Jacta: A Versatile Planner for Learning Dexterous and Whole-body Manipulation

Jan Bruedigam, Ali Adeeb Abbas, Maks Sorokin, Kuan Fang, Brandon Hung, Maya Guru, Stefan Georg Sosnowski, Jiuguang Wang, Sandra Hirche, Simon Le Cleac’h
Proceedings of The 8th Conference on Robot Learning, PMLR 270:994-1020, 2025.

Abstract

Robotic manipulation is challenging due to discontinuous dynamics, as well as high-dimensional state and action spaces. Data-driven approaches that succeed in manipulation tasks require large amounts of data and expert demonstrations, typically from humans. Existing planners are restricted to specific systems and often depend on specialized algorithms for using demonstrations. Therefore, we introduce a flexible motion planner tailored to dexterous and whole-body manipulation tasks. Our planner creates readily usable demonstrations for reinforcement learning algorithms, eliminating the need for additional training pipeline complexities. With this approach, we can efficiently learn policies for complex manipulation tasks, where traditional reinforcement learning alone only makes little progress. Furthermore, we demonstrate that learned policies are transferable to real robotic systems for solving complex dexterous manipulation tasks. Project website: https://jacta-manipulation.github.io/

Cite this Paper

BibTeX
@InProceedings{pmlr-v270-bruedigam25a, title = {Jacta: A Versatile Planner for Learning Dexterous and Whole-body Manipulation}, author = {Bruedigam, Jan and Abbas, Ali Adeeb and Sorokin, Maks and Fang, Kuan and Hung, Brandon and Guru, Maya and Sosnowski, Stefan Georg and Wang, Jiuguang and Hirche, Sandra and Cleac'h, Simon Le}, booktitle = {Proceedings of The 8th Conference on Robot Learning}, pages = {994--1020}, 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/bruedigam25a/bruedigam25a.pdf}, url = {https://proceedings.mlr.press/v270/bruedigam25a.html}, abstract = {Robotic manipulation is challenging due to discontinuous dynamics, as well as high-dimensional state and action spaces. Data-driven approaches that succeed in manipulation tasks require large amounts of data and expert demonstrations, typically from humans. Existing planners are restricted to specific systems and often depend on specialized algorithms for using demonstrations. Therefore, we introduce a flexible motion planner tailored to dexterous and whole-body manipulation tasks. Our planner creates readily usable demonstrations for reinforcement learning algorithms, eliminating the need for additional training pipeline complexities. With this approach, we can efficiently learn policies for complex manipulation tasks, where traditional reinforcement learning alone only makes little progress. Furthermore, we demonstrate that learned policies are transferable to real robotic systems for solving complex dexterous manipulation tasks. Project website: https://jacta-manipulation.github.io/} }
Endnote
%0 Conference Paper %T Jacta: A Versatile Planner for Learning Dexterous and Whole-body Manipulation %A Jan Bruedigam %A Ali Adeeb Abbas %A Maks Sorokin %A Kuan Fang %A Brandon Hung %A Maya Guru %A Stefan Georg Sosnowski %A Jiuguang Wang %A Sandra Hirche %A Simon Le Cleac’h %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-bruedigam25a %I PMLR %P 994--1020 %U https://proceedings.mlr.press/v270/bruedigam25a.html %V 270 %X Robotic manipulation is challenging due to discontinuous dynamics, as well as high-dimensional state and action spaces. Data-driven approaches that succeed in manipulation tasks require large amounts of data and expert demonstrations, typically from humans. Existing planners are restricted to specific systems and often depend on specialized algorithms for using demonstrations. Therefore, we introduce a flexible motion planner tailored to dexterous and whole-body manipulation tasks. Our planner creates readily usable demonstrations for reinforcement learning algorithms, eliminating the need for additional training pipeline complexities. With this approach, we can efficiently learn policies for complex manipulation tasks, where traditional reinforcement learning alone only makes little progress. Furthermore, we demonstrate that learned policies are transferable to real robotic systems for solving complex dexterous manipulation tasks. Project website: https://jacta-manipulation.github.io/
APA
Bruedigam, J., Abbas, A.A., Sorokin, M., Fang, K., Hung, B., Guru, M., Sosnowski, S.G., Wang, J., Hirche, S. & Cleac’h, S.L.. (2025). Jacta: A Versatile Planner for Learning Dexterous and Whole-body Manipulation. Proceedings of The 8th Conference on Robot Learning, in Proceedings of Machine Learning Research 270:994-1020 Available from https://proceedings.mlr.press/v270/bruedigam25a.html.

Related Material