基于WGAN的智能超表面辅助系统的信道估计研究  

Channel estimation study of reconfigurable intelligent surface aided system based on WGAN

作　　者：康晓非[1] 王甜 KANG Xiaofei;WANG Tian(College of Communication and Information Engineering,Xi’an University of Science and Technology,Xi’an 710600,China)

机构地区：[1]西安科技大学通信与信息工程学院,西安710600

出　　处：《电波科学学报》2025年第1期164-171,共8页Chinese Journal of Radio Science

基　　金：国家自然科学基金(61801372)。

摘　　要：针对智能超表面(reconfigurable intelligent surface,RIS)辅助的毫米波通信中系统复杂和难以获取准确信道状态信息(channel state information,CSI)的问题,设计了一种基于Chan-SRWGAN网络算法的信道估计方案。该方案采用混合有源/无源RIS架构,首先利用最小二乘(least square,LS)算法获取有源元件处信道估计值,再通过插值得到信道初步估计,最后利用Chan-SRWGAN深度学习网络将其重构为信道精确估计。仿真结果表明,所提方案的归一化均方误差(normalized mean squared error,NMSE)性能优于LS、正交匹配追踪(orthogonal matching pursuit,OMP)、同步OMP(simultaneous OMP,SOMP)、深度神经网络(deep neural network,DNN)、超分辨率卷积神经网络(super-resolution convolutional neural network,SRCNN)信道估计算法,证明了方案的可行性。Aiming at the problem that the system is complex and it is difficult to obtain accurate channel state information(CSI)in millimeter wave communication assisted by reconfigurable intelligent surface(RIS),this paper designs a channel estimation scheme of Chan-SRWGAN network algorithm.The scheme adopts a hybrid active/passive RIS architecture.Firstly,the least square(LS)algorithm is used to obtain the channel estimation value at the active elements,and then the preliminary channel estimation is obtained by interpolation.Finally,the Chan-SRWGAN deep learning network is used to reconstruct it into accurate channel estimation.Simulation results show that the proposed scheme outperforms LS,orthogonal matching pursuit(OMP),simultaneous orthogonal matching pursuit(SOMP),deep neural network(DNN),super-resolution convolutional neural network(SRCNN)channel estimation algorithms in terms of normalized mean squared error(NMSE)performance,thus confirming the feasibility of the approach.

关 键 词：智能超表面(RIS) 信道估计 深度学习 Wasserstein生成对抗网络(WGAN) 超分辨率卷积神经网络(SRCNN) 

分 类 号：TN929.5[电子电信—通信与信息系统]