Influence Research Of Serious And Major Road Traffic Accident Based On Bayesian Network

With the rapid development of transportation,the problem of traffic safety is becoming more and more serious.The traffic accidents that China facing,especially serious traffic accidents,will be even more severe.At present,the research on serious traffic accidents mainly focused on the statistical analysis of accident characteristics and influencing factors,the research depth and attention are not enough and lack of systematic analysis and regularity understanding of such accidents.Therefore,aiming at the current situation and existing problems of serious road traffic accidents in China,this paper conducts a preliminary study on the influencing factors of such accidents and their effects on crash severity.First of all,on the basis of defining the concept of serious traffic accidents,its distribution characteristics in terms of time,space,accident state and vehicle types were statistically analyzed.And the distribution differences of serious traffic accident and general road accidents were compared from the aspects of drivers,vehicles,roads and environment.Based on statistical analysis,from the perspective of single-car and multi-vehicles,correlation between different influencing factors and the number of serious traffic accidents,deaths and injuries were analyzed.According to the analysis results,main influencing factors were preliminarily screened.Then,based on the principle of maximum deviation,improve the traditional grey correlation method,weighted grey correlation was used to measure the correlation between different factors and serious traffic accidents.According to the weighted grey correlation,the key influencing factors can be obtained from many factors.Secondly,based on Bayesian network,the accident impact degree model for single-car and multi-vehicles were established respectively,to study the effects of different factors on the number of deaths and injuries in serious traffic accident.On the basis of structural learning,the hierarchical relationship between different factors and the number of deaths and injuries were obtained.And the model validation was verified through comparative analysis of probability learning results and the measured data.Then,based on the established model,the in-depth discussion of the impact of different factors on the consequences of serious traffic accidents were conducted through Bayesian network reasoning method.Applying joint tree algorithm,the probability distributions of deaths and injuries in single-car and multi-vehicle accidents affected by different factors were obtained.Using the interval number theory,the risk ranking of different influencing factors of serious traffic accidents were obtained,then the risk factors that have significant influence on the consequences of accidents are acquired.Finally,the impacts of various factors on the consequences of single-car and multi-vehicle accidents were analyzed.And based on the analysis results,active prevention for serious traffic accidents can be executed.The factor combinations that have a significant impact on the severity of traffic accidents were identified from a number of factors,which can be regarded as the key control object of road safety management work and has important guiding significance for reducing accidents.