The accident rate in urban area intersections: the use of neural artificial net to improve the road safety

The paper presents a research which tackles two issues. The first involves the study of different types of survey reports with reference to road accidents in urban areas: the aim is the identification of characteristics to make the survey reports more functional. The second one consists of defining the functional relationship between some variables which the hazard level of an intersection depends upon. Particularly, accident data are used to check, with a neural artificial net model, what type of dependence links the accident rate in urban areas to the traffic flows, the geometric features of the intersection and the environmental characteristics (road paving, weather conditions…). The pursued objectives contemplate a safety increase in the flow of traffic at urban intersections. An improved survey report layout implies a higher level of reliability of the data processing, and the use of neural nets allows a fast identification of the most suitable technical solutions to improve the safety of the intersections.