Automatic Detection of Radio Coverage and System Capacity Map for Edge Resource Federation

Network Slicing is one of the most promising possibilities that will be provided by 5G networks, allowing customer-specific services to run on a heterogeneous set of physical resources. In 5G, Bells Labs is aiming at the possibility of federating both 3GPP and non-3GPP resources, to create the infrastructure suited for Network Slicing. These resources must be organized and their coverage evaluated, solving two aspects: the problem of deduction (prediction of coverage interaction between resources starting from the relative positioning) and the problem of detection (estimation of resources’ signal quality). During the internship, a graph-based database has been built exploiting internal and external sources. Several techniques for both deduction and detection have been studied, among which the INF-score-based algorithm has been selected for deduction and the coverage radius estimation technique has been used to perform detection. A first baseline pipeline has been built to solve the coverage evaluation problem, showing promising results.