Researchers from Eth Zurich pushed Anymal, a quadruped robot, into a new territory: badminton court.
A new video released by the university shows a four-legged robot trying to return a shot using a racket, an advanced vision system, and a reinforced learning-based control. (See the video below.)
Built by Anybotics, Anymal was developed as a versatile quadruped animal platform and uses dual cameras to track Shuttlecocks in flight. The system looks for the shuttle color, calculates its trajectory, and instructs the robot's motor to move it to position and perform a return shot.
“Robot control is based on reinforcement learning,” the researchers explain in the video. “It means taking the robot and throwing it into the simulator, where you can do thousands of trial and error iterations and slowly learn the best strategy.”
The project highlights the difficulty of teaching robots to play fast-paced sports. Badminton requires rapid response, accurate limb adjustment, and fast dynamic control. This is all the challenges of today's robotic systems.
Researchers compare the robot's abilities to the ability of a 7-year-old child playing with his parents.
While robots are far from competing with human players, the movement is more than new. Eth Zurich says the lessons learned from controlling Ansymal under such demanding conditions are returning to broader robotics research, improving movement, coordination and system design.
For now, it's just a “friendly match”, but the team is working towards a future where robots could become serious opponents on the court.
This study is published in Science robot.
https://www.youtube.com/watch?v=2hq-rwsowdo
