Driver-Passenger Injury Severity Analysis Based On Probit Probability Model

Traffic accidents are happening all the time,and traffic safety issues are increasingly concerned.But existing accident data are difficult to fully record accident related information,such as drivers and passengers person state of mind,vehicles instantaneous speed at the time of the accident,etc.,these factors affecting traffic accident was invisible but actually exist,the damage result will be significantly different effect on the accident injury severity degree,namely the factor of heterogeneity and randomness.Objective factors such as vehicle,road and environment are easier to obtain than human attribute factors.When analyzing human factors,drivers are often only considered and passengers are ignored.This paper mainly improves and optimizes the traditional accident injury analysis model,considering the two key human factors,driver and passenger,and from the perspective of the heterogeneity,endogeneity and randomness of the influencing factors,improves the traditional fixed parameter setting,analyzes the advantages and disadvantages of each model,and compares the simulation results and evaluation index parameters of the accident model.Discrete selection model is often used to simulate and analyze the influence degree of road traffic accident factors.In order to simultaneously analyze the factors that affect the severity degree of driver and passenger collision injury in accidents under a single modeling framework,this paper first considers The Bivariate Probit Probability Model.In addition,due to the existence of potential factors that have not been observed,The Random Parameter Model is selected to deal with factor heterogeneity,and the selected model is combined and optimized to construct The Random Bivariate Probit Probability Model(RBP)studied in this paper.The Random Bivariate Probit Probability Model(RBP)comprehensively considers the advantages and disadvantages of Random Univariate Probit Probability Model(RUP)and Bivariate Probit Probability Model(BP).After improvement and optimization,it can not only study the correlation of common factors affecting the severity of driver and passenger injury.Moreover,the effect of unobserved heterogeneity factors can be captured.The case analysis is based on 3,665 motorcycle accidents in Hunan Province from 2014 to 2016 and 10,667 taxi passenger-carrying accidents in Hong Kong from 2015 to 2017.Factors from human,vehicle,road and environment are selected as independent variables,and the severity of driver and passenger accident injury is taken as dependent variable.RUP model,BP model and RBP model were selected for simulation,and evaluation indexes were selected to compare the applicability of the three models.Goodness of fit of the two cases all showed that RBP model had the best fit.For motorcycle accidents,the age and gender of passengers have a significant impact on the severity of driver’s injury.In addition,the impact object,safety facilities,drunk driving,light,speed and other factors have a significant impact on the severity of accident.For taxi accidents,gender and age of drivers have no significant random influence,but drivers,pedestrians,vehicles and low speed limit all have significant influence on the severity of drivers’ injuries.Verify the proposed model is presented in this paper to the research of various kinds of influence factors on the motorcycle taxi accident drivers,passengers and damage extent affect the applicability of the situation,summarizes the main factors that influence the road traffic accident,is not only beneficial to understand the accident damage mechanism,can also according to the analysis conclusion put forward measures and Suggestions of reducing accidents and reduce accident severity,develop targeted traffic safety strategies.