5G Multi-Tier Handover with Multi-Access Edge Computing: A Deep Learning Approach

Abstract

The research presented in this thesis discusses the potential enhancement of 5G multi-tier handover. This proposal will utilise two of 5G’s enabling technologies, multi-access edge computing (MEC) and machine learning (ML). MEC and ML techniques are believed to be the primary enablers for enhanced mobile broadband (eMBB) and ultra-reliable and low latency communication (URLLC). The subset of ML that was chosen for this research is deep learning (DL), as it is great at learning long-term dependencies. A variant of artificial neural networks called a long short-term memory (LSTM) network is used in conjunction with a lookup table (LUT), as part of the proposed solution. Subsequently, edge computing virtualisation methods are utilised to reduce handover latency and increase overall throughput of the network. In addition to the proposed, this thesis analyses the validity of various other potential solutions such as multi-connectivity, cloud centralised radio access networks (Cloud C-RAN) and artificial intelligence (AI).

To implement the proposed algorithm, a software simulation of a multi-tier 5G heterogeneous network is developed, based on the 3rd generation partnership project (3GPP) standards for: channel models, schedulers, and handovers. This simulator provided the tools for the author to analyse and evaluate the feasibility of the proposed solution.

The results gained from the research was promising. It showed a 40−60% improvement in overall throughput under high user densities. Although the proposed scheme may increase the number of handovers, it is effective in reducing the handover failure (HOF) and Ping-Ping rates in higher user density scenarios by 30%, and 86% respectively, compared to current state-of-the-art. In conclusion, a detailed analysis was undertaken, and the aims of the research were satisfied.