Research On Risk Evaluation Of Highway Traffic Safety Based On Bayesian Network

In order to meet the increasing demand of transportation,the mileage of highways in China has been expanding.In 2018,the total mileage of the highway was 4.8465 million kilometers,an increase of 73000 kilometers compared with 2017.The density of the highway is constantly improving,and the scope of the highway can be covered is also wider and wider.With the continuous growth of highway mileage,the number of traffic accidents is also increasing.In 2018,there were 244937 traffic accidents,an increase of 41888 compared with 2017.In 2018,the number of people killed by traffic accidents was 63194,which shows that China's traffic safety situation is more severe.In view of the increasingly severe situation of highway traffic safety,it is necessary to carry out risk assessment of highway traffic safety.Highway traffic safety risk assessment can effectively identify the hidden dangers of accidents,reduce the severity of traffic accidents,and evaluate the risk factors that affect highway traffic safety.Based on the evaluation results,targeted rectification suggestions can be made to greatly improve the safety of highway traffic.So as to reduce the occurrence of traffic accidents and reduce the harm of accidents,and better protect people's lives and property safety.The main research contents of this article include the following aspects:The main influencing factors of highway traffic safety were deeply explored,and the risk evaluation index system was established.The road transportation system is a dynamic and complex system.There are many influencing factors on road traffic safety,including road,environmental,vehicle,and human factors.The four influencing factors are studied separately to determine the influencing factors The specific indicators included finally build a highway risk evaluation index system consisting of a target layer,a criterion layer,and an indicator layer.Construct a highway safety risk assessment model based on Bayesian network.Using Bayesian network analysis software GeNIe,combining expert knowledge and machine learning,determining the relationship between nodes to construct a Bayesian network;using the data obtained from the questionnaire to determine the value of each initial node,and then constructing a complete Bayesian network Network security risk evaluation model;use the constructed Bayesian network highway security risk evaluation model to conduct reverse reasoning and sensitivity analysis,find out the main factors affecting highway traffic safety,and propose targeted rectification suggestionsCombined with examples to verify the constructed highway traffic safety risk assessment model.This paper takes provincial highway 238 as an example to carry out highway safety risk assessment,elaborate the highway traffic safety risk assessment process in detail,and finally get the result of risk assessment.At the same time,AHP was used to analyze the safety of Provincial Highway 238,and the obtained results were compared with the results of safety evaluation based on Bayesian network.The scientific,reliability and accuracy of the road safety risk evaluation model based on Bayesian network are verified by comparison.