论文部分内容阅读
以哈尔滨市干道路网为研究对象,收集到了该路网上468个路段和163个平面交叉口的道路交通数据,以及1999年至2004年所发生的8510起交通事故数据。分析了事故数据的统计分布特性,应用聚类分析技术确定了路段和交叉口的类别,并在此基础上分别建立了事故总体和分事故形态的预测模型。论文探讨了高峰时段的事故次数、事故率与路段v/c之间的定量关系。标定出了24个模型,并形成干道系统事故预测模型库。最后,运用所建立的事故预测模型选取了2010年哈尔滨规划路网的一部分进行实例分析,结果表明了预测模型是有效的。
Taking the dry road network of Harbin as the research object, the road traffic data of 468 intersections and 163 intersections on the road were collected, and 8,510 traffic accident data from 1999 to 2004 were collected. The statistical distribution characteristics of accident data are analyzed. The clustering analysis technique is used to determine the types of road sections and intersections. On the basis of this, the prediction models of overall accident and sub accident are established respectively. The paper discusses the number of accidents during peak hours, the accident rate and the quantitative relationship between road sections v / c. Twenty-four models were calibrated, and a database of trunk system accident prediction models was formed. Finally, using the accident prediction model established to select a part of the road network in Harbin in 2010 for an example analysis, the results show that the prediction model is effective.