SPAN-Nav: Generalized Spatial Awareness for Versatile Vision-Language Navigation

Jiahang Liu^*, Tianyu Xu^*, Jiawei Chen^*, Lu Yue^*, Jiazhao Zhang^*, Zhiyong Wang^*, Minghan Li, Qisheng Zhao, Anqi Li, Qi Su, Zhizheng Zhang, He Wang

CoRL 2026

arXiv Project bibtex

* Equal contribution

SPAN-Nav spatial awareness for vision-language, urban, and point-goal navigation.

SPAN-Nav learns spatial awareness by predicting 3D occupancy from RGB video, compresses the spatial prior into a single token, and incorporates it into navigation reasoning.

My contribution: Motion-planning trajectory collection, occupancy-data preparation, automatic outdoor scene generation in Isaac Sim, and evaluation on MetaUrban and UrbanSim.