主办：陕西省汽车工程学会
ISSN 1671-7988  CN 61-1394/TH
创刊：1976年
汽车实用技术

汽车实用技术 ›› 2025, Vol. 50 ›› Issue (5): 142-147.DOI: 10.16638/j.cnki.1671-7988.2025.005.027

• 标准·法规·管理 • 上一篇    

基于贝叶斯网络的海底隧道交通事故风险预测研究

邢英，胡雪芳，刘子萌   

  1. 青岛黄海学院 智能制造学院
  • 发布日期:2025-03-12
  • 通讯作者: 邢英
  • 作者简介:邢英（1994－），女，硕士，讲师，研究方向为道路交通安全，E-mail:xingying811@163.com
  • 基金资助:
    青岛黄海学院科技计划项目（2022KJ10）

Research on Risk Prediction of Submarine Tunnel Traffic Accidents Based on the Bayesian Network

XING Ying, HU Xuefang, LIU Zimeng   

  1. School of Intelligent Manufacturing, Qingdao Huanghai University
  • Published:2025-03-12
  • Contact: XING Ying

PDF

摘要： 为了准确预测海底隧道交通事故风险等级，提高内部行车安全，通过分析某海底隧道 2013 至 2018 年间发生的 1 653 起交通事故样本，得到了驾驶人、道路、环境、管理和车辆五 方面的风险因素，构造出贝叶斯网络，提出基于贝叶斯网络的海底隧道交通事故风险预测方 法。此方法利用该海底隧道 2013 至 2018 年交通事故数据进行参数学习，建立起贝叶斯预测 模型，得出该海底隧道交通事故风险处于重大风险状态的概率为 3%，处于较大风险状态的概 率为 4%，处于一般风险状态的概率为 14%，处于低风险状态的概率为 79%，结合实例验证了 模型的准确性，并通过敏感性分析针对影响显著的因素提出防范措施，降低海底隧道交通事 故风险。

关键词: 海底隧道；交通事故；风险因素；贝叶斯网络；风险预测

Abstract: In order to accurately predict the risk level of traffic accidents in the submarine tunnels and improve the safety of internal driving, by analyzing the 1 653 traffic accident samples that occurred in a certain submarine tunnel from 2013 to 2018, the risk factors of drivers, roads, environment, management and vehicles are obtained, a Bayesian network is constructed and the risk prediction method of traffic accidents is proposed based on the Bayesian network. This method utilizes the traffic accident data of the submarine tunnel from 2013 to 2018 for learning parameters to establish a Bayesian prediction model. The probability of the submarine tunnel traffic accident risk being in a major risk state is 3%, the probability of being in a high risk state is 4%, the probability of being in a general risk state is 14%, and the probability of being in a low-risk state is 79%. The accuracy of the model is verified through examples, and preventive measures are proposed for significant factors through sensitivity analysis to reduce the risk of submarine tunnel traffic accidents.

Key words: submarine tunnel; traffic accident; risk factors; Bayesian network; risk prediction