Analysis and classification of handovers in mobile telecommunications networks applying supervised learning algorithms

Abstract

The constant development of wireless network communications is transforming modern society, introducing new forms of interactivity. The 5G/New Radio (NR) network has enabled unprecedented levels of engagement, combining high transfer rates with a significant expansion in coverage area. However, it has also raised concerns about security and network transition procedures, commonly known as handovers. This paper proposes using machine learning in 5G mobile networks, specifically employing the Logistic Regression algorithm to predict handovers. Additionally, it examines a Dual Connectivity urban scenario between 5G/NR and 4G/Long Term Evolution (LTE), considering criteria such as Received Signal Strength Indicator (RSSI), Reference Signal Received Power (RSRP), distance, and Signal-to-Interference-Plus-Noise Ratio (SINR) for handover prediction using the K-Nearest Neighbor (KNN) algorithm. The primary goal of this study is to reduce the number of handovers in both 5G and 4G networks through predictions made by KNN and Logistic Regression. This implementation demonstrates the proposal’s feasibility, its impact on network performance, and an analysis of the relevant results.