Robot vs Human – ETH Zurich’s ANYmal Learns to Play Badminton with AI Precision!

At ETH Zurich, researchers have done something both impressive and slightly unexpected—they’ve taught a four-legged robot named ANYmal-D to play badminton. Yes, really.

Using reinforcement learning, the robot was trained to track a shuttlecock in flight, predict its trajectory, and return it with a robotic arm and racket. It’s equipped with a stereo camera for vision, and a highly mobile arm that can swing in almost any direction—basically, it’s got better hand-eye coordination than most of us on a Monday.

Badminton is no easy task—it demands quick reflexes, precise movements, and good footwork. So the team developed a clever control system to coordinate the robot’s legs and strokes, using a perception noise model to help it interpret real-time data, even in a noisy environment.

The result? A robot that can tilt, twist, and return shots like a mechanical pro. In tests, ANYmal-D successfully rallied with humans—managing up to ten consecutive strokes.

So while it might not be ready for Wimbledon just yet, it’s safe to say: the future of sport might be more robotic than we think.

And that’s your robot news update for today!
If you’re curious about how robotics can transform your business, join me for the weekly Live Robot Optimise Workshop — it’s packed with insights, trends, and practical tips. Don’t forget to subscribe so you stay in the loop with all the latest updates.

I’m RoboPhil from Robot Philosophy — thanks for watching, and I’ll see you next time!

Join our Robot Optimise Industry (ROI) Workshop: https://robophil.com/workshop

Sponsors:-

Robot Center: – https://robotcenter.co.uk/ – Buy Robot, Robot Buy, Robot consultancy, Robotics Consultancy

Robots of London: – https://robotsoflondon.co.uk/ – Robot Hire, Robot Rental, Rent Robot, Hire Robot, Robot Events

Robot Philosophy: – https://robophil.com/ – Robot Consultancy AND Robot Recruitment, Robot Advice, Robot Insights, Robot Ideas