M${}^{3}$Bench: Benchmarking Whole-Body <u>M</u>otion Generation for <u>M</u>obile <u>M</u>anipulation in 3D Scenes
We propose M <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${}^{3}$</tex-math></inline-formula> Bench, a new benchmark for whole-body motion generation in mobile manipulation tasks. Given a 3D scene context, M <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${}^{3}$</tex-math></inline-formula> Bench requires an embodied agent to reason about its configuration, environmental constraints, and task objectives to generate coordinated whole-body motion trajectories for object rearrangement. M <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${}^{3}$</tex-math></inline-formula> Bench features 30,000 object rearrangement tasks across 119 diverse scenes, providing expert demonstrations generated by our newly developed M <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${}^{3}$</tex-math></inline-formula> BenchMaker, an automatic data generation tool that produces whole-body motion trajectories from high-level task instructions using only basic scene and robot information. Our benchmark includes various task splits to evaluate generalization across different dimensions and leverages realistic physics simulation for trajectory assessment. Extensive evaluation analysis reveals that state-of-the-art models struggle with coordinating base-arm motion while adhering to environmental and task-specific constraints, underscoring the need for new models to bridge this gap. By releasing M <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${}^{3}$</tex-math></inline-formula> Bench and M <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${}^{3}$</tex-math></inline-formula> BenchMaker at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://zeyuzhang.com/papers/m3bench</uri> , we aim to advance robotics research toward more adaptive and capable mobile manipulation in diverse, real-world environments.
Authors
- Song‐Chun Zhu (ORCID: https://orcid.org/0009-0008-0876-8117)
- Xinggang Wang (ORCID: https://orcid.org/0000-0001-6732-7823)
- Hangxin Liu (ORCID: https://orcid.org/0000-0002-3003-8611)
- Zeyu Zhang (ORCID: https://orcid.org/0000-0001-8929-134X)
- Muzhi Han (ORCID: https://orcid.org/0000-0002-2649-4577)
- Sixu Yan
- Zaijin Wang (ORCID: https://orcid.org/0009-0007-7925-356X)
Institutions
- University of California, Los Angeles (US)
- Beijing Academy of Artificial Intelligence (CN)
- Beijing Institute for General Artificial Intelligence (CN)
- Huazhong University of Science and Technology (CN)
Publication Details
- Journal
- IEEE Robotics and Automation Letters
- Published
- 2025-05-30
- DOI
- https://doi.org/10.1109/lra.2025.3575013
- Primary Topic
- Interactive and Immersive Displays
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Natural Science Foundation of China