Unsupervised Driving Profile Detection Using Cooperative Vehicles' Data

摘要

C-ITS (Cooperative Intelligent Transport Systems) provide nowadays a very huge amounts of data either from vehicles, roadside units, operator servers or smart-phone applications. Data need to be exploited and analyzed. In this paper, we first study the communication logs containing network messages emitted by the vehicles and the infrastructures when they communicate. We used these logs to measure the latency and evaluate if it is consistent with data analysis. Then, we try to detect driving profile using unsupervised machine learning approaches. Results both in terms of latency and of driving profile detection reveal promising issues in this new area.