AI-Assisted Electromagnetic Radiation Reconfigurable Antennas for Energy-Efficient and Green 6G Wireless Networks: A Comprehensive Review

The evolution toward green and energy-efficient 6G wireless networks requires innovative antenna technologies capable of reducing power consumption while maintaining high communication performance. In this context, AI-assisted reconfigurable antennas have emerged as promising solutions for intelligent and sustainable wireless systems. By dynamically adapting their radiation patterns, these antennas can direct electromagnetic energy toward active users, thereby reducing unnecessary power dissipation and improving overall energy efficiency. Among these technologies, Reconfigurable Intelligent Surfaces (RIS) have gained significant attention as a new paradigm for programmable wireless environments. Indeed, RIS structures enable dynamic control of electromagnetic wave propagation, allowing efficient beam steering toward intended users while minimizing energy loss and enhancing spectral efficiency. This paper presents a comprehensive review of RIS based techniques for 6G wireless networks, with a particular focus on AI-assisted beamforming and energy-efficient antenna architectures. Thus, recent machine learning and deep learning approaches including Convolutional Neural Networks (CNN), CNN-LSTM, BF-Net, Deep Reinforcement Learning (DRL), and Soft Actor–Critic (SAC) with state-space exploration are examined. These techniques demonstrate strong potential for real-time beam steering, adaptive resource allocation, and mitigation of RIS hardware imperfections, contributing to the development of intelligent and energy-efficient 6G communication systems.