Energy Efficient Antenna Selection in Massive MIMO System Under Imperfect Channel State Information

Abstract

Massive MIMO systems have gained significant attention in recent years due to their potential to provide high data rates in wireless communication systems. However, the large number of antennas used in such systems results in high power consumption and decreased energy efficiency. Therefore, effective antenna selection schemes are required to reduce the number of active antennas and improve energy efficiency. Several antenna selection techniques have been proposed by researchers in the past few years, broadly classified into two categories: deterministic and stochastic. Deterministic algorithms aim to find the optimal subset of antennas by iteratively searching for the best combination of antennas using mathematical optimization techniques, while stochastic algorithms select a subset of antennas using a random search approach. Both techniques have limitations in terms of computational complexity and accuracy under rapidly changing channels. To overcome these limitations, this study proposes a deep learning-based approach that uses a deep neural network (DNN) for antenna selection. The proposed scheme in this thesis selects the antenna by taking into account the effect of imperfections in the channel state information and selects the antenna based on the factor less sensitive to channel imperfection. In the study, the impact of the distance between the user and the base station antenna and corresponding path losses on the energy efficiency of the system is evaluated. Then using this information as input to a deep learning scheme that considers the effect of SNR and channels the optimal number of antennas is determined. Finally, the accuracy of the proposed scheme is compared with some traditional machine learning techniques. The proposed scheme aimed to enhance the energy efficiency of massive MIMO systems while maintaining a certain level of performance.