A team of roboticists and AI specialists at the Robotics & Artificial Intelligence Lab in Korea has designed, built and successfully tested a four-legged robot that is capable of conducting high-speed parkour maneuvers. In their paper published in the journal Science Robotics, the group describes how they gave their robot a controller capable of both planning and tracking its own movements to allow it to freely traverse a range of environments.
Parkour is an obstacle course type athletic discipline that takes place in unpredictable, real-world, generally urban environments—it involves climbing walls, jumping between buildings, maneuvering around objects and running across difficult, uneven terrain. The objective is to get from one place to another without injury. To give their robot the ability to conduct parkour maneuvers, the team made one change right away—they gave it four legs.
The next thing they did was design and build a special kind of controller, one that was capable of planning the route to be taken and a tracker that told the robot where to place its feet and how to use its body to move forward safely.
The planner was designed using a neural network that was first trained and then used to generate and continuously update a map. Such a map, the team notes, was used to not only plan a route through a given terrain but to figure out where to put the robot’s feet in a way that would keep it to the desired route in a safe way. That data was used by the planner in conjunction with data from the camera and feedback sensors.
Testing of the robot, which the team called Raibo, involved first running simulations to ensure the robot’s components were working as intended. The next phase involved putting the robot in various lab environments to see not only how well it fared but also to study its mapping and course-plotting abilities.
They found their robot was capable of running vertically on a wall for short distances, jumping over 1.3-meter spans, and running across a course covered in stones. They also found it could walk up ramps and stairs, step up on boxes and step down on the other side.
The team plans to continue their work with the robot. They have already built and started testing Raibo 2, looking for ways to improve its abilities and ensure it behaves safely. They suggest that it could be used in disaster zones or other challenging environments.
More information:
Hyeongjun Kim et al, High-speed control and navigation for quadrupedal robots on complex and discrete terrain, Science Robotics (2025). DOI: 10.1126/scirobotics.ads6192
© 2025 Science X Network
Citation:
Robot navigates high-speed parkour with autonomous movement planning (2025, May 30)
retrieved 30 May 2025
from https://techxplore.com/news/2025-05-robot-high-parkour-autonomous-movement.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no
part may be reproduced without the written permission. The content is provided for information purposes only.
