We show that deep reinforcement learning can be used to learn policies for legged locomotion control tasks encountered in space exploration. such as three-dimensional re-orientation and landing of a quadruped robot exploring low-gravity celestial bodies. Using sim-to-real transfer, we deploy trained policies in the real world on the SpaceBok robot placed on an experimental testbed designed for two-dimensional micro-gravity experiments.
Title: Cat-like Jumping and Landing of Legged Robots in Low-gravity Using Deep Reinforcement Learning
Authors: Nikita Rudin, Hendrik Kolvenbach, Vassilios Tsounis and Marco Hutter
IEEE Transactions on Robotics (Early Access):
ieeexplore.ieee.org/document/...
DOI: 10.1109/TRO.2021.3084374
Also available here:
www.research-collection.ethz....
arxiv.org/abs/2106.09357
This work was supported by the European Space Agency (ESA) and Airbus DS in the framework of the Network Partnering Initiative 481-2016.
For more information visit:
rsl.ethz.ch
15 июн 2021