A MEC-based Extended Virtual Sensing for Automotive Services

Multi-access Edge Computing (MEC) promises to enable low-latency applications and to reduce the impact of edge service traffic on the core network. Leveraging on the extension of the popular OpenAir Interface (OAI) architecture to include MEC functionalities, in this paper we show the impact of edge computing resources on a crucial vertical domain, i.e., the automotive domain. As a key example, we focus on a relevant class of automotive services, namely, the Extended Virtual Sensing (EVS) services. With EVS, the network infrastructure collects and makes available measurements gathered by sensors aboard vehicles, as well as by smart city sensors, to improve road safety and passengers/driver comfort. Specifically, we select the EVS application that extends the vehicle sensing capability for supporting vehicle collision avoidance at intersections, and we describe its implementation within the OAI MEC platform. We evaluate the performance of the designed solution emulating the Cooperative Awareness Messages (CAMs) of several vehicles, using a Software Defined Radio (SDR) equipment. We then show experimentally that the MEC infrastructure is pivotal to meeting low-latency requirements and allows detecting all collisions between vehicles, thus proving to be of great benefit to the support of critical automotive services.