陆颖 刘裕发 束瑜 季小洁

摘 要：為提高高级车辆事故自动呼救（advanced automatic crash notification，AACN）系统的救援效率，提出了一种驾驶员伤情预测算法，基于算法进行了AACN系统终端的总体设计。首先，选取驾驶员年龄、性别、是否佩戴安全带、速度变化量、事故碰撞方向以及驾驶员侧安全气囊是否打开作为驾驶员伤情的影响因素;其次，对交通事故数据进行分析并构建Logistic回归模型，使用Hosmer-Lemeshow测试表对模型进行验证，通过敏感性分析获得最佳触发阈值;再次，对AACN系统终端进行设计;最后，通过实际案例检验伤情预测算法的准确性和AACN系统终端的有效性。结果表明，所提出的驾驶员伤情预测算法和AACN系统的准确性较高，能够有效预测驾驶员伤情，有助于救援中心制定积极有效的救援方案。研究结果可用于解决现有的集中式AACN系统呼叫效率不高且准确性受人为因素影响较大的问题，为分散式AACN系统中驾驶员伤情预测算法的设计提供参考。

关键词：公路运输其他学科;道路事故数据;Logistic回归模型;驾驶员伤情预测模型;高级车辆事故自动呼救系统

中图分类号：U461.91   文献标识码：A

doi：10.7535/hbkd.2021yx04002

收稿日期：2021-03-31;修回日期：2021-05-24;责任编辑：冯 民

基金项目：国家自然科学基金（51605197）;江苏省自然科学基金（BK20160524）;镇江市重点研发计划（社会发展）项目（SH2019054）

第一作者简介：陆 颖（1981—），男，江苏镇江人，副教授，博士，主要从事汽车被动安全方面的研究。

E-mail：luying@ujs.edu.cn

陆颖，刘裕发，束瑜，等.面向高级车辆事故自动呼救系统的伤情预测[J].河北科技大学学报，2021，42（4）：327-333.LU Ying，LIU Yufa，SHU Yu， et al.Injury prediction for advanced automatic crash notification system[J].Journal of Hebei University of Science and Technology，2021，42（4）：327-333.

Injury prediction for advanced automatic crash notification system

LU Ying1，LIU Yufa1，SHU Yu2，JI Xiaojie1

（1.School of Automotive and Traffic Engineering，Jiangsu University，Zhenjiang，Jiangsu 212013，China;2.The Fourth People′s Hospital of Zhenjiang City，Zhenjiang，Jiangsu 212000，China）

Abstract：In order to improve the rescue efficiency of the advanced automatic crash notification （AACN） system，a driver injury prediction algorithm was proposed，and the overall design of the AACN system terminal was conducted based on this algorithm.First，the amount of speed change，the direction of the accident，the driver′s age，gender，whether to wear the seat belt，and whether the driver′s side airbag inflated were selected as the influencing factors of the driver′s injury.Next，a Logistic regression model was analyzed and developed based on traffic accident data.The effectiveness of the model was verified by using the Hosmer-Lemeshow test table，and the best trigger threshold was obtained through sensitivity analysis.Then，the AACN system terminal was designed.Finally，an actual case was used to test the accuracy of the injury prediction algorithm and the effectiveness of AACN system terminal.The results of the case study show that the proposed driver injury prediction algorithm and the AACN system are highly accurate，which can effectively predict the driver′s injury and help the rescue center work out an active rescue plan.The research results can be used to solve the problem that existing centralized AACN systems′ efficiency is not high and their accuracy is greatly affected by human factors.Consequently，they provide a reference for the design of driver injury prediction algorithm in the decentralized AACN system.

Keywords：

other subjects of highway transportation;road accident data;Logistic regression model;driver injury prediction model;the advanced automatic crash notification system

高级车辆事故自动呼救（advanced automatic crash notification，AACN）系统是一种事故后车辆安全技术。……