arXivDaily arXiv每日学术速递 周一至周五更新

期刊&会议

International Conference on Robotics and Automation · 会议 · Robotics

2026-05-20 至 2026-05-20 共收录 5
2605.20082 2026-05-20 cs.CV cs.AI

VL-DPO: Vision-Language-Guided Finetuning for Preference-Aligned Autonomous Driving

VL-DPO:基于视觉语言的偏好对齐自动驾驶微调

Zhefan Xu, Ghassen Jerfel, Marina Haliem, Qi Zhao, Jeonhyung Kang, Khaled S. Refaat

机构 * Waymo

AI总结 本文提出VL-DPO,一种基于视觉语言模型的框架,通过零样本推理生成偏好对来微调自动驾驶模型,以提升与人类驾驶偏好的对齐程度,实验表明该方法在RFS和ADE指标上均优于基线模型。

Comments Published in International Conference on Robotics and Automation (ICRA), 2026 8 pages, 6 figures, 4 tables

详情

展开后加载摘要…

URL PDF HTML 收藏
2601.20529 2026-05-20 cs.RO cs.MA

A Practical Framework of Key Performance Indicators for Multi-Robot Lunar and Planetary Field Tests

多机器人月球和行星实地测试的关键绩效指标实用框架

Julia Richter, David Oberacker, Gabriela Ligeza, Valentin T. Bickel, Philip Arm, William Talbot, Marvin Grosse Besselmann, Florian Kehl, Tristan Schnell, Hendrik Kolvenbach, Rüdiger Dillmann, Arne Roennau, Marco Hutter

机构 * Robotic Systems Lab (RSL), ETH Zürich(罗伯特系统实验室(RSL),苏黎世联邦理工学院) FZI Research Center for Information Technology(信息技术研究所以及中心) Machine Intelligence and Robotics Lab (MaiRo), Karlsruhe Institute for Technology (KIT)(机器智能与机器人实验室(MaiRo),卡尔斯鲁厄理工学院) Department of Environmental Sciences, University of Basel(巴塞尔大学环境科学系) European Space Agency/ESTEC(欧洲航天局/ESTEC) Center for Space and Habitability, University of Bern(伯尔尼大学空间与宜居性中心) Space Instruments Group, University of Zürich(苏黎世大学空间仪器组) Space Science and Technology, ETH Zürich(苏黎世联邦理工学院空间科学与技术)

AI总结 本文提出了一种基于多机器人月球场景的KPI框架,用于评估和比较不同实地测试的性能,强调效率、鲁棒性和精度的场景依赖性优先级,并通过实地测试验证了其在实际应用中的有效性。

Comments Presented at ICRA 2026 Workshop on Multi-Agent Robotic Systems: Real-World Collaboration and Interaction

详情

展开后加载摘要…

URL PDF HTML 收藏
2605.19690 2026-05-20 cs.RO

D-CLING: Prior-Preserving Depth-Conditioned Fine-Tuning for Navigation Foundation Models

D-CLING: 保留先验知识的深度条件细调方法用于导航基础模型

Shintaro Nakaoka, Takayuki Kanai, Kazuhito Tanaka

机构 * Frontier Research Center, Toyota Motor Corporation(丰田电机公司前沿研究中心)

AI总结 本文提出了一种新的细调方法,通过利用大规模预训练同时高效学习新环境或相机配置等新设置,从而在保留预训练知识的同时提升导航模型的鲁棒性和准确性。

Comments This paper has been accepted to the 2026 IEEE International Conference on Robotics and Automation (ICRA 2026), which will be held in Vienna, Austria, from June 1 to 5, 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2605.19104 2026-05-20 cs.RO cs.AI

Neural Operators for Design-Space Surrogate Modeling of Tendon-Actuated Continuum Robots

神经运算符用于腱驱动连续机器人设计空间的代理建模

Branden Frieden, James M. Ferguson, Alan Kuntz, Varun Shankar

机构 * The Robotics Center and the Kahlert School of Computing at the University of Utah(犹他大学机器人中心和Kahlert计算学院) The Departments of Computer Science and Electrical and Computer Engineering at Vanderbilt University(范德比大学计算机科学与电气与计算机工程系)

AI总结 本文提出了一种基于神经运算符的学习方法,用于腱驱动连续机器人的设计空间代理建模,通过映射机器人设计参数和腱驱动输入到最终配置,实现跨大量机器人设计的泛化能力。

Comments Accepted to ICRA 2026

详情

展开后加载摘要…

URL PDF HTML 收藏
2509.14787 2026-05-20 cs.RO

COMPASS: Confined-space Manipulation Planning with Active Sensing Strategy

COMPASS:基于主动感知策略的受限空间操作规划

Qixuan Li, Chen Le, Dongyue Huang, Jincheng Yu, Xinlei Chen

机构 * Shenzhen International Graduate School, Tsinghua University, Shenzhen, China(清华大学深圳国际研究生院,中国深圳) Department of Electronic Engineering, and the Institute for Embodied Intelligence and Robotics, Tsinghua University, Beijing, China(清华大学电子工程系,以及 embodied intelligence and robotics 研究院,中国北京) The School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore(南洋理工大学电子与电气工程学院,新加坡)

AI总结 本文提出COMPASS框架,通过主动感知策略在受限和杂乱环境中实现安全操作,提高了操作成功率。

Comments Accepted to the 2026 IEEE International Conference on Robotics and Automation (ICRA 2026)

详情

展开后加载摘要…

URL PDF HTML 收藏