DESIGN AND ASSESSMENT OF A CLOUD-BASED REAL-TIME GEO-REFERENCED VEHICULAR NOTIFICATION SYSTEM

The recent need of a fully autonomous car has pushed researchers into developing applications, services and technologies capable of connecting the vehicle to its surroundings. A real-time geo-referenced notification system coupled with aggregation of data and optimization of fog-computations in the edge network allows for a fast and reliable traffic orchestration in critical areas. Modern literature showcases the efficiency and responsiveness of cellular vehicle-to-network in areas like collision avoidance and traffic control. This thesis aims at developing an architecture capable of aggregating and disseminating real-time notification messages through a multi-MAC infrastructure. Furthermore, an assessment on its efficiency and responsiveness is carried out through several metrics. After a quick overview on the state of the art, the overall architecture is discussed. This features a brief overview on the MEC infrastructure that recognizes and elaborates information on the MEC node. The focus of the analysis revolves around the cloud architecture algorithms and the message flows traveling from the road to the users. Due to stringent requirements, experimental data are collected on the end-to-end latency; this serves as a benchmark for evaluating the reliability and effectiveness of such a solution. The collected data are then aggregated in a live Citizen Application that displays information about vehicles and alerts on the critical zones where possible hazards have been detected.