基于车道集计交通流数据的事故风险评估分析

Application of Aggregated Lane Traffic Data from Dualloop Detector to Crash Risk Evaluation

摘要:

事故风险评估分析提取事故及对应非事故状况下的交通运行数据作为自变量，以事故发生与否作为0或1因变量，采用数理统计模型分析事故发生与交通流状态的关联性.为探索上海市城市快速路的车道集计交通流数据用于事故风险评估分析的可行性，基于条件Logistic回归，使用该数据与事故数据构建事故风险评估模型，并检验模型预测精度.结果表明低运行速度、车道间流量的差异性会增加事故发生的概率；上海市城市快速路的车道集计交通流数据可用于事故风险评估研究.

Abstract:

In crash risk evaluation study, the relevance between crash occurrence and traffic conditions was analyzed through statistical model. In the model, traffic conditions prior to each crash and corresponding noncrashes were extracted as independent variables, while 0 or 1 was taken as dependent variable for crashes or noncrashes. For the purpose of searching the validity of aggregated lane traffic data from dualloop detectors of Shanghai Urban Expressway System applied to crash risk evaluation, aggregated lane traffic data and crash data were used to develop crash risk evaluation model by condition logistic regression, and the predictive performance of the model was tested. The results show that lower speed and higher volume variation crossing lanes will increase the likelihood of crash occurrence; and the aggregated lane traffic data from dualloop detectors of Shanghai Urban Expressway System can be applied to crash risk evaluation study.