A small crew of roboticists at Robotic Techniques Lab, ETH Zurich, in Switzerland, has designed, constructed and examined a four-legged robotic able to enjoying badminton with human gamers.
Of their examine, printed within the journal Science Robotics, the group used a reinforcement learning-based controller to present the robotic the flexibility to trace, predict and reply to the motion of a shuttlecock in play, demonstrating the feasibility of utilizing multi-legged robots in dynamic sports activities situations.
Badminton is a sport much like tennis, the primary distinction being using a shuttlecock fairly than a tennis ball. The aim is identical: to hit the shuttlecock over a internet positioned midcourt to an awaiting opponent.
Taking part in badminton requires skillful footwork to maneuver shortly right into a desired place and dexterity of the arms and arms to precisely hit the shuttlecock and ship it over the online to a desired location. Giving a robotic such talents required some diversifications.
The principle adaptation was giving it 4 legs as a substitute of the 2 utilized by people. The configuration provides the robotic far more stability and suppleness in actions.
To provide their robotic such talents, the researchers gave it a stereo digicam and a dynamic arm. In addition they gave it a reinforcement-learning-based controller to permit the robotic to place itself and to reply appropriately to the shuttlecock.
Serving to the robotic was the implementation of what the crew describes as a “notion noise mannequin” to check real-world knowledge from the digicam with knowledge from the knowledge database it constructed throughout coaching. This allowed the robotic to have interaction in human-like badminton participant actions, like following via on strikes and pitching (tilting its base backward or ahead) to maintain the shuttlecock in view.
-
Credit score: 2025 Yuntao Ma, Robotic Techniques Lab, ETH Zurich
-
Badminton robotic. Credit score: 2025 Yuntao Ma, Robotic Techniques Lab, ETH Zurich
Testing confirmed the coaching paid off; the robotic, which the crew has named ANYmal-D, was in a position to transfer across the court docket moderately nicely and preserve rallies with human gamers for as much as 10 photographs.
Extra info:
Yuntao Ma et al, Studying coordinated badminton expertise for legged manipulators, Science Robotics (2025). DOI: 10.1126/scirobotics.adu3922
© 2025 Science X Community
Quotation:
4-legged robotic performs badminton with people (2025, Might 29)
retrieved 29 Might 2025
from https://techxplore.com/information/2025-05-legged-robot-plays-badminton-humans.html
This doc is topic to copyright. Other than any truthful dealing for the aim of personal examine or analysis, no
half could also be reproduced with out the written permission. The content material is supplied for info functions solely.