《计算机应用》唯一官方网站 ›› 2026, Vol. 46 ›› Issue (8): 2548-2554.DOI: 10.11772/j.issn.1001-9081.2025070859

• 网络与通信 • 上一篇    下一篇

基于深度学习的STAR⁃RIS辅助混合场通信系统信道估计

郭晓金1,2, 隋旭洋1,2(), 周柯男1,2   

  1. 1.重庆邮电大学 通信与信息工程学院,重庆 400065
    2.移动通信技术重庆市重点实验室(重庆邮电大学),重庆 400065
  • 收稿日期:2025-07-31 修回日期:2025-09-25 接受日期:2025-09-29 发布日期:2025-11-05 出版日期:2026-08-10
  • 通讯作者: 隋旭洋
  • 作者简介:郭晓金(1974—),男,山东新泰人,副教授,博士,主要研究方向:移动通信物理层算法
    隋旭洋(2000—),男,安徽马鞍山人,硕士研究生,主要研究方向:移动通信物理层算法、深度学习
    周柯男(2001—),男,云南曲靖人,硕士研究生,主要研究方向:移动通信物理层算法、信道估计。
  • 基金资助:
    重庆市自然科学基金创新发展联合基金(中国星网)资助项目(CSTB2023NSCQ-LZX0114)

Channel estimation of STAR-RIS-assisted hybrid-field communication system based on deep learning

Xiaojin GUO1,2, Xuyang SUI1,2(), Kenan ZHOU1,2   

  1. 1.School of Communication and Information Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,China
    2.Chongqing Key Laboratory of Mobile Communication Technology (Chongqing University of Posts and Telecommunications),Chongqing 400065,China
  • Received:2025-07-31 Revised:2025-09-25 Accepted:2025-09-29 Online:2025-11-05 Published:2026-08-10
  • Contact: Xuyang SUI
  • About author:GUO Xiaojin, born in 1974, Ph. D., associate professor. His research interests include mobile communication physical layer algorithms.
    ZHOU Kenan, born in 2001, M. S. candidate. His research interests include mobile communication physical layer algorithms, channel estimation.
  • Supported by:
    Innovation and Development Joint Fund of Chongqing Natural Science Foundation (China Satellite Network)(CSTB2023NSCQ-LZX0114)

摘要:

针对同时透射与反射可重构智能表面(STAR-RIS)辅助混合场通信系统在信道估计中面临的导频开销高和精度低的问题,提出一种基于多尺度特征感知融合的超分重建信道估计网络(MFSRN)。该网络将信道估计问题建模为信道矩阵的重建任务,首先,从上采样的信道矩阵中提取公共特征;随后,利用注意力门限分离各用户的特定信道特征;最后,并行地完成多用户信道矩阵的重建。为了提升网络性能,设计增强卷积模块(ECB)和双流特征感知(DSFA)模块,以有效提取混合场信道的局部和全局特征,并引入特征金字塔注意力(FPA)模块增强模型对多尺度特征的聚合能力。实验结果表明,所提方案在实现更低导频开销的同时,与经典的DRSN和U-MLP方案相比,归一化均方误差(NMSE)在不同信噪比(SNR)下分别平均降低了3.62 dB和1.28 dB,表现出更好的估计性能。

关键词: 深度学习, 同时透射与反射可重构智能表面, 信道估计, 混合场, 注意力机制

Abstract:

To address the problems of high pilot overhead and low accuracy in channel estimation of Simultaneously Transmitting and Reflecting Reconfigurable Intelligent Surface (STAR-RIS)-assisted hybrid-field communication systems, a Multi-scale Feature-aware Fusion Super Resolution Network (MFSRN) was proposed. In the network, channel estimation problem was modeled as a channel matrix reconstruction task. First, common features were extracted from an upsampled channel matrix; then, an attention gate was utilized to separate user-specific channel features; finally, reconstruction of multi-user channel matrices was completed in parallel. To enhance network performance, an Enhanced Convolution Block (ECB) and a Dual-Stream Feature-Aware (DSFA) block were designed to extract both local and global features of hybrid-field channel effectively, and a Feature Pyramid Attention (FPA) block was introduced to strengthen model’s capability for multi-scale feature aggregation. Experimental results demonstrate that compared with classic DRSN and U-MLP schemes, the proposed scheme achieves a lower pilot overhead and has the Normalized Mean Square Error (NMSE) reduced by an average of 3.62 dB and 1.28 dB respectively at different Signal-to-Noise Ratio (SNR), showing better estimation performance.

Key words: deep learning, Simultaneously Transmitting and Reflecting Intelligent Surface (STAR-RIS), channel estimation, hybrid-field, attention mechanism

中图分类号: