A group of researchers has taken robotics to an entirely new level by teaching robots how to skateboard.
RELATED: P-Rod Stands Strong on Calling New Gen Skaters Like Robots w/ No Style
According to reports, Scientists at the University of Michigan's Computational Autonomy and Robotics Laboratory, along with Southern University of Science and Technology, have been working to create legged robots that can handle the complexities of real-world movement, and apparently, that includes rolling down the street on a skateboard.
At first glance, this might seem like just a fun experiment, but there’s some serious science behind it.
The goal is to develop machines that can switch between different movement styles seamlessly.
Walking is one thing, but skateboarding requires an entirely different skill set - balancing on a moving platform, shifting weight, and reacting to unpredictable surfaces.
All of these factors make it a perfect testing ground for more advanced robotic mobility.
Why Skateboarding?
Most robots are built to handle either smooth, predictable environments or rough terrain where they can carefully plan each step.
Skateboarding throws all of that out the window. A robot on a skateboard has to make constant, real-time adjustments while maintaining balance, similar to how a human skater constantly shifts their weight.
This requires a combination of precise physics calculations, quick reflexes, and real-world adaptability.
By tackling something as dynamic as skateboarding, these researchers are figuring out how to make robots that can handle chaotic environments without falling flat on their faces—or, in this case, their sensors.
The ability to ride a skateboard could translate to more practical uses, like helping robots move efficiently through urban environments or rough terrain without relying solely on walking.
How Does It Work?
The key challenge here is that traditional robotics control systems aren’t built for activities like skateboarding. The scientists had to create a model that allowed the robot to predict and respond to the shifting forces under its wheels.
They combined machine learning with physics-based modeling, teaching the robot how to balance and move forward, much like how a beginner skater learns to push off and stay upright.
The robot was trained using a combination of simulations and real-world testing. First, it practiced in a virtual environment where it could “learn” the mechanics without risking any expensive falls.
Once it showed progress in simulations, researchers let it loose in the real world. And yes, just like a human learning to skate, there were plenty of wipeouts.
But over time, it started getting better at adjusting its stance, shifting its weight, and even making minor turns.
What’s Next?
With the basics of skateboarding down, the next step is figuring out more advanced maneuvers. Can a robot learn to carve? Drop in?
Maybe even kickflip one day? While those questions remain unanswered for now, this research could have much broader applications.
Balancing on a moving platform while adjusting to rapid changes isn’t just useful for skateboarding - it could be essential for making robots that can navigate disaster zones, work in unpredictable environments, or even assist in everyday tasks where surfaces aren’t always stable.
For now, though, the sight of a robot rolling through a skatepark is enough to turn some heads.
The idea that one day a robot might be able to hit the streets with a crew of skaters isn’t as far-fetched as it once seemed. Who knows - maybe the future of skating includes a few metal-footed shredders rolling alongside us.