Hazard Analysis Method Of Highway Based On Case Base

The continuous improvement of the highway network has provided tremendous help and convenience for economic development,but the large number of safety problems it has caused cannot be ignored.In the process of highway operation,safety analysis and hazard analysis are an important means to improve the level of highway safety management.It is of great practical significance to accurately analyze the hazard of highway for accident prevention.The frequent occurrence of highway accidents has accumulated a large amount of accident data and information,which provides strong support for studying and predicting the relationship between the severity of accident consequences and influencing factors.In recent years,the continuous development of information technology such as geographic information technology(GIS),database technology,and machine learning technology has provided new ideas for the study of highway safety issues.This article summarizes the existing safety analysis and assessment methods,applies GIS technology,database technology,machine learning technology and knowledge base technology to the study of highway safety issues,and designs and develops highway hazard analysis based on the case library.system.The main work of this paper is as follows:(1)This paper collects a large number of historical accident case data and research literature on highway safety at home and abroad,summarizes the research methods in the literature,and puts forward a highway hazard analysis method based on case base.(2)Based on the analysis of the causes of highway accidents,combined with the domestic statistical data chart of highway accidents,this paper expounds the influence of drivers,vehicles,roads and environment on the accident consequences;Then,the factors of the accident case are defined,layered and expressed;Finally,based on My SQL database technology,the fields of influencing factors and consequence severity are designed,the Expressway accident case database is established,and the hierarchical storage,management and transfer of accident case data are realized.(3)Based on the highway accident case base and combined with the characteristics of highway accidents,the accident association model is established.Apriori algorithm is selected as the inference engine to mine the accident data.1387 association rules are deduced and assembled into a knowledge base which can be used for hazard analysis.Some of the knowledge is explained.(4)Through the comparison of nine algorithms,Random Forest is determined as the main algorithm,decision tree and naive Bayes as the control algorithm for data learning and hazard analysis.With this method as the core,a highway hazard analysis system based on Arc GIS is designed and developed.The system can realize the management of accident case base,display of accident point distribution,analysis of accident attribute chart and highway hazard analysis.(5)Taking 3500 accident cases of Wenzhou section of Ningbo Taizhou Wenzhou expressway as an example,11 factors that have a significant impact on the severity of Expressway accident are obtained as key factors through machine learning,among which the first five key factors are vehicle condition(big truck),vehicle condition(Minibus),line condition(straight),Driving behavior(not keeping a safe distance),and traffic conditions(Level 4);finally,the accidents under the influence of 11 key factors(Only property loss accidents,injury accidents and fatal accidents)occurrence situation was analyzed.