基于图像超分辨率生成对抗网络的MIMO信道估计方法

Channel Estimation Method for Massive MIMO Based onImage Super-resolution Generative Adversarial Network

中文关键词:  大规模MIMO  信道估计  图像超分辨率  生成对抗网络

英文关键词:massive MIMO  channel estimation  image super-resolution  generative adversarial network

中文摘要:

      在毫米波大规模多输入多输出系统中,针对传统信道估计需要借助信道和噪声的统计特性来提高精度的弊端, 提出了一种基于图像超分辨率生成对抗网络的信道估计算法(SRGAN),将信道估计建模为图像超分辨率恢复问题。首 先,采取最小二乘算法得到导频位置处的信道信息;其次,通过二维线性插值得到分辨率较低的信道矩阵信息作为所提 SRGAN 网络的输入;最后,通过训练恢复出真实信道频率响应。仿真实验表明:文中所提信道估计算法的性能较传统信 道估计算法有较大的提升,并且恢复的信道更符合真实信道。

英文摘要:

      In millimeter-wave massive multiple-input multiple-output systems, for the traditional channel estimation the accuracy can be improved by using the statistical characteristics of the channel and noise. In this paper a channel estimation algorithm based on image super-resolution generative adversarial network (SRGAN) is proposed, which models channel estimation as the problem of image super-resolution recovery. Firstly, the least square algorithm is used to obtain the channel information at the pilot position, and then the channel matrix information with low resolution is obtained by two-dimensional linear interpolation, and finally it is used as the input of the SRGAN network proposed in this paper to recover the true channel frequency response of the channel through training. The simulation results show that the performance of the proposed channel estimation algorithm is greatly improved compared with the traditional channel estimation algorithm, and the recovered channel is more consistent with the real channel.