T(R,O) Grasp++: Cross-Embodiment Dexterous Grasping in Cluttered Environments at Scale
A simulation benchmark of about 0.5B grasp labels over eight dexterous hands and ~50K objects, plus a diffusion model for cross-embodiment grasps in clutter: 88% average success, 0.3 s inference.构建约 5 亿条抓取标注、覆盖 8 种灵巧手和约 5 万个物体的仿真基准，并提出跨本体的杂乱场景灵巧抓取扩散模型：平均成功率 88%，推理 0.3 秒。
BibTeX
@misc{fei2026t,
title = {{T(R,O) Grasp++: Cross-Embodiment Dexterous Grasping in Cluttered Environments at Scale}},
author = {Xin Fei and Zhixuan Xu and Huaicong Fang and Fangjun Chen and Zhaoyu Mu and Rui Yang and Zhexi Luo and Lin Shao},
year = {2026},
eprint = {2610.25410},
archivePrefix = {arXiv},
primaryClass = {cs.RO},
url = {https://arxiv.org/abs/2610.25410},
}Manipulation & Dexterous Hands操作与灵巧手Dexterous hand灵巧手dexterous-graspingcross-embodimentclutterdiffusionbenchmark