Crash Narrative-Guided Countermeasure Recommendation Using Large Language Models: A Retrieval-Augmented Generation Framework for Intersection Safety
Crash Narrative-Guided Countermeasure Recommendation Using Large Language Models: A Retrieval-Augmented Generation Framework for Intersection Safety
基于事故叙述引导的大语言模型对策推荐：一种用于交叉路口安全的检索增强生成框架
Abstract: Improving safety at intersections requires identifying crash mechanisms and recommending appropriate countermeasures. However, this process traditionally relies on expert judgment, making it labor-intensive, difficult to scale, and dependent on the availability of experienced traffic safety engineers. Although crash narratives contain rich description of crash mechanisms, this unstructured information remains largely underutilized in safety analyses.
摘要： 提升交叉路口的安全性需要识别事故机制并推荐相应的对策。然而，这一过程传统上依赖于专家判断，导致其不仅劳动密集、难以扩展，还受限于经验丰富的交通安全工程师的可用性。尽管事故叙述中包含了关于事故机制的丰富描述，但这些非结构化信息在安全分析中仍未得到充分利用。
This study presents a crash narrative-guided retrieval-augmented generation (RAG) framework that translates narrative-derived crash mechanisms into site-specific countermeasure recommendations. Key mechanism attributes including traffic control, signal indication, driver fault, vehicle movement, and travel direction were extracted from crash narratives and linked to evidence-based treatments from the FHWA Proven Safety Countermeasures and the CMF Clearinghouse.
本研究提出了一种由事故叙述引导的检索增强生成（RAG）框架，该框架能将从叙述中提取的事故机制转化为针对特定地点的对策建议。研究从事故叙述中提取了关键机制属性，包括交通管制、信号指示、驾驶员过失、车辆移动和行驶方向，并将其与来自美国联邦公路管理局（FHWA）“经验证的安全对策”及“事故修正因子（CMF）信息库”的循证处理方案相关联。
The framework integrates embedding-based retrieval of historically similar intersections, association-rule mining, statistical guidance on the expected number of relevant countermeasures, and an engineering reasoning guidance that directs LLM through a domain-consistent decision process before selecting countermeasures.
该框架整合了基于嵌入（Embedding）的历史相似交叉路口检索、关联规则挖掘、关于相关对策预期数量的统计指导，以及工程推理指导，引导大语言模型（LLM）在选择对策前经历一个符合领域逻辑的决策过程。
Evaluated on 312 fatal and serious-injury crashes across 115 intersections in Lake and Sumter Counties, Florida, using five-fold cross-validation, the framework achieved a precision of 0.82, recall of 0.85, and F1-score of 0.82, while recommending an average of 3.91 countermeasures per location with 3.14 matching, closely matching the actual average (3.86).
通过对佛罗里达州莱克县和萨姆特县 115 个交叉路口的 312 起致命及重伤事故进行五折交叉验证评估，该框架实现了 0.82 的精确率、0.85 的召回率和 0.82 的 F1 分数。同时，该框架平均每个地点推荐 3.91 项对策，其中 3.14 项与实际情况吻合，非常接近实际平均值（3.86）。
Overall, the proposed framework demonstrates the potential of retrieval-augmented LLMs as an interpretable and scalable decision-support tool for transportation agencies for translating crash narratives into countermeasure recommendations.
总体而言，所提出的框架展示了检索增强型大语言模型作为一种可解释且可扩展的决策支持工具的潜力，能够帮助交通部门将事故叙述转化为有效的对策建议。