SR-VLN: Implicit Spatial Reasoning Vision-and-Language Navigation
Sensors · 15.06.2026 · 26 (12) · 3809
DOI: 10.3390/s26123809 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 15.06.2026. Verfasst von Ruolin Zhu und 2 weiteren Autor:innen. Beteiligte Einrichtung: School of Information and Communication Engineering, Communication University of China, Beijing 100024, China. Die Arbeit nennt 49 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Vision-and-language navigation (VLN) traditionally relies on explicit reasoning chains, which, despite being interpretable, impose severe constraints on inference efficiency and scalability in long-range environments. Existing multimodal large language models (MLLMs) frequently encounter latency bottlenecks due to the generation of verbose textual narratives during decision-making. To address these limitations, we propose spatial reasoning vision-and-language navigation (SR-VLN), a novel framework that shifts the paradigm from explicit chain-of-thought (CoT) to an implicit spatial representation space. SR-VLN introduces a pyramidal hierarchical history framework integrated with perceptual compression to condense historical trajectories into multi-scale representations, effectively minimizing token overhead while preserving critical spatial semantics. Rather than generating verbose textual reasoning steps, SR-VLN employs compact, learnable spatial tokens (S-Tokens) to perform agile inference directly within the latent feature space. To establish robust causal mappings between these implicit states and navigational actions, we employ a hybrid training strategy that combines sparse reward supervision with reinforcement learning via GRPO. Extensive evaluations on the R2R, REVERIE, and SOON datasets demonstrate that SR-VLN achieves state-of-the-art overall navigation performance, while maintaining a comparable balance between accuracy and efficiency. Compared to explicit reasoning baselines, our method reduces token consumption by 68% and achieves a 4.1× speedup in inference while reaching a 76.02% success rate and a 73.80% SPL on the R2R unseen split, thereby facilitating near-real-time action prediction in long-range navigation environments.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- School of Information and Communication Engineering, Communication University of China, Beijing 100024, China.
- School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing 100088, China.
ORCID-Profile: Ruolin Zhu
Zitieren
APA
Zhu, R., Li, S., & Yang, M. (2026). SR-VLN: Implicit Spatial Reasoning Vision-and-Language Navigation. Sensors, 26(12), 3809. https://doi.org/10.3390/s26123809
Vancouver
Zhu R, Li S, Yang M. SR-VLN: Implicit Spatial Reasoning Vision-and-Language Navigation. Sensors. 2026;26(12):3809. doi:10.3390/s26123809
BibTeX
@article{Zhu2026,
author = {Zhu, Ruolin and Li, Shaobin and Yang, Min},
title = {SR-VLN: Implicit Spatial Reasoning Vision-and-Language Navigation},
journal = {Sensors},
year = {2026},
volume = {26},
number = {12},
pages = {3809},
doi = {10.3390/s26123809}
}
EUVIMED