Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Enhancements for 5G NR PRACH Reception: An AI/ML Approach

Rohit Singh, Anil Kumar Yerrapragada, Jeeva Keshav S, Radha Krishna Ganti

TL;DR

The paper addresses robust RAPID detection and TA estimation for 5G NR PRACH, where correlation-based methods struggle in multipath and low SNR. It introduces two parallel ResNet-like classifiers that process raw frequency-domain OFDM data to classify RAPID among $10$ values and TA among $12$ delay steps, leveraging the phase rotations from RAPID ($C_v$ with $L_{RA}=139$) and TA. Evaluations on simulated TDLC channels and OTA hardware show that the NN-based receiver outperforms correlation-based receivers and generalizes across channel conditions; mixed simulation+OTA training further improves OTA robustness. This approach reduces initial attach latency and retransmissions in 5G deployments and provides a path for extending PRACH reception to additional sequences and multi-user scenarios.

Abstract

Random Access is an important step in enabling the initial attachment of a User Equipment (UE) to a Base Station (gNB). The UE identifies itself by embedding a Preamble Index (RAPID) in the phase rotation of a known base sequence, which it transmits on the Physical Random Access Channel (PRACH). The signal on the PRACH also enables the estimation of propagation delay, often known as Timing Advance (TA), which is induced by virtue of the UE's position. Traditional receivers estimate the RAPID and TA using correlation-based techniques. This paper presents an alternative receiver approach that uses AI/ML models, wherein two neural networks are proposed, one for the RAPID and one for the TA. Different from other works, these two models can run in parallel as opposed to sequentially. Experiments with both simulated data and over-the-air hardware captures highlight the improved performance of the proposed AI/ML-based techniques compared to conventional correlation methods.

Enhancements for 5G NR PRACH Reception: An AI/ML Approach

TL;DR

The paper addresses robust RAPID detection and TA estimation for 5G NR PRACH, where correlation-based methods struggle in multipath and low SNR. It introduces two parallel ResNet-like classifiers that process raw frequency-domain OFDM data to classify RAPID among values and TA among delay steps, leveraging the phase rotations from RAPID ( with ) and TA. Evaluations on simulated TDLC channels and OTA hardware show that the NN-based receiver outperforms correlation-based receivers and generalizes across channel conditions; mixed simulation+OTA training further improves OTA robustness. This approach reduces initial attach latency and retransmissions in 5G deployments and provides a path for extending PRACH reception to additional sequences and multi-user scenarios.

Abstract

Random Access is an important step in enabling the initial attachment of a User Equipment (UE) to a Base Station (gNB). The UE identifies itself by embedding a Preamble Index (RAPID) in the phase rotation of a known base sequence, which it transmits on the Physical Random Access Channel (PRACH). The signal on the PRACH also enables the estimation of propagation delay, often known as Timing Advance (TA), which is induced by virtue of the UE's position. Traditional receivers estimate the RAPID and TA using correlation-based techniques. This paper presents an alternative receiver approach that uses AI/ML models, wherein two neural networks are proposed, one for the RAPID and one for the TA. Different from other works, these two models can run in parallel as opposed to sequentially. Experiments with both simulated data and over-the-air hardware captures highlight the improved performance of the proposed AI/ML-based techniques compared to conventional correlation methods.
Paper Structure (19 sections, 4 equations, 10 figures, 1 table)

This paper contains 19 sections, 4 equations, 10 figures, 1 table.

Table of Contents

  1. Introduction
  2. Our Contributions
  3. Background on 5G NR Physical Random Access Channel
  4. 5G Random Access Procedures
  5. PRACH Sequence Generation and Transmission
  6. Existing receivers for 5G NR PRACH
  7. Conventional approaches
  8. AI/ML approaches
  9. Proposed AI/ML model for the PRACH receiver
  10. Neural Network Architectures
  11. Dataset Generation
  12. Simulated datasets
  13. Over-the-Air (OTA) Hardware captured datasets
  14. Training and testing
  15. Training and testing on same channel scenario (Simulated Datasets)
  16. ...and 4 more sections

Figures (10)

  • Figure 1: Contention Based Random Access (CBRA) Procedure
  • Figure 2: Example correlation plot, showing division into preamble windows and shifting of the peaks due to propagation delay and fading channel effects.
  • Figure 3: The proposed AI/ML architecture replaces correlation operations with 2 neural networks that can run in parallel.
  • Figure 4: NN Architecture
  • Figure 5: IIT-Madras 5G testbed setup with the Remote Radio Head used as a receiver and the VSG used as a transmitter.
  • ...and 5 more figures