Comprehensive Evaluation and Bottleneck Problem Analysis for the Traffic Safety of Freeway
|School||East China Jiaotong University|
|Course||Transportation Planning and Management|
|Keywords||Highway Traffic Safety Comprehensive Evaluation Support Vector Machine (SVM) Accident-prone points Bottlenecks|
With the growing highway mileage of highway traffic safety situation is worsening traffic fatality rate highest in the world over the past decade. Highway accidents occur frequently not only a serious threat to the safety of people's lives and property, but also severely limits the normal functioning of the highway performance. Therefore, the scientific evaluation of highway traffic safety, analyze the bottleneck of traffic safety, as well as looking for traffic safety improvement strategy is urgent of the important issues. This article is based on highway traffic safety, transport facilities and traffic conditions of the site investigation, comprehensive traffic safety research, the use of AHP, matter-element evaluation method, support vector machine (SVM) and factor analysis Highway Traffic Safety Comprehensive Evaluation and bottlenecks analysis. Mainly from the following aspects: First, based on the status of highway traffic safety analysis, five people, vehicles, road traffic environment and traffic management from the point of view of the traffic engineering Highway Traffic Safety influencing factors and their interactions are discussed . Secondly, highway traffic safety impact factors, based on the evaluation criterion for screening and grading standards, the establishment of comprehensive evaluation index system Highway Traffic Safety. Analytic Hierarchy Process to determine the index weight, matter-element evaluation method to select learning samples to construct a based on SVM highway traffic safety comprehensive evaluation model. And the use of partial binary tree SVM multi-classification algorithm based on matrix similarity parameters selected method, with the help of MATLAB to solve the SVM comprehensive evaluation model the learning sample testing and inspection, verification SVM comprehensive evaluation of the effectiveness of the model. Again, the combination of the statistical characteristics of the frequency of accidents, explore suitable for highway traffic accident prone locations to identify the cumulative frequency curve method. And then identify the accident more than fat point, analysis triggered the accident-prone points all aspects of factors, the use of factor analysis to the accident and more hair Causes of points is divided into the major causes and secondary causes, identify bottlenecks the crux of the problem is located, and according to the bottleneck governance principles, on Highway Traffic Safety bottlenecks management strategies. Finally, an empirical study of the Shanghai-Kunming expressway (Chang Zhang segment) traffic safety. Analysis of the overall situation of the Chang Zhang traffic safety, mining Chang Zhang traffic accident distribution characteristics, the use of the overall evaluation based on the comprehensive evaluation model of SVM Highway Traffic Safety Chang Zhang traffic safety, Chang Zhang accident cumulative frequency curve method to identify multiple Analysis of accident-prone Causes of points to locate the bottleneck problem, combined with the comprehensive evaluation and bottlenecks analysis and factor analysis results of Chang Zhang Traffic Safety Improvement Strategy. Empirical studies have shown the the built SVM comprehensive evaluation model and selection bottleneck analysis method feasibility and practicality, so as to establish a people, vehicles, roads, traffic environment, traffic management to achieve the theoretical framework of coordination, improve highway the level of traffic safety provide a theoretical basis.