Metraipolis › archive › dc_3da37a31d062
Geo-VLA: Geometry-Aware Vision-Language-Action Planning via Internalization of Map Semantics — arXiv:2608.21440
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.21440
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T07:42:13.871Z
- SHA-256 of the archived text
6132a141a47f6ef929279fdb86f31c0d19d7b6521216e57ec46dc5024b3e0fe0- Stable URL of this page
https://metraipolis.com/documento/dc_3da37a31d062
How to cite
Geo-VLA: Geometry-Aware Vision-Language-Action Planning via Internalization of Map Semantics — arXiv:2608.21440 — arXiv (en), https://arxiv.org/abs/2608.21440 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_3da37a31d062 (Metraipolis archive dc_3da37a31d062, sha256 6132a141a47f6ef929279fdb86f31c0d19d7b6521216e57ec46dc5024b3e0fe0)
BibTeX
@misc{dc3da37a31d062,
title = {Geo-VLA: Geometry-Aware Vision-Language-Action Planning via Internalization of Map Semantics — arXiv:2608.21440},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.21440},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_3da37a31d062; sha256 6132a141a47f6ef929279fdb86f31c0d19d7b6521216e57ec46dc5024b3e0fe0},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.21440. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Vision-language-action (VLA) models have advanced end-to-end autonomous driving by leveraging foundation models for semantic reasoning and long-tail generalization. However, their planning performance remains limited in complex driving environments because image-only representations inadequately capture planning-relevant road geometry and topology. In this paper, we propose Geo-VLA, a plug-and-play framework that enhances VLA models by learning geometry-aware visual representations. During training, Geo-VLA internalizes geometric map semantics to strengthen road-structure representations, while requiring no HD maps or additional lane information during inference. To support this approach, we introduce Geo-QA, a geometry-focused question-answering dataset that injects road geometry into vision-language representations through contrastive learning and instruction tuning. Experiments on NAVSIM v1 demonstrate that Geo-VLA consistently improves VLA planners with distinct action-generation architectures, achieving 92.1 PDMS and establishing a new state-of-the-art among single-camera VLA planners. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-25 in cs.AI). Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.21440 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-18T07:37:52.811Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).