| Road traffic accident has the enormous impact on economic society and people's lives, especially the number of motor vehicle is booming in recent years. With the increasingly outstanding contradiction between traffic demand and supply, traffic accident, traffic congestion and traffic pollution become chronic illness of contemporary society, and even more worse. The existing traffic accident analysis theories stay in the static analysis levels and is difficult to objectively reflect the traffic accident process, trends and influencing factors, with the little regard to the variable changes as well as interaction and mutual restriction in the process of changes in these variables. To improve the effectiveness of traffic accident prevention measures, road traffic accident dynamic analysis is particularly important.With the deeply absorbing domestic and overseas research of traffic accident analysis and warning methods, through the main line of dynamic analysis, urban arterial network and trunk highway network is respectively analyzed from the angle of road network. Furthermore, by using macro and micro analysis methods, traffic accident dynamic analysis and warning issues are systematic studied in this paper. The purpose of the research on these theories and methods, is to provide some reference and basis for traffic accident analysis, warning and safety improvement.Firstly, based on the definition of road network, combined with economics and statistics theory, the dynamic analysis indicators and methods were proposed, the differences and connections of dynamic and static analysis were analyzed. Traffic accident analysis methods and hierarchy were defined through analyzing the characteristics of traffic accident. Considering the dynamic analysis and traffic accident analysis principle, the objects and contents of Traffic Accident Dynamic Analysis (TADA) were proposed, and the framework of TADA was constructed.Through classified TADA indicators into level indicators, speed indicators and change indicators, the definitions were proposed, and the dynamic indicators applied to Traffic Accident Distribution Feature (TADF) study were put forward. The TADF of absolute and average distribution were dynamic analyzed, and the space distribution features of overall and peak traffic accidents were studied based on GIS platform, through constructing traffic accident time series. By designing quantitative indicators of traffic accident form distribution, the dynamic analysis indicators were applied for TADF, and the associated distribution was also analyzed.On the basis of analyzing objective levels of traffic accident causes and macro causes constitution, the Level of Sensitivity (LOS) method was put forwad, significant macro causes were dynamic analyzed by the method. In connection with micro causes of single accident and accident spots, the Fault Tree Analysis (FTA) and Bayesian Network (BN) was advanced respectively. The quantitative analysis indicators of direct causes and related causes were proposed, and the quantitative analysis model of related causes were constructed, changes and features of traffic accident causes were analyzed. With the basis of historical accident data time series, grey Verhulst dynamic analysis model of single factor and regression analysis model of multiple factors were established, the grey Verhulst dynamic analysis model was selected for analyzing traffic accident death toll trends. BP neural network model was established for road sections and intersections based on traffic accident trends analysis, the spatial distribution features were achieved based on GIS platform. In the perspective of the potential accidents risk of road network, the Accident Probability Prediction (APP) model of traffic accident was established based on empirical Bayes, the trends of traffic accident were dynamic analyzed by using APP indicator.In order to apply the researchs of TADA, traffic accident warning indicators system and evaluation range of road network were constructed, which from the aspects of traffic accident distribution features, traffic accident causes, traffic accident trends of road network. In connection with warning indicators, two-level fuzzy comprehensive evaluation model was applied to judge the level of traffic accident warning. The process of evaluation, triggering, identification and disposal was also put forwad. At last, using the data of Heilongjiang trunk highway network and the G301 national highway, traffic accident macro and micro warning were applied, and the warning signal dynamic change figures were created.
|
PDF Full Text Request