深度学习辅助的5G OFDM系统的信道估计OACSTPCD

The traditional channel estimation algorithm is difficult to meet the requirement of high speed and low delay in 5G system.For this problem,the authors propose a channel estimation method based on image restoration technology by considering the time-frequency response of communication channel as a two-dimensional image.First,parameters are set to generate a channel data information data set of physical downlink shared channel(PDSCH)based on 5G new radio(NR)standard,and the generated channel matrix is treated as a two-dimensional image.Then,an image restoration network based on convolutional neural network is constructed,and residual connection is incorporated to improve the performance of the network.Finally,the trained network model is used for channel estimation.The simulation results show that the performance of the proposed channel estimation algorithm is significantly improved compared with those of the Least Square(LS),Practical Channel Estimation(PCE)and Image-based Super-resolution ChannelNet network.