Robust PRACH Preamble Detection Using GATv2-Based Graph Learning for 5G/6G Networks

  • Le, Viet Anh-Huy; 
  • Tran, Anh-Tien; 
  • Dang, Thanh Thien-An; 
  • Cho, Sungrae
Citations

SCOPUS

0

초록

The evolution from 5G toward 5G-Advanced and future 6G systems is expected to support increasingly dense, heterogeneous, and delay-sensitive wireless access. In this context, the random access (RA) procedure is a critical entry point because each user equipment (UE) must transmit a Physical Random Access Channel (PRACH) preamble before uplink synchronization and connection establishment can proceed. However, conventional correlation-based PRACH detection can become unreliable under low signal-to-noise ratio (SNR), multipath fading, Doppler spread, and dense-access interference. This paper proposes a graph-based PRACH preamble detection algorithm using Graph Attention Network version 2 (GATv2). The proposed method models the power delay profile (PDP) as a graph and uses dynamic attention to emphasize structurally consistent signal components while suppressing noise-induced peaks. Simulation results show that the proposed detector improves detection probability compared with the conventional correlation-based method, particularly in the low-SNR region.

키워드

5G; GATv2; graph neural network; PRACH; random access
제목
Robust PRACH Preamble Detection Using GATv2-Based Graph Learning for 5G/6G Networks
저자
Le, Viet Anh-Huy; Tran, Anh-Tien; Dang, Thanh Thien-An; Cho, Sungrae
DOI
10.1109/ICUFN69619.2026.11628565
발행일
2026-07
유형
Conference Paper
저널명
International Conference on Ubiquitous and Future Networks, ICUFN
페이지
228 ~ 232