How machine learning can help resolve mobility constraints in D2D communications

Device-to-Device (D2D) communications are based on geographic proximity. This chapter introduces D2D communications and research on future mobile networks. The centralized approach involves one or more network entities in the discovery process. As the network operators have a broader view of global traffic and the context of UE mobility, centralized discovery approaches aim to make use of the resources available to mobile operator networks on micro- and macromobility, in order to supply more precise information for detection. To predict user density in real time and thus enhance the performance of the discovery process for dynamic environments, a new approach is proposed, which is based on machine learning tools: Support Vector Machines and Support Vector Regression. The chapter then examines machine learning (ML) techniques and the applications of ML. It describes the approach that is proposed and presents the experimental results obtained.