Make a Donut: Hierarchical EMD-Space Planning for Zero-Shot Deformable Manipulation with Tools

Published in RA-L,IROS, 2025

Yang You, Bokui Shen, Congyue Deng, Haoran Geng, Songlin Wei, He Wang, Leonidas Guibas

arXiv / PDF / project page / BibTeX

Abstract

Deformable object manipulation is a challenging area in robotics, often relying on demonstrations to learn task dynamics. However, obtaining suitable demonstrations for long-horizon tasks is difficult and can limit model generalization. We propose a demonstration-free hierarchical planning approach for complex long-horizon tasks without training. Using large language models (LLMs), we create a high-level, stage-by-stage plan for a task, specifying tools and generating Python code for intermediate subgoal point clouds. With these, we employ a closed-loop model predictive control strategy using Differentiable Physics with Point-to-Point correspondence (DiffPhysics-P2P) loss in the earth mover distance (EMD) space. Our method outperforms benchmarks in dough manipulation tasks and generalizes well to novel tasks without demonstrations, validated through real-world robotic experiments.

BibTeX

@article{you2025make,
  title     = {Make a Donut: Hierarchical EMD-Space Planning for Zero-Shot Deformable Manipulation with Tools},
  author    = {Yang You and Bokui Shen and Congyue Deng and Haoran Geng and Songlin Wei and He Wang and Leonidas Guibas},
  journal   = {IEEE Robotics and Automation Letters (RA-L)},
  year      = {2025},
  eprint    = {2311.02787},
  archivePrefix = {arXiv}
}

Recommended citation: You, Y., Shen, B., Deng, C., Geng, H., Wei, S., Wang, H., & Guibas, L. (2025). Make a Donut: Hierarchical EMD-Space Planning for Zero-Shot Deformable Manipulation with Tools. IEEE Robotics and Automation Letters, 10(4), 3270-3277.