Implementation of Deep Learning in Spatial Multiplexing MIMO Communication

Abstract

Research in Multiple Input Multiple Output (MIMO) communication system has been developed rapidly in order to improve the effectiveness of communication among users. However, trade-off phenomenon between performance and computational complexity always become the hugest dilemma suffered by researchers. As an alternative solution, this paper proposes an optimization in 3x3 spatial multiplexing MIMO communication system using end-to-end based learning, specifically, it adapts autoencoder based model with the knowledge of Channel State Information (CSI) in the receiver side, make it fairly compared with the baseline method. The proposed models were evaluated in one of the most common channel impairment which is fast Rayleigh fading with additional Additive White Gaussian Noise (AWGN). By appropriately determining hyperparameters and the help of PReLU (Parametric Rectified Linear Unit), the results show that this autoencoder based MIMO communication system results in very promising results by exceeding the baseline methods (methods widely used in conventional MIMO communication) by reaching BER lower than at SNR 22.5 dB.