Analysis And Prediction System Of The Influence Of Meteorological Factors On Road Traffic Safety

Road traffic safety is a major issue in the world.With the increase of roads and vehicles,the frequency of traffic accidents is also increasing.This paper mainly analyses the traffic accidents caused by weather factors and other indirect factors including road surface and lighting conditions.Firstly,Based on the survey data of different weather conditions(rain,snow,etc.),different road surfaces(snow,frost/ice,etc.),and different lighting(night(no light),night(street light not lit,daylight)conditions in two Datasets(Leeds and London),and through collation and statistical analysis,this paper analyses the impact of weather conditions,road surface conditions and lighting conditions of traffic accidents.Secondly,we make a correlation analysis of road traffic accidents under specific weather conditions,road surface conditions and lighting conditions.At last,the number of traffic accidents is predicted under specific conditions.The main contributions of this paper are as follows:(1)According to analyze the correlation between the traffic accident and meteorological data,this article explores the principal component factors affecting traffic accidents.We mainly use linear regression,logistic regression and comprehensive analysis techniques to analyze the impact of various weather conditions,road surface conditions and lighting conditions on road traffic accidents.The degree of influence of weather conditions,road surface conditions and lighting conditions on road traffic accidents is obtained through analysis.(2)A road traffic accident impact factor(Road Traffic Accident Impact Factor,RTAIF)was proposed in this paper to reflect the severity of traffic accident.This paper analyzes the impact of RTAIF’s changes on road traffic accidents in all weather and analyzes the relationship between RTAIF changes and road traffic accidents under road or lighting conditions.(3)An traffic accident prediction model based on Long Short Term Memory Network is proposed to predict the number of road traffic accidents under multiple factors.The prediction model is mainly constructed by using Double Scale Equation(DS),LSTM and ARIMA,which is called DS-LSTM-ARIMA for short.In order to predict the amount of road traffic accidents in a single factor,a mixed traffic accident prediction model is constructed by combing the algorithm ELM and SVM,which is called ELM-SVM for short.The experiments show that the two models proposed have accurate predictions for the number of traffic accidents in both multi-factor and single-factor situations.(4)Finally,based on the research results above,we developed a system for analyzing the impact of meteorological conditions on road traffic safety and predicting the amount of road traffic accidents.This system is based on Python and PyQt5 for system architecture design,algorithm modular encapsulation,and user interaction interface design.