Four-Legged Robot Masters High-Speed Parkour with Autonomous Navigation
A team of researchers at the Robotics & Artificial Intelligence Lab in Korea has developed a four-legged robot capable of performing high-speed parkour using autonomous movement planning.
The robot, named Raibo, was designed with a controller that allows it to plan and track its own movements, enabling it to navigate complex environments without human input.
Parkour, an athletic discipline involving obstacle navigation in unpredictable urban settings, requires agility and precision. To meet these demands, the team equipped Raibo with four legs for enhanced stability and mobility.
The robot’s controller includes a planner and a tracker. The planner, powered by a neural network, generates and updates a map of the terrain, helping Raibo determine safe and efficient foot placements.
This system integrates data from cameras and feedback sensors, allowing Raibo to adjust its movements in real time while maintaining its course.
Initial testing began with simulations to verify the robot’s components. Raibo was then tested in various lab environments to evaluate its mapping and navigation capabilities.
The robot demonstrated the ability to run vertically on walls for short distances, jump over 1.3-metre gaps, and traverse uneven terrain such as stone-covered courses.
Raibo also successfully navigated ramps, stairs, and boxes, stepping up and down with precision.
The team has already begun work on an upgraded version, Raibo 2, aiming to enhance its performance and ensure safe operation.
Researchers suggest the robot could be deployed in disaster zones or other hazardous environments in the future.
Korean researchers developed a four-legged robot named Raibo for high-speed parkour
Raibo uses a neural network-based planner and real-time tracking for autonomous navigation
It can jump 1.3 metres, run on walls, and traverse uneven terrain
Source: TechXplore
