A behavioural model for the level crossing collision risk assessment

Level crossings (LC) are identified as being critical security points in both road and rail infrastructures. Statistics show that more than three hundred people are killed every year in Europe in more than 1200 accidents occurring at level crossings. In this paper, we propose and apply a systematic method to establish a global model involving both rail and road traffic in the level crossing area. The model building approach proposes a progressive integration of elementary models, each of which describing the behaviour of a part in the whole LC environment. We are interested more precisely in a particular phenomenon which may cause collisions at level crossings, and which corresponds to the accumulation of vehicle waiting queues on the LC exit zone. As a model, we use Petri Nets (PN). For the sake of precision and in order to obtain a realistic and fine behavioural model, quantitative aspects are integrated in the form of functions (stochastic distributions, deterministic values, probabilities, etc.) associated to the transitions of our PN model Stochastic PN).