Integration of Massive MIMO and Machine Learning in the Present and Future of Power Consumption in Wireless Networks: A Review

Abstract

The steady increase in data traffic rates and systems’ complexity have contributed to the information and communication technologies (ICT) sector’s increased energy consumption and CO2 emissions. These pose a significant challenge to the telecommunication industry and the environment. This challenge has necessitated considering energy efficiency as a critical design pillar in 5G and future wireless networks. As a result, current research efforts for future wireless networks focus on minimising energy usage and improving efficiency. This work investigates several energy optimisation techniques in the present and future wireless networks, their contributions, advantages, and limitations. Based on the review of different techniques, we discuss the architecture of the massive MIMO (mMIMO) technique, including its operation and requirements. We also present the performance evaluation of mMIMO using different precoding algorithms, which is crucial for energy efficiency in future wireless networks. We further review incorporating intelligence using a Machine Learning (ML) approach in switching off underused mMIMO arrays to minimise energy usage. Finally, we discuss several critical open research issues in mMIMO and ML that make future research and implementation possible in next-generation wireless networks.