Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

5G Positioning Advancements with AI/ML

Mohammad Alawieh, Georgios Kontes

TL;DR

The paper surveys AI/ML-based direct positioning in 5G NR, focusing on challenging NLOS/complex LOS scenarios and building on TR38.843 to define an AI/ML Life Cycle Management (LCM) framework for training, deployment, inference, monitoring, and data handling. It synthesizes TR38.843 evaluations showing sub-meter accuracy in many conditions but highlights degradation under strong NLOS, synchronization errors, and SNR mismatches, while emphasizing the value of time-domain inputs such as CIR/PDP/DP. It also discusses countermeasures including optimized measurement reporting, data collection strategies, and proactive model/functionality switching to balance performance with signaling and data costs. Overall, AI/ML-based direct positioning has strong potential as a future 3GPP-enabled positioning solution, provided robust LCM, data strategies, and signaling models are established.

Abstract

This paper provides a comprehensive review of AI/ML-based direct positioning within 5G systems, focusing on its potential in challenging scenarios and conditions where conventional methods often fall short. Building upon the insights from the technical report TR38.843, we examine the Life Cycle Management (LCM) with a focus on to the aspects associated direct positioning process. We highlight significant simulation results and key observations from the report on the direct positioning under the various challenging conditions. Additionally, we discuss selected solutions that address measurement reporting, data collection, and model management, emphasizing their importance for advancing direct positioning.

5G Positioning Advancements with AI/ML

TL;DR

The paper surveys AI/ML-based direct positioning in 5G NR, focusing on challenging NLOS/complex LOS scenarios and building on TR38.843 to define an AI/ML Life Cycle Management (LCM) framework for training, deployment, inference, monitoring, and data handling. It synthesizes TR38.843 evaluations showing sub-meter accuracy in many conditions but highlights degradation under strong NLOS, synchronization errors, and SNR mismatches, while emphasizing the value of time-domain inputs such as CIR/PDP/DP. It also discusses countermeasures including optimized measurement reporting, data collection strategies, and proactive model/functionality switching to balance performance with signaling and data costs. Overall, AI/ML-based direct positioning has strong potential as a future 3GPP-enabled positioning solution, provided robust LCM, data strategies, and signaling models are established.

Abstract

This paper provides a comprehensive review of AI/ML-based direct positioning within 5G systems, focusing on its potential in challenging scenarios and conditions where conventional methods often fall short. Building upon the insights from the technical report TR38.843, we examine the Life Cycle Management (LCM) with a focus on to the aspects associated direct positioning process. We highlight significant simulation results and key observations from the report on the direct positioning under the various challenging conditions. Additionally, we discuss selected solutions that address measurement reporting, data collection, and model management, emphasizing their importance for advancing direct positioning.
Paper Structure (27 sections, 7 figures, 1 table)

This paper contains 27 sections, 7 figures, 1 table.

Table of Contents

  1. Introduction
  2. Direct Positioning Requirements and Anticipated Challenges
  3. AI/ML model life cycle management
  4. Model management complexity
  5. Data Management
  6. Model training
  7. Model validation
  8. Model Deployment
  9. Model Inference
  10. Model Monitoring and Management
  11. TR38.843 Direct Positioning Evaluations
  12. Evaluations assumptions
  13. Observations on AI/ML Model Performance
  14. Multisource Accuracy Evaluations
  15. Time-Domain Model Inputs
  16. ...and 12 more sections

Figures (7)

  • Figure 1: Example realizations on AI/ML direct and assisted positioning.
  • Figure 2: Illustrative representation of a hybrid indoor factory environment, emphasizing different reception conditions across different zones (UE-A to UE-D).
  • Figure 3: AI/ML direct positioning model life cycle management, as discussed within tr38843 and adapted from ashmore2021assuring.
  • Figure 4: Overview of the semi-supervised learning setup studied within 3GPP tr38843.
  • Figure 5: Results from multiple sources in tr38843 for direct positioning evaluations, excluding generalization aspects and without label errors.
  • ...and 2 more figures