首页|Findings from University of Freiburg Yields New Findings on Robotics and Automat ion (Language-grounded Dynamic Scene Graphs for Interactive Object Search With M obile Manipulation)
Findings from University of Freiburg Yields New Findings on Robotics and Automat ion (Language-grounded Dynamic Scene Graphs for Interactive Object Search With M obile Manipulation)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
A new study on Robotics-Robotics and Automation is now available. According to news reporting originating in Freibur g, Germany, by NewsRx journalists, research stated, "To fully leverage the capab ilities of mobile manipulation robots, it is imperative that they are able to au tonomously execute long-horizon tasks in large unexplored environments. While la rge language models (LLMs) have shown emergent reasoning skills on arbitrary tas ks, existing work primarily concentrates on explored environments, typically foc using on either navigation or manipulation tasks in isolation." Funders for this research include Toyota Motor Europe (TME), Nvidia Corporation. The news reporters obtained a quote from the research from the University of Fre iburg, "In this work, we propose MoMa-LLM, a novel approach that grounds languag e models within structured representations derived from open-vocabulary scene gr aphs, dynamically updated as the environment is explored. We tightly interleave these representations with an object-centric action space. Given object detectio ns, the resulting approach is zero-shot, open-vocabulary, and readily extendable to a spectrum of mobile manipulation and household robotic tasks. We demonstrat e the effectiveness of MoMa-LLM in a novel semantic interactive search task in l arge realistic indoor environments."
FreiburgGermanyEuropeRobotics and AutomationRoboticsUniversity of Freiburg
NETL
NSTL
