Teaching robots how to walk across a narrow path without losing balance is easier said than done. Researchers have come up with an autonomous path planning approach for Atlas humanoid robot that enables it to sense the terrain with LiDAR and build a map of planar regions.
The path planning algorithm is used to allow the robot to plan its footsteps with its destination in mind. According to the researchers, the robot is about 50% successful over this type of terrain.
[HT]