Scholarlink

<p>As 5G networks continue to expand, one challenge stands out delivering fast, reliable internet through ultra-dense deployments while balancing coverage and capacity. Self-Organizing Networks (SON) offer a practical solution by automating key network tasks configuration, optimization, healing, and protection. This study takes a close look at how existing self-optimization techniques, particularly those powered by machine learning, can improve 5G coverage and capacity. We dig into the root causes of poor network performance, from coverage holes to signal overshooting and undershooting, and examine how SON-based strategies tackle these problems. We also look at how Key Performance Indicators (KPIs) shape optimization decisions across ML approaches from supervised and unsupervised learning to reinforcement learning. Beyond reviewing what works, this survey points out where current research falls short and highlights gaps that still need attention. This survey paper aims to provide an informative resource for both industry and the research field regarding the importance of coverage and capacity integration for the ML-driven SON system for the next generation. Moreover, this survey investigates the limitations of the current studies and outlines unaddressed research areas that require additional investigation. This survey paper aims to provide an informative resource for both industry and the research field regarding the importance of coverage and capacity integration for the ML-driven SON system for the next generation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A comprehensive review of machine learning-driven self-organizing networks for 5G coverage and capacity

  • 期刊论文

  • 发表日期：

    2026年2月25日

查看全文
  • Sharmin Sharmin,
  • Ismail Ahmedy,
  • Bryan Raj Peter Jabaraj,
  • Muhammad Umair Munir,
  • Rafidah Md Noor,
  • Muneer Ahmad

摘要

As 5G networks continue to expand, one challenge stands out delivering fast, reliable internet through ultra-dense deployments while balancing coverage and capacity. Self-Organizing Networks (SON) offer a practical solution by automating key network tasks configuration, optimization, healing, and protection. This study takes a close look at how existing self-optimization techniques, particularly those powered by machine learning, can improve 5G coverage and capacity. We dig into the root causes of poor network performance, from coverage holes to signal overshooting and undershooting, and examine how SON-based strategies tackle these problems. We also look at how Key Performance Indicators (KPIs) shape optimization decisions across ML approaches from supervised and unsupervised learning to reinforcement learning. Beyond reviewing what works, this survey points out where current research falls short and highlights gaps that still need attention. This survey paper aims to provide an informative resource for both industry and the research field regarding the importance of coverage and capacity integration for the ML-driven SON system for the next generation. Moreover, this survey investigates the limitations of the current studies and outlines unaddressed research areas that require additional investigation. This survey paper aims to provide an informative resource for both industry and the research field regarding the importance of coverage and capacity integration for the ML-driven SON system for the next generation.