Denoising Deep Sparse Learning Based Channel Estimation for MU-MIMO Systems

Channel estimation for multi-user multiple input multiple output (MIMO) systems has been recognized as a key issue in next generation wireless communication. The wireless channel is approximately sparse due to the transmission noise effect, which limits the performance of the existing sparse channel estimation method. To address this problem, the denoising deep learning based channel estimation method for MU-MIMO system is proposed in this paper. Utilizing the denoising algorithm to remove noise perturbations in channel estimation, the proposed method can obtain the accurate sparse feature of wireless channels in MU-MIMO system. Moreover, the estimation accuracy and spectrum efficiency can be further improved by fully utilizing the spatial correlation among the massive MIMO channel. Simulation results demonstrated that, the proposed method can improve the channel estimation accuracy and robustness of massive MIMO channel compared with the existing benchmarks.