Analysis Of The Effect Of Road Lighting On Vehicle Characteristics At Intersections

With the rapid development of the economic,the number of vehicles in China shows a trend of steady increase,which leads to the increase of security risks on road.Person,vehicle,road and environment would cause different levels of traffic accidents.Studies showed that bad night lighting is an important reason for traffic accident at night,so that nighttime traffic accident accounts for a larger proportion of the total number of traffic accidents.Studies also showed that traffic accidents happened more frequently in urban road intersections because of more conflicts at the intersections,bigger blind areas and relatively more complex management,when compared to other traffic conditions.Therefore,this study is dedicated to explore the influence of lighting on intersection safety at night.Data collected by questionnaire survey and by the intersection of unmanned aerial vehicle(UAV)was used in this study in order to analyze the driver’s subjective feelings and recognition of vehicle driving state.In addition,multiple choice models and artificial neural network model were used to explore the influence of illumination on the driving characteristics of the driver through the intersection is analyzed.Questionnaire analysis was used in this study to explore the driver’s subjective feelings,mainly including personal basic information,driving information,driving behavior intention and intention to limit driving.In the third part of the questionnaire,a five-level likert scale design was used to complete the scale of data reliability analysis and factor analysis.The purpose of factor analysis was to achieve the factors of drivers’ choice tendency under different lighting conditions by compositing eight variables instead of the original using software stata to deal with the variable data analysis with the multinomial logit model.Then,the driver’s choice tendency was reified under different lighting conditions.UAV aerial vehicles were also used in this study to collect the speed of signal control intersection point data Software tracker was used to extract the video data In the process of collecting data,a laser range finder was used to measure the information such as street light distance and the street lamp spacing at the intersection.Matlab was then used to calculate video speed data and plot distance scatterplots,which could reflect the velocity distribution characteristics of intersection point;By comparing the average speed of the whole road adjacent section,mainly between the adjacent two section two independent sample mean,the average cumulative distribution deviation as well as the percentage change were used to react the vehicle speed driving laws.Illuminance data at the intersections were collected by the illuminance meter.Finally,speed variation was proposed to represent the change in pilot speed after the intersection.An artificial neural network model was established with explanatory variables like speed,the total number of traffic volume,straight vehicles,street lamps and street lamp spacing distance from sidewalks,and dependent variable like speed variation.The model is solved by the inverse propagation algorithm and obtained through training.The MSE is used as the evaluation criterion for the convergence of the model.Result of model fitting effect is good after the completion of training.Output of the model also showed the degree of influence on speed variation size of the five variables.The intensity of effect of illuminance on the speed variation curve was then analyzed by variable control method.Results provide suggestions for management department related to intersection safety.