New paper release: Learning Agile Perceptive Traversal of Sparse 3D Structures for Humanoids
Paper https://t.co/LlrDsajHqx
Videos https://t.co/jXLtG68oUg
We study how to do learning and sim2real for 3d traversal, such as monkey bars and overhanging obstacles.
The World Humanoid Robot Games have been getting a lot of attention lately! 🤖🏃🎾
This reminds me of SMPLOlympics, our “Humanoid Olympics” work from 2024: https://t.co/exFl63yNws
We built physically simulated environments and trained life-like RL policies for 10+ sports, including high jump, long jump, hurdles, tennis, table tennis, soccer, basketball, boxing, and more, covering many of the sports we’re now seeing at the Robot Games!
Here’s one fun example: high jump. Interestingly, the running style looks surprisingly similar 👀
Project page: https://t.co/mZNX6Gb7rP
Lab Update: Meet our tiny bipedal robot。XGO-Walker.
Driven by small servos and trained via Reinforcement Learning, Luwu's mini biped is taking its baby steps in the lab today. Small size, huge potential!
Researchers at Light Origins have created LightParkour, a single AI system that lets their Lightbot 0 humanoid robot handle everything from walking to climbing to vaulting.
It works using just a depth camera and a velocity command, deciding on its own which movement fits each obstacle.
The system even adapts to obstacle shapes and variations it has never encountered before.