Scholarlink

<p>Modern 5G New Radio (NR) networks require accurate positioning, which is essential to provide applications in emergency response, urban planning, and resource management. Nevertheless, traditional technologies like the Global Positioning System (GPS) and Assistant-GPS (A-GPS) cannot work in dense urban and indoor areas because they cannot see signals, and power consumption is too high. Conversely, other positioning methods, including Time of Arrival (TOA), Time Difference of Arrival (TDOA), and Enhanced Cell ID (E-CID), are limited in accuracy and prone to multipath interference. This paper introduces a new User Equipment (UE) positioning solution based on signal-level measurements and the Timing Advance (TA) parameter in 5G networks. The algorithm improves localization accuracy, and no additional hardware costs are required because the Machine Learning (ML) model is combined with features such as signal strength, signal quality, SINR, TA, and throughput. The statistical results indicate that localization accuracy in a complex setting with signal-level data and TA measurements, complemented by predictive modeling, can be notably enhanced by 33-55 percent relative to the baseline approach. Moreover, integrating ML-based position refines model accuracy better and optimally, achieving improvement range from 98% to 99%, offering a scalable and effective solution for modern 5G positioning challenges.</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

Predictive and Machine Learning-Based Approaches for UE Positioning in 5G NR Networks

  • 期刊论文

  • 发表日期：

    2026年1月9日

查看全文
  • Zaenab Shakir,
  • Abbas Al-Thaedan,
  • Josko Zec,
  • Monera Salah,
  • Fitzroy Nembhard

摘要

Modern 5G New Radio (NR) networks require accurate positioning, which is essential to provide applications in emergency response, urban planning, and resource management. Nevertheless, traditional technologies like the Global Positioning System (GPS) and Assistant-GPS (A-GPS) cannot work in dense urban and indoor areas because they cannot see signals, and power consumption is too high. Conversely, other positioning methods, including Time of Arrival (TOA), Time Difference of Arrival (TDOA), and Enhanced Cell ID (E-CID), are limited in accuracy and prone to multipath interference. This paper introduces a new User Equipment (UE) positioning solution based on signal-level measurements and the Timing Advance (TA) parameter in 5G networks. The algorithm improves localization accuracy, and no additional hardware costs are required because the Machine Learning (ML) model is combined with features such as signal strength, signal quality, SINR, TA, and throughput. The statistical results indicate that localization accuracy in a complex setting with signal-level data and TA measurements, complemented by predictive modeling, can be notably enhanced by 33-55 percent relative to the baseline approach. Moreover, integrating ML-based position refines model accuracy better and optimally, achieving improvement range from 98% to 99%, offering a scalable and effective solution for modern 5G positioning challenges.