arXiv Robotics — research abstracts· Yuehao Huang, Yunzi Wu, Xiaotao Zhang, Xinhai Li, Jiankun Dong, Sunyao Zhou, Jiajun Lv, Chi Zhang, Chenjia Bai, Yong Liu, Xuelong Li·· 2 days agoSignalEditorial score82
WNM-3D: A World Navigation Model with 3D Scene Conditioning for Closed-Loop VLN
WNM-3D: A World Navigation Model with 3D Scene Conditioning for Closed-Loop VLN
Summary
WNM-3D is a generative model for vision-language navigation that uses 3D scene geometry to improve navigation actions. It outperforms 2D versions in simulation and real-world deployment, with component ablations supporting its design.
Editorial context
This paper introduces WNM-3D, a generative world navigation model that conditions navigation actions on 3D scene geometry, improving performance over 2D counterparts in both simulation and real-world deployment.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
Source：arXiv Robotics — research abstracts · arxiv.org