Modelling, Analysis and Optimisation of Ultra-Dense Networks in Future Smart Cities

点击翻译按钮获取中文摘要

The objective of the project is to develop a generalised analytical framework based on novel multivariate statistical theory that can be used to study ultra-dense modern wireless communication paradigms, including (but not limited to) fog-RAN, large-scale distributed antenna systems and dynamical systems that describe information propagation through networks. Following future 5G paradigms, the study is focused on a network where small-cells are massively deployed and are allowed to store a limited set of contents in a local cache. Traditionally, a macroscopic approach for analysing such networks have been used wherein the users and/or remote radio heads are randomly deployed, often following a homogeneous Poisson distribution such that well established stochastic geometry theory can be applied. However, it is more likely to experience spatially non-homogeneous content requests generation due to urban physical impendence such as building, road or walking pavement pattern. Our work is focused therefore on finding the best network set up in terms spatial deployment and content caching probabilistic model of small-cells. As result of our study, we will be able to both increase the success content delivery probability, saving energy (by triggering an idle state for some access points) and alleviating the interference saturation of the whole network (due to a massive deployment of nodes). The study stands therefore as an investigation of the limits in massively deploying transmitting nodes in future 5G mobile networks, against the generally shared view of remarkably increase the number of small cells across a network. This study will provide valuable insight on the network scaling laws under different communication technologies and elegantly quantify the trade-off between diversity and interference as the network density increases. Relevance to EPSRC thematic areas: "RF and microwave communications", "Digital signal processing", "Artificial intelligence technologies", "Future Intelligent Technologies".