The article presents a method to predict the manoeuvres of vehicles at an urban intersection. The core intelli- gence of this cooperative system is located in the traffic light controller. In order to estimate manoeuvres, the algorithms use probe vehicle data, the current and future status of the traffic light control and data provided by infrastructure-based sensors. A prototype of the system was tested and validated first using data provided by a microscopic traffic simulation and second at a real intersection with equipped vehicle transmitting probe data. In the research project SAFESPOT, this approach was integrated successfully as a core component of the application IRIS to identify safety critical situations before they occur and to warn the impacted driver.     «

The article presents a method to predict the manoeuvres of vehicles at an urban intersection. The core intelli- gence of this cooperative system is located in the traffic light controller. In order to estimate manoeuvres, the algorithms use probe vehicle data, the current and future status of the traffic light control and data provided by infrastructure-based sensors. A prototype of the system was tested and validated first using data provided by a microscopic traffic simulation and second a...     »