Towards Road Profiling with Cooperative Intelligent Transport Systems

Cooperative Intelligent Transport Systems (C-ITS) used to generate large amounts of data from communications achieved through V2V (vehicle-to-vehicle) or V2I (vehicle-to-infrastructure). We have collected these data from a dedicated smartphone application and we have analyzed them. Useful information about road profiles and driver profiles have been provided. Various works have been dedicated to driver profiles since a decade but road profile was not of interest. We propose in this article to analyse data contained in the Cooperatives Awareness Messages (CAM) coming from anonymous participants to define road profiles considering various parameters like speed, location, day-time. The trajectories of individual vehicles are classified into various classes using four different algorithms: K-means, Agglomerative Clustering, DBSCAN and BIRCH. We have worked on 609 different trajectories and found different classes for each algorithm. Our contribution is to analyze the correlations between the different classes provided by each algorithm.