Bridging vision-language reasoning and learned control for legged navigation in the wild
Science Robotics · 18.11.2026 · 11 (120)
DOI: 10.1126/scirobotics.aee1549 ↗Kurz gefasst
Fachartikel in der Zeitschrift Science Robotics, erschienen am 18.11.2026. Verfasst von Takahiro Miki und 8 weiteren Autor:innen. Beteiligte Einrichtung: ETH Zurich, Zurich, Switzerland. Die Arbeit nennt 86 Literaturangaben. Gefördert von: Horizon Therapeutics, Alzheimer Nadační Fond.
Abstract (Originalsprache)
Autonomous navigation in diverse real-world environments remains one of the greatest challenges in robotics. It demands both robust low-level motion control to handle diverse terrain and high-level reasoning to interpret environmental cues. We present ROAM (Robust Open-world Autonomous Mobility), a system that enables legged robots to perform kilometer-scale navigation in complex alpine and urban settings. ROAM combines the world knowledge of a vision-language model (VLM), pretrained on internet-scale data, with robust locomotion and local planning policies learned in simulation via reinforcement learning (RL). The VLM provides common-sense reasoning and semantic interpretation of the scene, assessing risks, interpreting signs, and producing semantic labels with associated navigation costs. An open-vocabulary segmentation module converts these labels into a dense, real-time risk map, which conditions two pretrained RL policies responsible for agile locomotion and local navigation: effectively coupling high-level VLM reasoning with low-level physical control. Real-world experiments showed the system’s ability to autonomously navigate a wide range of complex real-world environments. The robot completed kilometer-long hikes with minimal intervention, interpreted signs to avoid danger, and evaluated its environment to identify safe areas on which to step. This work represents a step toward the reliable deployment of systems in unstructured environments. However, it also underscores the challenges that remain in VLM reliability and spatial reasoning, indicating that foundational models are a crucial component but not yet a complete solution for navigation in the wild.
Autor:innen und Institutionen
- ETH Zurich, Zurich, Switzerland.
- Apple, Cupertino, CA, USA.
ORCID-Profile: Takahiro Miki, Robin Schmid, Manthan Patel, Turcan Tuna, Julia Richter, Julian Nubert, Joonho Lee, Matthias Müller, Marco Hutter
Förderung
- Horizon Therapeutics
- Alzheimer Nadační Fond
Zitieren
APA
Miki, T., Schmid, R., Patel, M., Tuna, T., Richter, J., Nubert, J., Lee, J., Müller, M., & Hutter, M. (2026). Bridging vision-language reasoning and learned control for legged navigation in the wild. Science Robotics, 11(120). https://doi.org/10.1126/scirobotics.aee1549
Vancouver
Miki T, Schmid R, Patel M, Tuna T, Richter J, Nubert J, et al. Bridging vision-language reasoning and learned control for legged navigation in the wild. Science Robotics. 2026;11(120). doi:10.1126/scirobotics.aee1549
BibTeX
@article{Miki2026,
author = {Miki, Takahiro and Schmid, Robin and Patel, Manthan and Tuna, Turcan and Richter, Julia and Nubert, Julian and Lee, Joonho and Müller, Matthias and Hutter, Marco},
title = {Bridging vision-language reasoning and learned control for legged navigation in the wild},
journal = {Science Robotics},
year = {2026},
volume = {11},
number = {120},
doi = {10.1126/scirobotics.aee1549}
}
EUVIMED