Jev-AD：通过语言世界模型实现安全的语言引导自动驾驶
Jev with Foresight: Safe Language-Guided Autonomous Driving via a Language World Model
AI 导读
Jev-AD 将以概率回答预设问题的类型化决策模型 Jev 与语言世界模型结合，为每个候选动作预测未来两秒内附近车辆的运动。在 500 回合仿真高速公路汇入中，前瞻将碰撞率从 26.0% 降至 2.4%，每步决策耗时 0.25 秒，自由指令意图达成率 93%；在 500 回合 NGSIM 真实回放数据中，碰撞率从 48.2% 降至 2.8%，世界模型可预见 93% 的危险动作。
正文 · 原文
Abstract:Large language models understand passengers' driving commands, but as driving agents they reason in generated text for seconds and must infer each action's consequences implicitly from the scene. We present Jev-AD, which pairs Jev, a typed decision model that answers predefined questions with probabilities, not generated text, with a language world model that predicts for every candidate action how nearby vehicles move over the next two seconds. Code turns these predictions into foreseen consequences, which Jev compares instead of inferring them from coordinates. In simulated highway merging (500 car-episodes), this foresight lowers the same decision model's collision rate from 26.0% to 2.4% at 0.25 seconds per decision, and Jev-AD achieves the passenger's intent in 93% of free-form command trials. In 500 replayed real-world NGSIM episodes, foresight lowers the collision rate from 48.2% to 2.8%, and the world model, learned from this traffic, foresees 93% of dangerous actions. Project page: this https URL.
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.05448 [cs.RO] |
| (or arXiv:2611.05448v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2611.05448 arXiv-issued DOI via DataCite |
Submission history
From: Yansong Qu [view email]
[v1]
Tue, 3 Nov 2026 17:36:47 UTC (1,173 KB)
来源：arXiv cs.AI · arxiv.org