Cell-free massive MIMO signal detector fused with multi-head attention by graph neural network

IF 1.9 4区计算机科学 Q3 ENGINEERING, ELECTRICAL & ELECTRONIC

Physical Communication Pub Date : 2026-03-01 Epub Date: 2026-01-29 DOI:10.1016/j.phycom.2026.103023

Shihao Guo , Xiaohui Zhang , Gaoyuan Zhang

{"title":"Cell-free massive MIMO signal detector fused with multi-head attention by graph neural network","authors":"Shihao Guo , Xiaohui Zhang , Gaoyuan Zhang","doi":"10.1016/j.phycom.2026.103023","DOIUrl":null,"url":null,"abstract":"<div><div>The Cell-free massive multiple-input multiple-output (CF-mMIMO) system utilizes multiple access points (APs) to receive user information and has the advantage of uniform coverage. During uplink transmission, the aliased multi-user signals are aggregated by each AP to the central processing unit (CPU). Designing efficient detection algorithms is the key to eliminating multi-user interference and ensuring collaborative gain. In this paper, aiming at the limitations of the independent Gaussian approximation in expectation propagation (EP), and using the 2D multi-head attention (MHA) module to obtain the temporal and frequency correlations, a Graph neural network (GNN) detector based on EP and MHA is proposed. It is called a GNN-MHA-EP detector. The simulation results show that in different low-precision quantization systems, or when there are errors in channel estimation, the proposed detection algorithm is significantly superior to the traditional EP algorithm, belief propagation (BP) algorithm and minimum mean square error (MMSE) algorithm, and also significantly improves the performance of the GNN-EP detection algorithm and maintains the same order of computational complexity with it.</div></div>","PeriodicalId":48707,"journal":{"name":"Physical Communication","volume":"75 ","pages":"Article 103023"},"PeriodicalIF":1.9000,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Physical Communication","FirstCategoryId":"94","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1874490726000327","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2026/1/29 0:00:00","PubModel":"Epub","JCR":"Q3","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}

引用次数: 0

Abstract

The Cell-free massive multiple-input multiple-output (CF-mMIMO) system utilizes multiple access points (APs) to receive user information and has the advantage of uniform coverage. During uplink transmission, the aliased multi-user signals are aggregated by each AP to the central processing unit (CPU). Designing efficient detection algorithms is the key to eliminating multi-user interference and ensuring collaborative gain. In this paper, aiming at the limitations of the independent Gaussian approximation in expectation propagation (EP), and using the 2D multi-head attention (MHA) module to obtain the temporal and frequency correlations, a Graph neural network (GNN) detector based on EP and MHA is proposed. It is called a GNN-MHA-EP detector. The simulation results show that in different low-precision quantization systems, or when there are errors in channel estimation, the proposed detection algorithm is significantly superior to the traditional EP algorithm, belief propagation (BP) algorithm and minimum mean square error (MMSE) algorithm, and also significantly improves the performance of the GNN-EP detection algorithm and maintains the same order of computational complexity with it.

查看原文本刊更多论文

基于图神经网络融合的无小区海量MIMO信号检测器

无小区大规模多输入多输出（CF-mMIMO）系统利用多个接入点（ap）接收用户信息，具有均匀覆盖的优点。在上行链路传输过程中，混叠的多用户信号由每个AP汇聚到中央处理器（CPU）。设计高效的检测算法是消除多用户干扰、保证协同增益的关键。针对独立高斯近似在期望传播（EP）中的局限性，利用二维多头注意（MHA）模块获取期望传播的时间和频率相关性，提出了一种基于EP和MHA的图神经网络（GNN）检测器。它被称为GNN-MHA-EP探测器。仿真结果表明，在不同的低精度量化系统中，或信道估计存在误差时，所提出的检测算法明显优于传统的EP算法、信念传播（BP）算法和最小均方误差（MMSE）算法，并且显著提高了GNN-EP检测算法的性能，并与之保持相同的计算复杂度。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Physical Communication ENGINEERING, ELECTRICAL & ELECTRONICTELECO-TELECOMMUNICATIONS

CiteScore

5.00

自引率

9.10%

发文量

212

审稿时长

55 days

期刊介绍： PHYCOM: Physical Communication is an international and archival journal providing complete coverage of all topics of interest to those involved in all aspects of physical layer communications. Theoretical research contributions presenting new techniques, concepts or analyses, applied contributions reporting on experiences and experiments, and tutorials are published. Topics of interest include but are not limited to: Physical layer issues of Wireless Local Area Networks, WiMAX, Wireless Mesh Networks, Sensor and Ad Hoc Networks, PCS Systems; Radio access protocols and algorithms for the physical layer; Spread Spectrum Communications; Channel Modeling; Detection and Estimation; Modulation and Coding; Multiplexing and Carrier Techniques; Broadband Wireless Communications; Wireless Personal Communications; Multi-user Detection; Signal Separation and Interference rejection: Multimedia Communications over Wireless; DSP Applications to Wireless Systems; Experimental and Prototype Results; Multiple Access Techniques; Space-time Processing; Synchronization Techniques; Error Control Techniques; Cryptography; Software Radios; Tracking; Resource Allocation and Inference Management; Multi-rate and Multi-carrier Communications; Cross layer Design and Optimization; Propagation and Channel Characterization; OFDM Systems; MIMO Systems; Ultra-Wideband Communications; Cognitive Radio System Architectures; Platforms and Hardware Implementations for the Support of Cognitive, Radio Systems; Cognitive Radio Resource Management and Dynamic Spectrum Sharing.