An Artificial Intelligence-based Handover Triggering and Management Mechanism for 5G Ultra-dense Networks to Improve Handover Authentication

Abstract

The emergence of 5G ultra-dense networks has gained considerable attention, as solutions of this kind enable rapid and intelligent device connectivity, thus ushering in a new era of high-speed communications. Nevertheless, the process of managing mobility across varying inter-frequency strategies increases interference and complexity. The development of a reliable handover algorithm is crucial for high-quality service, especially in ultra-dense networks with small cells. However, frequent handovers, ping-pong effects, and load-balancing issues arise due to the random and dense deployment of small cells. Additionally, ensuring secure and smooth handover authentication is critical, due to an increased risk of frequent transitions of users across different networks. In such a context, this research focuses on triggering handovers and managing 5G mobile networks, all while protecting sensitive data. We introduce an artificial intelligence-based approach aimed at improving handover initiation and management processes, leveraging Boruta random forest optimization (BRFO) to fine-tune handover margins and identify optimal trigger points for handovers. In addition, an impulsive graph neural network (IGNN) is utilized as a decision framework to predict and minimize unnecessary handovers, thus improving stability in small cell environments. Simulation results demonstrate that the proposed methodology significantly enhances handover management, strengthens authentication, and effectively mitigates a variety of attacks in 5G ultra-dense networks.