愿景协 driver：利用LLM和HUD增强驾驶员风险感知
人机交互
2025-11-19 v1
摘要
驾驶员对危险情境的感知始终是驾驶中的一个挑战。现有的风险检测方法擅长识别碰撞，但在评估非碰撞情境下路用户行为方面面临挑战。本文引入Visionary Co-Driver系统，利用大语言模型识别非碰撞路侧风险并基于驾驶员的眼动信号发出警报。具体而言，系统结合视频处理算法和大语言模型来识别潜在的风险路用户。这些风险在适应性的头显界面上动态指示，以增强驾驶员的注意力。41名驾驶员的用户研究确认，Visionary Co-Driver提高了驾驶员的风险感知，支持其识别路侧风险。
引用
@article{arxiv.2511.14233,
title = {Visionary Co-Driver: Enhancing Driver Perception of Potential Risks with LLM and HUD},
author = {Wei Xiang and Ziyue Lei and Jie Wang and Yingying Huang and Qi Zheng and Tianyi Zhang and An Zhao and Lingyun Sun},
journal= {arXiv preprint arXiv:2511.14233},
year = {2025}
}
备注
Accepted for publication in IEEE Transactions on Intelligent Transportation Systems (T-ITS)