Motion Policy Networks

Adam Fishman, Adithyavairavan Murali, Clemens Eppner, Bryan Peele, Byron Boots, Dieter Fox
Proceedings of The 6th Conference on Robot Learning, PMLR 205:967-977, 2023.

Abstract

Collision-free motion generation in unknown environments is a core building block for robot manipulation. Generating such motions is challenging due to multiple objectives; not only should the solutions be optimal, the motion generator itself must be fast enough for real-time performance and reliable enough for practical deployment. A wide variety of methods have been proposed ranging from local controllers to global planners, often being combined to offset their shortcomings. We present an end-to-end neural model called Motion Policy Networks (M$\pi$Nets) to generate collision-free, smooth motion from just a single depth camera observation. M$\pi$Nets are trained on over 3 million motion planning problems in more than 500,000 environments. Our experiments show that M$\pi$Nets are significantly faster than global planners while exhibiting the reactivity needed to deal with dynamic scenes. They are 46% better than prior neural planners and more robust than local control policies. Despite being only trained in simulation, M$\pi$Nets transfer well to the real robot with noisy partial point clouds. Videos and code are available at https://mpinets.github.io

Cite this Paper

BibTeX
@InProceedings{pmlr-v205-fishman23a, title = {Motion Policy Networks}, author = {Fishman, Adam and Murali, Adithyavairavan and Eppner, Clemens and Peele, Bryan and Boots, Byron and Fox, Dieter}, booktitle = {Proceedings of The 6th Conference on Robot Learning}, pages = {967--977}, year = {2023}, editor = {Liu, Karen and Kulic, Dana and Ichnowski, Jeff}, volume = {205}, series = {Proceedings of Machine Learning Research}, month = {14--18 Dec}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v205/fishman23a/fishman23a.pdf}, url = {https://proceedings.mlr.press/v205/fishman23a.html}, abstract = {Collision-free motion generation in unknown environments is a core building block for robot manipulation. Generating such motions is challenging due to multiple objectives; not only should the solutions be optimal, the motion generator itself must be fast enough for real-time performance and reliable enough for practical deployment. A wide variety of methods have been proposed ranging from local controllers to global planners, often being combined to offset their shortcomings. We present an end-to-end neural model called Motion Policy Networks (M$\pi$Nets) to generate collision-free, smooth motion from just a single depth camera observation. M$\pi$Nets are trained on over 3 million motion planning problems in more than 500,000 environments. Our experiments show that M$\pi$Nets are significantly faster than global planners while exhibiting the reactivity needed to deal with dynamic scenes. They are 46% better than prior neural planners and more robust than local control policies. Despite being only trained in simulation, M$\pi$Nets transfer well to the real robot with noisy partial point clouds. Videos and code are available at https://mpinets.github.io} }
Endnote
%0 Conference Paper %T Motion Policy Networks %A Adam Fishman %A Adithyavairavan Murali %A Clemens Eppner %A Bryan Peele %A Byron Boots %A Dieter Fox %B Proceedings of The 6th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2023 %E Karen Liu %E Dana Kulic %E Jeff Ichnowski %F pmlr-v205-fishman23a %I PMLR %P 967--977 %U https://proceedings.mlr.press/v205/fishman23a.html %V 205 %X Collision-free motion generation in unknown environments is a core building block for robot manipulation. Generating such motions is challenging due to multiple objectives; not only should the solutions be optimal, the motion generator itself must be fast enough for real-time performance and reliable enough for practical deployment. A wide variety of methods have been proposed ranging from local controllers to global planners, often being combined to offset their shortcomings. We present an end-to-end neural model called Motion Policy Networks (M$\pi$Nets) to generate collision-free, smooth motion from just a single depth camera observation. M$\pi$Nets are trained on over 3 million motion planning problems in more than 500,000 environments. Our experiments show that M$\pi$Nets are significantly faster than global planners while exhibiting the reactivity needed to deal with dynamic scenes. They are 46% better than prior neural planners and more robust than local control policies. Despite being only trained in simulation, M$\pi$Nets transfer well to the real robot with noisy partial point clouds. Videos and code are available at https://mpinets.github.io
APA
Fishman, A., Murali, A., Eppner, C., Peele, B., Boots, B. & Fox, D.. (2023). Motion Policy Networks. Proceedings of The 6th Conference on Robot Learning, in Proceedings of Machine Learning Research 205:967-977 Available from https://proceedings.mlr.press/v205/fishman23a.html.

Related Material