ClutterGen: A Cluttered Scene Generator for Robot Learning

Yinsen Jia, Boyuan Chen
Proceedings of The 8th Conference on Robot Learning, PMLR 270:1782-1796, 2025.

Abstract

We introduce ClutterGen, a physically compliant simulation scene generator capable of producing highly diverse, cluttered, and stable scenes for robot learning. Generating such scenes is challenging as each object must adhere to physical laws like gravity and collision. As the number of objects increases, finding valid poses becomes more difficult, necessitating significant human engineering effort, which limits the diversity of the scenes. To overcome these challenges, we propose a reinforcement learning method that can be trained with physics-based reward signals provided by the simulator. Our experiments demonstrate that ClutterGen can generate cluttered object layouts with up to ten objects on confined table surfaces. Additionally, our policy design explicitly encourages the diversity of the generated scenes for open-ended generation. Our real-world robot results show that ClutterGen can be directly used for clutter rearrangement and stable placement policy training.

Cite this Paper


BibTeX
@InProceedings{pmlr-v270-jia25a, title = {ClutterGen: A Cluttered Scene Generator for Robot Learning}, author = {Jia, Yinsen and Chen, Boyuan}, booktitle = {Proceedings of The 8th Conference on Robot Learning}, pages = {1782--1796}, year = {2025}, editor = {Agrawal, Pulkit and Kroemer, Oliver and Burgard, Wolfram}, volume = {270}, series = {Proceedings of Machine Learning Research}, month = {06--09 Nov}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v270/main/assets/jia25a/jia25a.pdf}, url = {https://proceedings.mlr.press/v270/jia25a.html}, abstract = {We introduce ClutterGen, a physically compliant simulation scene generator capable of producing highly diverse, cluttered, and stable scenes for robot learning. Generating such scenes is challenging as each object must adhere to physical laws like gravity and collision. As the number of objects increases, finding valid poses becomes more difficult, necessitating significant human engineering effort, which limits the diversity of the scenes. To overcome these challenges, we propose a reinforcement learning method that can be trained with physics-based reward signals provided by the simulator. Our experiments demonstrate that ClutterGen can generate cluttered object layouts with up to ten objects on confined table surfaces. Additionally, our policy design explicitly encourages the diversity of the generated scenes for open-ended generation. Our real-world robot results show that ClutterGen can be directly used for clutter rearrangement and stable placement policy training.} }
Endnote
%0 Conference Paper %T ClutterGen: A Cluttered Scene Generator for Robot Learning %A Yinsen Jia %A Boyuan Chen %B Proceedings of The 8th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2025 %E Pulkit Agrawal %E Oliver Kroemer %E Wolfram Burgard %F pmlr-v270-jia25a %I PMLR %P 1782--1796 %U https://proceedings.mlr.press/v270/jia25a.html %V 270 %X We introduce ClutterGen, a physically compliant simulation scene generator capable of producing highly diverse, cluttered, and stable scenes for robot learning. Generating such scenes is challenging as each object must adhere to physical laws like gravity and collision. As the number of objects increases, finding valid poses becomes more difficult, necessitating significant human engineering effort, which limits the diversity of the scenes. To overcome these challenges, we propose a reinforcement learning method that can be trained with physics-based reward signals provided by the simulator. Our experiments demonstrate that ClutterGen can generate cluttered object layouts with up to ten objects on confined table surfaces. Additionally, our policy design explicitly encourages the diversity of the generated scenes for open-ended generation. Our real-world robot results show that ClutterGen can be directly used for clutter rearrangement and stable placement policy training.
APA
Jia, Y. & Chen, B.. (2025). ClutterGen: A Cluttered Scene Generator for Robot Learning. Proceedings of The 8th Conference on Robot Learning, in Proceedings of Machine Learning Research 270:1782-1796 Available from https://proceedings.mlr.press/v270/jia25a.html.

Related Material