Early detection traffic flow incidents model on road networks

In this article, an early detection incident model is proposed. This model is based on a statistical joint study of traffic density and the level of services defined for the basic freeway segments. This joint study has been calculated through a real traffic data set collected by a sensor located on a Spanish freeway stretch where a lot of high level congestion situations happen. The results achieved with this model have been compared with the incidents posted on the website of the Spanish Traffic Department. The global performance of the proposed model shows a favorable outcome.