Metraipolis › archive › dc_352120ec5fb9

Sometimes You Gotta Run Before You Can Walk: Run-then-Walk Scheduling Strategy for VLM Autonomous Driving — arXiv:2609.25831

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.25831
License (as archived)
CC0-1.0
Archived
2026-10-08T16:43:08.922Z
SHA-256 of the archived text
992dad6ef921f5353a6486e22df515971ff58c9ae387d127e6c6a07842f5b240
Stable URL of this page
https://metraipolis.com/documento/dc_352120ec5fb9

How to cite

Sometimes You Gotta Run Before You Can Walk: Run-then-Walk Scheduling Strategy for VLM Autonomous Driving — arXiv:2609.25831 — arXiv (en), https://arxiv.org/abs/2609.25831 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_352120ec5fb9 (Metraipolis archive dc_352120ec5fb9, sha256 992dad6ef921f5353a6486e22df515971ff58c9ae387d127e6c6a07842f5b240)
BibTeX
@misc{dc352120ec5fb9,
  title = {Sometimes You Gotta Run Before You Can Walk: Run-then-Walk Scheduling Strategy for VLM Autonomous Driving — arXiv:2609.25831},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.25831},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_352120ec5fb9; sha256 992dad6ef921f5353a6486e22df515971ff58c9ae387d127e6c6a07842f5b240},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.25831. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Recent VLM-based autonomous driving planners adopt GRPO-style reinforcement learning to optimize driving performance. However, existing GRPO recipes either optimize driving efficiency, risking progress-seeking but unsafe behavior, or enforce early safety constraints, leading to overly conservative behavior; both require lengthy training. To solve these problems, we first reveal two distinct RL regimes: a progress regime (Run-GRPO) that aggressively explores high progress, and a safety regime (Walk-GRPO) that restores safety under stable progress. Based on this finding, we propose $\textit{Run-then-Walk}$, a simple yet effective two-stage reward scheduling strategy for GRPO, achieving both better performance and faster convergence. Unlike one-stage RL, which may focus on progress, safety, or a mixture of both within a single training phase, this schedule explicitly separates progress discovery from safety repair. In the $\textit{Run}$ phase, we focus on progress, allowing the policy to escape the conservative bias and discover high-progress modes. In the subsequent $\textit{Walk}$ phase, we introduce endpoint and safety strategy to repair unsafe behaviors from the Run phase. This reversed schedule overcomes the conservatism of Walk-first methods and the unsafe progress-seeking of joint optimization. We validate it with various VLM-based planners on multiple benchmarks: NAVSIMv1, NAVSIMv2, Navhard, and nuScenes. Extensive experiments demonstrate improved driving performance while requiring 40--50\% fewer RL training epochs than the baselines.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-23 in cs.CV).
Categorie arXiv: Robotics (cs.RO) ; Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.25831 — 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 2026-10-08T15:56:18.242Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-23?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).