Cellular communications between vehicles and everything(C-V2X): integration and first prototype

C-V2X is a technology introduced by 3gpp 4G standards and designed to allow vehicles to communicate with everything, i.e., other vehicles (V2V), pedestrians (V2P), infrastructure (V2I), and the cloud network (V2N). The technology uses cellular network infrastructure that is already providing service to UEs, with architectural enhancement. C-V2X defines two complementary transmission modes: Direct communication and Network-based communication. Direct communication, necessary for V2V and V2P, is based on the LTE device-to-device communication over the so-called PC5 interface and it can operate both under and out of the coverage of base stations. All other communication modes, instead, leverage the cellular network infrastructure (i.e., the LTE-Uu interface), and require a base station (i.e., an eNB) as inter mediator to communicate with an application server in the mobileoperator network or in the cloud network.V2X communication is intended to improve road and passenger safety, traffic efficiency and infotainment services. C-V2X technology enhances safety providing information even beyond the driver's line of sight to notify and warn drivers of possible hazards and of oncoming vehicles. Ultimately, this feature will play a vital role in supporting fully autonomous vehicle deployment. Road safety is one of the use cases in Intelligent Transport System (ITS) among warning applications (i.e, collision risk and others).In my thesis, I collaborated to develop a working first prototype of a collision avoidance application implementing intersection warning. The prototype is designed to operate considering vehicles and pedestrians (UEs), whose mobility is simulated using sumo, a well-established mobility simulator. The system relies on Cooperative Awareness Messages(CAMs), which are beacons sent by UEs and Decentralized Environment Notification Messages (DENMs), which are alerts sent by the application server. According to ETSI standards, cams are transmitted with a minimum frequency of 1hz and maximum of 10hz. In order to detect a potential crash accident, the prototype uses state-of-the-art algorithms based on the trajectory of UEs. The collision avoidance algorithm resides on an application server and it is run when, after receiving a cam, it is determined that the sender of the message is on a collision course with another node. The collision detection algorithm requires as input the position and speed (including heading) of the current vehicle, as well as the latest cam sent by each vehicle in the scenario. The collision avoidance algorithm can distinguish between cams sent by pedestrians and cams sent by vehicles. This gives us a double advantage. First, when the server receivesa cam from a vehicle, it looks for possible collisions with both cars and pedestrians, while on the contrary, with a message sent from a pedestrian, the algorithm skips the analysis for pedestrian-with-pedestrian collisions. The second advantage involves the possibility to set different parameters for the collision detection algorithm, according to the type of entity which sent the cam. This allows a better performance of the algorithm, in terms of false positives and false negatives. The experimental study features a realistic scenario including both human driver and autonomous vehicles in an urban environment and it tries to establish the effectiveness of the whole system in preventing collisions between vehicles and between vehicles and pedestrians.