Skip to content

论文arXiv 2610.0208927 天前

HumanoidToolBench:从选择到移动执行的人形机器人工具使用基准测试 ​

原文链接:HumanoidToolBench: Benchmarking Humanoid Tool Use from Selection to Mobile Execution(arXiv 2610.02089 · 发布日期:2026-10-01)

HumanoidToolBench: Benchmarking Humanoid Tool Use from Selection to Mobile Execution

提出HumanoidToolBench和ToolBook,评测人形机器人从工具选择到移动执行的工具使用,并在仿真和真实Unitree G1上验证。

核心要点 ​

  • 动机:现有基准未在人形机器人上联合评估工具选择、操作与移动能力。
  • 方法:构建18任务HumanoidToolBench,并配套ToolBook演示数据集。
  • 实验:评估仿真七种、真机三种策略,显示选对工具与完成任务仍有差距。
  • 代码和数据已公开;探针显示未见工具选择准确率下降,并会在无关指令下继续执行。

验证范围与局限:验证范围为18个工具使用任务,覆盖仿真与一台真实Unitree G1。

英文摘要 (Abstract) ​

As robotic hardware and learning methods advance, humanoids need tools to perform tasks beyond their inherent physical limits. Successful tool use requires selecting a suitable tool and coordinating manipulation and, when needed, locomotion to complete the task. Existing benchmarks do not jointly evaluate these capabilities on a humanoid. We introduce HumanoidToolBench, an 18-task benchmark spanning three scenarios, three execution levels, and two tool-set modes, together with ToolBook, a dataset of 3.1k demonstrations collected in simulation and on a real Unitree G1. Evaluation of seven policies in simulation and three on the real robot reveals substantial gaps between selecting a suitable tool and completing the task. Focused GR00T N1.7 probes show reduced selection accuracy on unseen tools and continued task execution under unrelated instructions. Code and data are available at https://snu-pi.github.io/HumanoidToolBench/.

作者:Kyochul Jang, Seohyeon Park, Ohchul Kwon, Sangjun Park, Junhyeok Choi, Seungyeop Yi, Chaeyun Kim, Sangkyu Lee, Idan Szpektor, Avi Caciularu

来源

本页事实对应的公开页面,以原文为准。

  1. arXiv 2610.02089:HumanoidToolBench: Benchmarking Humanoid Tool Use from Selection to Mobile Execution访问 2026-10-21