基于粒子群优化BP网络的山区高速公路交通事故预测模型  

Mountain Highway Traffic Accident Prediction Model Based on BP Neural Network Based on Particle Swarm Optimization

作　　者：季芳 熊坚[1] 杨文臣 郭凤香[1] JI Fang ,XIONG Jian , YANG Wenchen, GUO Fengxiang(1. School of Transportation Engineering, Kunming University of Science and Technology, Kunming 650500, China; 2. Broadvision Engineering Consultants,Kunming 650500,China)

机构地区：[1]昆明理工大学交通工程学院,云南昆明650500 [2]云南省交通规划设计研究院,云南昆明650500

出　　处：《电子科技》2018年第10期64-68,共5页Electronic Science and Technology

基　　金：云南省教育厅基金(2015Y082);陆地交通气象灾害防治技术国家工程实验室开放研究基金(NELJA201605)

摘　　要：针对传统预测模型在多因素方面下无法对交通事故做出有效预测的问题,文中采用基于粒子群优化的BP神经网络建立了交通事故严重等级预测模型。该模型是以"影响因素-事故严重等级"的输入输出形式来实现对事故的预测的。文中通过昆石高速公路的事故数据对模型进行预测验证,得到模型的预测结果准确率为89. 6%,证明该模型可用于事故严重等级的预测;然后,在该模型的基础上,进行了设定情境的仿真模拟,预测得出设定情境下的事故严重等级;最后,通过模型的仿真结果分析得出一定的事故发生规律。仿真结果分析表明,在以下几种情况更易发生严重事故:行车时间为凌晨0:00~8:00;天气状况为雨天和雾天;道路曲线半径在0~600 m;车辆类型为重型货车等。Aiming at the problem that conventional forecasting models cannot ettectively predict traffic accidentsunder multi - factors, this paper uses BP neural network based on particle swami optimization to establish a traffic acci-dent severity rating model. The model is based on the ＂ influencing factors - accident severity level＂ input and outputto achieve the prediction of the accident. In this paper, the model is predicted and verified by the accident data of theKunshi Expressway, and the accuracy rate of the predicted result of the model is 89.6% , which proves that the modelcan be used to predict the severity of the accident. On the basis of the model, a simulation simulation of the set situation is perfomled to predict the severity of the accident in the set situation. Finally, through the simulation results ofthe model, the law of occuncence of certain accidents is obtained. Analysis of simulation results shows that serious acci-dents are more likely to occur in the following situations travel time is 0：00 - 8：00 in the early morning; weather con-ditions are rainy and loggy; road cmwe radius is 0 -600 meters; vehicle type is heavy truck Wait.

关 键 词：山区高速公路 BP神经网络 粒子群算法 交通事故预测 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]