Analysis And Prediction Of Factors Affecting The Severity Of Expressway Traffic Accidents

Expressway has become the most popular way for people to travel because of its fast speed and comfortable driving experience,but the traffic safety problems of expressway have come with it.While speed has become a major advantage of expressway,its disadvantages are also obvious,that is,the consequences of traffic accidents are more serious than ordinary highways.In order to solve this problem,scholars at home and abroad have carried out a series of studies on the cause and prediction of accidents,and their research results provide ideas for the research content of this paper.Therefore,this paper will conduct relevant research on the factors affecting the severity of expressway traffic accidents in China.First of all,the data of expressway traffic accidents were statistically analyzed.Based on the previous accident data,this paper divided the data into three categories according to the collision category:unilateral collision,rear-end collision and side collision,selected 12categories of factors related to the driver,vehicle,road and environment as independent variables,and divided the severity of the accident into three categories of non-injury,injury and death as dependent variables,and preliminarily analyzed the distribution of each factor under different collision categories through frequency statistics.Through Pearson correlation analysis,the independent variables with strong correlation were found,and after removing them,10 categories of factors were left as the research object.Secondly,a standard multiple logit model and a random parameter logit model were built for the severity of expressway traffic accidents,and their influencing factors were analyzed.In the modeling results of standard multiple logit and random parameter logit,the factors that have significant impact on casualties in three types of collision accidents were compared,and the results showed that the significance of the same factor in different types of collision accidents is different.At the same time,AIC,BIC and Mcfadden R~2 were used to test and evaluate the two models,and the results showed that the fitting effect of the random parameter logit model was better than the standard multiple logit model.Then,the marginal effect under the logit model with random parameters was calculated,and the impact of each factor on the severity of the accident was quantitatively analyzed.Thirdly,a genetic algorithm optimized BP neural network(GA-BP neural network)model was established to predict the severity of expressway traffic accidents.According to the analysis results in Chapter 3,the variables that are significantly related to the severity of the three types of collision accidents were respectively included as input information into the BP neural network to predict the severity of the accident.However,due to the shortcomings of BP neural network such as low computational efficiency and easy to fall into local extremum,this paper proposed to add genetic algorithm to BP neural network,established GA-BP neural network model to optimize its shortcomings,and realized the prediction of accident severity through the GA-BP neural network model.The article also compared the prediction accuracy of BP neural network and GA-BP neural network.The result showed that the accuracy of the neural network added with genetic algorithm is 4%-6%higher than the original model.The ROC curve verified that the prediction effect of the optimized GA-BP neural network is better than that of BP neural network.Finally,in view of the risk source in the above research sections,improvement measures were proposed from three aspects of traffic management,traffic engineering and the construction of smart expressways.Regular emergency response achievement test for older drivers,timely intervention for sections with potential safety hazards,and strengthened management and control can effectively reduce the incidence of accidents.At the same time,an emergency response system for accident rescue will be established,with multi department linkage and information sharing,so as to timely respond to accidents and minimize accident injuries..The research results of this paper can provide reference for traffic management departments,and can also be used for accident research and early warning.At the same time,the vision of making travel safer and more efficient can be realized by building smart high-speed.