← Back to the feed
Space & aviationResearch findings

Safety4DM · A Vision-Language Framework for Risk-Aware Dynamic Maps in Autonomous Driving

Original published 2026.11.02 00:34 KSTAdded to Blackdeck 2026.11.03 15:23 KST

What has changed?

A pipeline that converts real-world road observations into fine-grained risk entities in dynamic maps is proposed

Key points

  1. Dynamic maps serve as a core information infrastructure in vehicle-road-cloud cooperative autonomous driving
  2. A pipeline converting real-world road observations into fine-grained risk entities is introduced
  3. A natural language interface enables accurate risk queries and generates actionable driving advice

Connect it to your work

A vision-language model is fine-tuned on real-world road images to generate structured risk representations directly usable by dynamic maps

What remains unproven

This research note is based on an abstract; check the original for test conditions and application scope.

Originals used for this note

Links open the original publisher in a new tab.

  1. arxivSafety4DM: A Vision-Language Framework for Risk-Aware Dynamic Maps in Vehicle-Road-Cloud Collaborative Driving