Learning a Contact-Adaptive Controller for Robust, Efficient Legged Locomotion

Xingye Da, Zhaoming Xie, David Hoeller, Byron Boots, Anima Anandkumar, Yuke Zhu, Buck Babich, Animesh Garg
Proceedings of the 2020 Conference on Robot Learning, PMLR 155:883-894, 2021.

Abstract

We present a hierarchical framework that combines model-based control and reinforcement learning (RL) to synthesize robust controllers for a quadruped (the Unitree Laikago). The system consists of a high-level controller that learns to choose from a set of primitives in response to changes in the environment and a low-level controller that utilizes an established control method to robustly execute the primitives. Our framework learns a controller that can adapt to challenging environmental changes on the fly, including novel scenarios not seen during training. The learned controller is up to 85 percent more energy efficient and is more robust compared to baseline methods. We also deploy the controller on a physical robot without any randomization or adaptation scheme.

Cite this Paper

BibTeX
@InProceedings{pmlr-v155-da21a, title = {Learning a Contact-Adaptive Controller for Robust, Efficient Legged Locomotion}, author = {Da, Xingye and Xie, Zhaoming and Hoeller, David and Boots, Byron and Anandkumar, Anima and Zhu, Yuke and Babich, Buck and Garg, Animesh}, booktitle = {Proceedings of the 2020 Conference on Robot Learning}, pages = {883--894}, year = {2021}, editor = {Kober, Jens and Ramos, Fabio and Tomlin, Claire}, volume = {155}, series = {Proceedings of Machine Learning Research}, month = {16--18 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v155/da21a/da21a.pdf}, url = {https://proceedings.mlr.press/v155/da21a.html}, abstract = {We present a hierarchical framework that combines model-based control and reinforcement learning (RL) to synthesize robust controllers for a quadruped (the Unitree Laikago). The system consists of a high-level controller that learns to choose from a set of primitives in response to changes in the environment and a low-level controller that utilizes an established control method to robustly execute the primitives. Our framework learns a controller that can adapt to challenging environmental changes on the fly, including novel scenarios not seen during training. The learned controller is up to 85 percent more energy efficient and is more robust compared to baseline methods. We also deploy the controller on a physical robot without any randomization or adaptation scheme.} }
Endnote
%0 Conference Paper %T Learning a Contact-Adaptive Controller for Robust, Efficient Legged Locomotion %A Xingye Da %A Zhaoming Xie %A David Hoeller %A Byron Boots %A Anima Anandkumar %A Yuke Zhu %A Buck Babich %A Animesh Garg %B Proceedings of the 2020 Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2021 %E Jens Kober %E Fabio Ramos %E Claire Tomlin %F pmlr-v155-da21a %I PMLR %P 883--894 %U https://proceedings.mlr.press/v155/da21a.html %V 155 %X We present a hierarchical framework that combines model-based control and reinforcement learning (RL) to synthesize robust controllers for a quadruped (the Unitree Laikago). The system consists of a high-level controller that learns to choose from a set of primitives in response to changes in the environment and a low-level controller that utilizes an established control method to robustly execute the primitives. Our framework learns a controller that can adapt to challenging environmental changes on the fly, including novel scenarios not seen during training. The learned controller is up to 85 percent more energy efficient and is more robust compared to baseline methods. We also deploy the controller on a physical robot without any randomization or adaptation scheme.
APA
Da, X., Xie, Z., Hoeller, D., Boots, B., Anandkumar, A., Zhu, Y., Babich, B. & Garg, A.. (2021). Learning a Contact-Adaptive Controller for Robust, Efficient Legged Locomotion. Proceedings of the 2020 Conference on Robot Learning, in Proceedings of Machine Learning Research 155:883-894 Available from https://proceedings.mlr.press/v155/da21a.html.

Related Material