Gentle Manipulation of Tree Branches: A Contact-Aware Policy Learning Approach

Jay Jacob, Shizhe Cai, Paulo Vinicius Koerich Borges, Tirthankar Bandyopadhyay, Fabio Ramos
Proceedings of The 8th Conference on Robot Learning, PMLR 270:631-648, 2025.

Abstract

Learning to interact with deformable tree branches with minimal damage is challenging due to their intricate geometry and inscrutable dynamics. Furthermore, traditional vision-based modelling systems suffer from implicit occlusions in dense foliage, severely changing lighting conditions, and limited field of view, in addition to having a significant computational burden preventing real-time deployment.In this work, we simulate a procedural forest with realistic, self-similar branching structures derived from a parametric L-system model, actuated with crude spring abstractions, mirroring real-world variations with domain randomisation over the morphological and dynamic attributes. We then train a novel Proprioceptive Contact-Aware Policy (PCAP) for a reach task using reinforcement learning, aided by a whole-arm contact detection classifier and reward engineering, without external vision, tactile, or torque sensing. The agent deploys novel strategies to evade and mitigate contact impact, favouring a reactive exploration of the task space. Finally, we demonstrate that the learned behavioural patterns can be transferred zero-shot from simulation to real, allowing the arm to navigate around real branches with unseen topology and variable occlusions while minimising the contact forces and expected ruptures.

Cite this Paper

BibTeX
@InProceedings{pmlr-v270-jacob25a, title = {Gentle Manipulation of Tree Branches: A Contact-Aware Policy Learning Approach}, author = {Jacob, Jay and Cai, Shizhe and Borges, Paulo Vinicius Koerich and Bandyopadhyay, Tirthankar and Ramos, Fabio}, booktitle = {Proceedings of The 8th Conference on Robot Learning}, pages = {631--648}, 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/jacob25a/jacob25a.pdf}, url = {https://proceedings.mlr.press/v270/jacob25a.html}, abstract = {Learning to interact with deformable tree branches with minimal damage is challenging due to their intricate geometry and inscrutable dynamics. Furthermore, traditional vision-based modelling systems suffer from implicit occlusions in dense foliage, severely changing lighting conditions, and limited field of view, in addition to having a significant computational burden preventing real-time deployment.In this work, we simulate a procedural forest with realistic, self-similar branching structures derived from a parametric L-system model, actuated with crude spring abstractions, mirroring real-world variations with domain randomisation over the morphological and dynamic attributes. We then train a novel Proprioceptive Contact-Aware Policy (PCAP) for a reach task using reinforcement learning, aided by a whole-arm contact detection classifier and reward engineering, without external vision, tactile, or torque sensing. The agent deploys novel strategies to evade and mitigate contact impact, favouring a reactive exploration of the task space. Finally, we demonstrate that the learned behavioural patterns can be transferred zero-shot from simulation to real, allowing the arm to navigate around real branches with unseen topology and variable occlusions while minimising the contact forces and expected ruptures.} }
Endnote
%0 Conference Paper %T Gentle Manipulation of Tree Branches: A Contact-Aware Policy Learning Approach %A Jay Jacob %A Shizhe Cai %A Paulo Vinicius Koerich Borges %A Tirthankar Bandyopadhyay %A Fabio Ramos %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-jacob25a %I PMLR %P 631--648 %U https://proceedings.mlr.press/v270/jacob25a.html %V 270 %X Learning to interact with deformable tree branches with minimal damage is challenging due to their intricate geometry and inscrutable dynamics. Furthermore, traditional vision-based modelling systems suffer from implicit occlusions in dense foliage, severely changing lighting conditions, and limited field of view, in addition to having a significant computational burden preventing real-time deployment.In this work, we simulate a procedural forest with realistic, self-similar branching structures derived from a parametric L-system model, actuated with crude spring abstractions, mirroring real-world variations with domain randomisation over the morphological and dynamic attributes. We then train a novel Proprioceptive Contact-Aware Policy (PCAP) for a reach task using reinforcement learning, aided by a whole-arm contact detection classifier and reward engineering, without external vision, tactile, or torque sensing. The agent deploys novel strategies to evade and mitigate contact impact, favouring a reactive exploration of the task space. Finally, we demonstrate that the learned behavioural patterns can be transferred zero-shot from simulation to real, allowing the arm to navigate around real branches with unseen topology and variable occlusions while minimising the contact forces and expected ruptures.
APA
Jacob, J., Cai, S., Borges, P.V.K., Bandyopadhyay, T. & Ramos, F.. (2025). Gentle Manipulation of Tree Branches: A Contact-Aware Policy Learning Approach. Proceedings of The 8th Conference on Robot Learning, in Proceedings of Machine Learning Research 270:631-648 Available from https://proceedings.mlr.press/v270/jacob25a.html.

Related Material