Research On Traffic Accident Prediction Model Of Negative Binomial By Introducing Pavement Condition

Road conditions are an important cause of traffic accidents.Among them,traffic volume and geometrical alignments are the key road traffic conditions which causes accidents,and the research on this aspect has achieved many results.In addition,pavement condition is also an important road conditions affecting accidents,but because of the difficult availability of data,the research progress is not much.At present,the development of large-scale data and the normalization of road detection technology make the data of pavement condition can be obtained,so this paper constructs a freeway accident prediction model based on road traffic condition and pavement condition.Common pavement condition indicators include five categories: road quality index,road damage index,pavement structure strength index,skid resistance index and road rut depth index,according to the data,the range of indicators from 0 to 100,which representing the theoretically worst and the best of pavement condition.The accident prediction model is established by analyzing the freeway traffic condition and pavement condition.Firstly,taking the Shenshan freeway of Liaoning Province as the research object,the traffic volume data,the data of geometrical linear,pavement condition index and accident are analyzed.This paper chooses the homogeneity method to divide the highway section by analyzing the common road section dividing method,and raises the basic evaluation unit for accident prediction.Secondly,the characteristics and applicability of common accident prediction models are analyzed,and according to the statistic distribution characteristics of freeway accidents,the model of accident prediction in this paper is negative binomial.The paper chooses the traffic volume,geometrical alignment index and pavement condition index as the independent variable of the model,and the annual accident rate as the dependent variable,according to 95% confidence level to construct the accident prediction model based on negative binomial form finally.Then,according to the principle of sensitivity analysis,the relative sensitivity of the factors affecting the freeway accidents is analyzed,and the influence degree of the variables is analyzed according to the re sult.According to the relationship between geometrical alignment index and pavement condition index and accident rate,the optimal value of the index is obtained.Finally,taking part of the data of Shenshan freeway as an example,this paper validates the accuracy of accident prediction model.Taking the data of Yuegan freeway as an example,this paper applies the accident prediction model and verifies the accuracy of the model through actual accident number,which indicates that the model has some adaptability,and then predicts the number of accidents in Yuegan freeway in the future.