Boston Dynamics and Hyundai developed the School of Football program to teach their Atlas robot to play football, using human biomechanics and movement patterns from elite players as training data. The robot first learns movements in simulation before attempting them in the real world, learning timing, force generation, coordination, rotational movement, weight transfer, and full body control. This approach demonstrates that sports like football, which demand balance, coordination, precision, and adaptation simultaneously, serve as effective training grounds for developing natural movement capabilities in humanoid robots, ultimately making them more adaptive and capable of navigating complex real-world environments.
Inmersión profunda
Prerrequisito
- No hay datos disponibles.
Próximos pasos
- No hay datos disponibles.
Inmersión profunda
School of Football | Making Of | Boston Dynamics x HyundaiAñadido:
The biggest challenge in robotics is getting robots to move more naturally.
To move naturally, a robot must learn balance, timing, coordination, and adaptation simultaneously. So, we looked for an activity that demands all of them.
Football, a sport built on balance, coordination, precision, the world's most complete school of movement.
So, we asked ourselves, could football teach a robot to move more naturally?
Hyundai and Boston Dynamics present School of Football, pushing the boundaries of robotics through football.
Together with Hyundai, we've developed a program where we're teaching Atlas true football.
Football is the height of human physical capability.
We see human athletes perform amazing acts of physicality every day. That's a bar that we want to achieve for our robots as well.
First, we drew inspiration from the biomechanics and movement patterns of some of football's greatest players. We translated them into trainable motion data and movement protocols for Atlas.
Atlas then trained these movements in simulation before attempting them in the real world.
>> The best way to prepare for a future robotics is to push Atlas to perform motions that is never done before.
>> From kicking a ball, Atlas learns timing, force generation, and coordination.
From complex kicks, it learns rotational movement, weight transfer, and full body control.
These movement capabilities extend far beyond football, helping Atlas navigate complex real world environments.
>> It was clear early on that we couldn't just give a soccer ball to Atlas and expect it to kick it. It took a lot of work developing a lot of new tools and strategies to really show off what the physical hardware is able to do. A simple task for us like kicking a ball is very very difficult for a humanoid like like us. It has to first figure out where the ball is, move towards the ball, and then perform a kicking motion.
So, how do we combine all that together?
So, we're learning from it.
>> I think this is the first time maybe ever someone's done a raona on a on a robot. Certainly one quite as physical as this one. So, we're very proud.
So, the key steps to developing this move is the first thing is we capture a human performing the action, so kicking.
Then, we transform that human motion into the Atlas kinematics. And then we use AI to train a policy to to mimic the same behavior and then we deploy it on the hardware and actually test it.
>> Play can be really important for a field like robotics both because you know it allows us to push the limits with the robots in ways that we wouldn't if we're just trying to solve useful functional problems.
The things that we've had to build to allow at least to learn soccer are going to become part of the process going forward. Teaching humanoids to play soccer will definitely make them more adaptive and more aware of its environment. Achieving good football will push robotics forwards. At Hyundai, we believe every big step in technology should be anchored in humanity. And what's more human than football?
Videos Relacionados
Beyond Robotics | European Rover Challenge 2026
beyondrobotics
189 views•2026-06-01
Beatbot Sora70: JetPulse Technology and AI obstacle avoidance and navigation!
DroidModderX
26K views•2026-06-02
Tesla FSD 14.3.3 Hits Phoenix Streets - FIRST LOOK
anthonystesla
114 views•2026-05-29
NVIDIA Just Unveiled a 75 DOF Humanoid Robot
DPCcars
74K views•2026-06-01
Elon Musk Just Revealed Fremont Line for Optimus Gen 3 Mass Production
TheAINexusOfficial
180 views•2026-05-30
人機一体「零式人機 ver.2」 子ども企画【おもしろ発見!モビリティー】 #乗り物 #automobile #robot #shorts
KyodoNews
1K views•2026-05-28
China’s New Luna AI Robot Looks Shockingly Human...
NextGenHumanoids
850 views•2026-05-28
Reachy Mini: the $300 open source robot you can actually hack — Andres Marafioti, Hugging Face
aiDotEngineer
662 views•2026-05-29











