Artificial intelligence linear regression model for mobility robustness optimization algorithm in 5G cellular networks*

Sawsan Ali Saad

doi:10.1016/j.aej.2024.01.014

Artificial intelligence linear regression model for mobility robustness optimization algorithm in 5G cellular networks*

2024-02-01

5

OA

AI

DOI：10.1016/j.aej.2024.01.014

原文链接

查看原文

摘要

En 中文

Ensuring reliable and stable communication links between User Equipment (UE) and serving cellular networks during UE movement is one of the significant difficulties facing the deployment of the Fifth Generation (5G) and Sixth Generation (6G) of cellular networks. Therefore, the Handover Parameters Self-Optimization (HPSO) function has been introduced in modern cellular networks to address mobility management issues. Its main purpose to automatically optimize Handover Control Parameters (HCPs) settings. But the probability of estimating suboptimal HCP settings remains an issue, leading to a critical impact on the network performance. This results in an increase in Handover Probability (HOP), Handover Ping-Pong Probability (HPPP), and Radio Link Failure (RLF). The challenge becomes even more critical with the advent of 5G and 6G in cellular networks, owing to various factors. These factors include the extensive deployment of small base stations and the massive growth of connected devices. Despite the development of several HPSO algorithms, the existing techniques fall short of providing optimal solutions. Furthermore, the issue of suboptimal parameters remains unaddressed. This paper introduces an Artificial Intelligence Multiple Linear Regression (AI-MLR) model as an algorithm for optimizing mobility robustness in 5G cellular networks. The objective of the AI-MLR model is to automatically optimize HCP settings based on network experiences, leveraging the Instantaneous Indication Measure (IIM) Function. The AI-MLR model dynamically and instantly estimates HCP settings for UE by utilizing the IIM function. This function evaluates UE experiences through instantaneous Signal-to-Interference-plus-Noise Ratio (SINR) levels from both the target and serving base stations. Initially, the UE captures instantaneous measurements of SINR levels, serving as input parameters for the IIM function. The function then produces an output that acts as an indicator for automating the optimization process of the HCP settings. The proposed algorithm undergoes a comprehensive investigation and validation against various benchmark methods found in the literature, encompassing different mobility speed scenarios over a 5 G cellular network. This study involved the development of a simulation model using Matlab software. Performance evaluation utilized a range of Key Performance Indicators (KPIs), including HOP, HPPP, and RLF. The simulation results demonstrate that the proposed solution achieves significant improvements in terms of HOP, HPPP, and RLF under diverse movement speed scenarios, compared to the algorithms examined in the literature.

Keyword:

Handover control parameters

Mobility robustness optimization

Handover optimization algorithm

Adaptive handover parameters

Self-optimization

5G

6G

LTE-advanced system

Heterogenous networks

Artificial intelligence linear regression model for mobility robustness optimization algorithm in 5G cellular networks*

摘要

AI总结

期刊

机构

引用论文