This Robotic Hand Walks on Its Fingertips

A hand with places to be. We taught a robotic hand to use its fingers as legs and put those same fingers to work interacting with the world. Chapters: 00:00 A hand with places to be 00:13 Untethered locomotion 01:20 Steering 01:32 Fall recovery 01:58 Object pushing 02:11 Keyboard pressing Paper: https://arxiv.org/abs/2609.17172 Project page: https://srl-ethz.github.io/website-fingers-as-legs/ Using reinforcement learning, the hand learns to crawl, steer, recover from falls, press keys and push objects while supporting its own weight. This video brings together untethered locomotion across indoor and outdoor surfaces, fall recovery, keyboard interaction and object-pushing demonstrations. The skills use separate, task-specific policies trained in simulation, with power and policy inference onboard. Keyboard interaction runs without visual feedback, while object pushing uses an overhead camera. Fingers as Legs: Learning Self-Supported Locomotion and Manipulation with an Anthropomorphic Hand Amirhossein Kazemipour · Hehui Zheng · Robert Katzschmann Soft Robotics Lab, ETH Zurich #Robotics #RobotLearning #ReinforcementLearning Буферизовано с YouTube (https://www.youtube.com/watch?v=93BLyBPAfYs) на время плохой работы YouTube.