New traffic models considering complex geometries and data

传统的交通安全分析方法--基于历史事故数据--受到限制，即必须发生足够数量的事故才能识别事故黑点并采取适当措施。因此，传统方法的预测可能性是有限的，应由其他可能性加以扩展，如基于模型的预测，或由带有安全指标的道路安全分析所取代。交通模型提供了基于模型预测路段上的事故数量的可能性。虽然高速公路路段的交通模型已经很先进，但由于驾驶行为的变化而风险更大的更复杂的几何形状(例如交叉口)的模型还没有同样的可用的程度。因此，本项目的目标是将以前关于高速公路路段交通行为的结果和数据转移到复杂的高速公路交叉口几何形状。新开发的模型和数值方法将与迄今已有的结果相结合，并辅之以随机方法(不确定性)，以获得整个高速公路的多尺度随机交通预测模型。此外，将安全指标传递给交织交通，并辅之以不确定的参数。

The traditional approach of traffic safety analysis – based on historical accident data – is subject to the restriction that a sufficient number of accidents must have occurred in order to be able to identify accident black spots and take appropriate measures. The predictive possibilities of the traditional approach are therefore limited and should be extended by other possibilities such as model-based predictions or replaced by road safety analyses with safety indicators. Traffic modelling offers the possibility of predicting the number of accidents on a road section based on models. While traffic models of motorway sections are well advanced, models for more complex geometries (e.g. junctions), which are more risky due to changed driving behaviour, are not yet available to the same extent.The goal of this project is therefore to transfer the previous results and data on traffic behaviour on motorway sections to complex geometries of motorway junctions. The newly developed models and numerical methods will be combined with the results obtained to date and supplemented by stochastic approaches (uncertainty) in order to obtain a multi-scale stochastic traffic forecast model for the entire motorway. In addition, the safety indicators will be transferred to weaving traffic and complemented by uncertain parameters.