Visual-Locomotion: Learning to Walk on Complex Terrains with Vision

Wenhao Yu, Deepali Jain, Alejandro Escontrela, Atil Iscen, Peng Xu, Erwin Coumans, Sehoon Ha, Jie Tan, Tingnan Zhang
Proceedings of the 5th Conference on Robot Learning, PMLR 164:1291-1302, 2022.

Abstract

Vision is one of the essential perception modalities for legged robots to safely and efficiently navigate uneven terrains, such as stairs and stepping stones. However, training robots to effectively understand high-dimensional visual input for locomotion is a challenging problem. In this work, we propose a framework to train a vision-based locomotion controller which enables a quadrupedal robot to traverse uneven environments. The key idea is to introduce a hierarchical structure with a high-level vision policy and a low-level motion controller. The high-level vision policy takes as inputs the perceived vision signals as well as robot states and outputs the desired footholds and base movement of the robot. These are then realized by the low level motion controller composed of a position controller for swing legs and a MPC-based torque controller for stance legs. We train the vision policy using Deep Reinforcement Learning and demonstrate our approach on a variety of uneven environments such as randomly placed stepping stones, quincuncial piles, stairs, and moving platforms. We also validate our method on a real robot to walk over a series of gaps and climbing up a platform.

Cite this Paper


BibTeX
@InProceedings{pmlr-v164-yu22a, title = {Visual-Locomotion: Learning to Walk on Complex Terrains with Vision}, author = {Yu, Wenhao and Jain, Deepali and Escontrela, Alejandro and Iscen, Atil and Xu, Peng and Coumans, Erwin and Ha, Sehoon and Tan, Jie and Zhang, Tingnan}, booktitle = {Proceedings of the 5th Conference on Robot Learning}, pages = {1291--1302}, year = {2022}, editor = {Faust, Aleksandra and Hsu, David and Neumann, Gerhard}, volume = {164}, series = {Proceedings of Machine Learning Research}, month = {08--11 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v164/yu22a/yu22a.pdf}, url = {https://proceedings.mlr.press/v164/yu22a.html}, abstract = {Vision is one of the essential perception modalities for legged robots to safely and efficiently navigate uneven terrains, such as stairs and stepping stones. However, training robots to effectively understand high-dimensional visual input for locomotion is a challenging problem. In this work, we propose a framework to train a vision-based locomotion controller which enables a quadrupedal robot to traverse uneven environments. The key idea is to introduce a hierarchical structure with a high-level vision policy and a low-level motion controller. The high-level vision policy takes as inputs the perceived vision signals as well as robot states and outputs the desired footholds and base movement of the robot. These are then realized by the low level motion controller composed of a position controller for swing legs and a MPC-based torque controller for stance legs. We train the vision policy using Deep Reinforcement Learning and demonstrate our approach on a variety of uneven environments such as randomly placed stepping stones, quincuncial piles, stairs, and moving platforms. We also validate our method on a real robot to walk over a series of gaps and climbing up a platform.} }
Endnote
%0 Conference Paper %T Visual-Locomotion: Learning to Walk on Complex Terrains with Vision %A Wenhao Yu %A Deepali Jain %A Alejandro Escontrela %A Atil Iscen %A Peng Xu %A Erwin Coumans %A Sehoon Ha %A Jie Tan %A Tingnan Zhang %B Proceedings of the 5th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2022 %E Aleksandra Faust %E David Hsu %E Gerhard Neumann %F pmlr-v164-yu22a %I PMLR %P 1291--1302 %U https://proceedings.mlr.press/v164/yu22a.html %V 164 %X Vision is one of the essential perception modalities for legged robots to safely and efficiently navigate uneven terrains, such as stairs and stepping stones. However, training robots to effectively understand high-dimensional visual input for locomotion is a challenging problem. In this work, we propose a framework to train a vision-based locomotion controller which enables a quadrupedal robot to traverse uneven environments. The key idea is to introduce a hierarchical structure with a high-level vision policy and a low-level motion controller. The high-level vision policy takes as inputs the perceived vision signals as well as robot states and outputs the desired footholds and base movement of the robot. These are then realized by the low level motion controller composed of a position controller for swing legs and a MPC-based torque controller for stance legs. We train the vision policy using Deep Reinforcement Learning and demonstrate our approach on a variety of uneven environments such as randomly placed stepping stones, quincuncial piles, stairs, and moving platforms. We also validate our method on a real robot to walk over a series of gaps and climbing up a platform.
APA
Yu, W., Jain, D., Escontrela, A., Iscen, A., Xu, P., Coumans, E., Ha, S., Tan, J. & Zhang, T.. (2022). Visual-Locomotion: Learning to Walk on Complex Terrains with Vision. Proceedings of the 5th Conference on Robot Learning, in Proceedings of Machine Learning Research 164:1291-1302 Available from https://proceedings.mlr.press/v164/yu22a.html.

Related Material