Identification Method And Experimental Study Of Traffic Accident Prone Location

With the rapid economic growth,the in-depth development of urbanization,the increasingly perfect road traffic infrastructure,the continuous growth of urban motor vehicle ownership,the role of road traffic in supporting and guiding economic and social development is becoming more and more obvious.The following problem is that the number of road traffic accidents remains high,and road traffic safety has increasingly become a key issue related to the safety of people's lives and property.In addition,it will also have a negative impact on the sustained growth of economic development.At present,road traffic safety has become a strategic issue related to national security in the new era.Studying traffic accidents from the perspective of geography can reveal the spatial distribution characteristics of traffic accident data,help traffic police departments to have a deeper understanding of the distribution of traffic accidents in the study area,summarize the characteristics,find out the geographical factors affecting the occurrence of traffic accidents,and control the number of traffic accidents.On the basis of summarizing the definition of traffic accident-prone location at home and abroad and the identification method of traffic accident-prone location,this paper discusses the traditional method of identifying traffic accident-prone location based on mathematical statistics.However,the accuracy and integrity of accident data are highly demanded and neglected in this method.Therefore,the spatial aggregation law of points and the way of road section division can not be unified,and the identification results have some defects such as"splitting analysis".In view of its shortcomings,this paper proposes two methods based on spatial statistics to identify the location of traffic accidents.Traffic accident itself belongs to a kind of spatial event,and the nature of individual accidents is basically independent.It can be regarded as a spatial point for analysis.Traffic accidents have the characteristics of aggregation in space,showing the distribution pattern of "accident hot spots".One kind of spatial point pattern analysis is to study the spatial distribution pattern of points by using the density and frequency of points on the basis of aggregation.In this paper,based on the two methods of plane nuclear density estimation and network nuclear density estimation,the idea and principle of the two methods are discussed,and the concrete steps of applying the two methods to traffic accident location identification analysis are given.By using the plane kernel density estimation method and the fixed window width search window,the density contribution value of each point event to each grid cell in the window is calculated,and the corresponding density estimation value is obtained by summing up the density contribution value of the point event in each grid cell search window.The smoother density estimation can be obtained by using this method.Plane representation is used to mine the spatial distribution characteristics of traffic accident-prone locations;on the basis of plane kernel density estimation method,the model is improved,which not only fully takes into account the influence of geospatial location on identification results,but also further restricts the research area to network space,taking section units as the research object,road network.As the expression of spatial distance,the shortest path distance under constraint is used to analyze the linear distribution characteristics of traffic accidents,which is more reliable and more in line with the actual situation than the results obtained by the general plane kernel density estimation method.Finally,this paper takes the actual accident data of Suzhou Industrial Park as an example to verify the research method of this paper.In this paper,the existing historical traffic accident records(2015-2016)are used to analyze the identification results using the plane nuclear density estimation method and the network nuclear density estimation method respectively.The validity of the method is verified by comparing with the results obtained by mathematical statistics,and the reliability and accuracy of the method are verified.