AI-Driven Spectrum Intelligence and Energy Optimization Techniques for 6G Networks: A Comprehensive Review

As wireless technology moves toward sixth generation (6G) networks, future wireless communication systems must be intelligent, self-optimizing, and energy-efficient. To address increasing pressure on spectrum, systems will need to integrate AI (artificial intelligence) and ML (machine learning) technologies to facilitate dynamic spectrum access, efficient energy distributions, and max data throughput. This review examines key AI-driven techniques for spectrum intelligence and energy optimization in 6G networks. The applications of spectrum prediction employing deep learning, reinforcement learning for dynamic spectrum allocation of resources, and federated learning approaches that provide distributed decision-making capabilities in heterogeneous wireless environments are highlighted. The roles of intelligent multiple-input multiple-output (MIMO) antennae and reconfigurable surfaces as design tools for maximizing the spectral efficiency and energy efficiency of systems is also examined. Recent designs in MIMO operating in sub-6 GHz have been employed as reference points for novel systems. Comparative analyses show that the AI enabled spectrum optimization techniques vastly outperform conventional static allocation in spectral efficiency, latency and energy efficiency. Finally, the review concludes with a discussion on the outstanding research challenges corresponding to trustworthy AI frameworks, sustainable energy frameworks, spectrum sharing etc., which will lead us to towards fully autonomous and energy efficient 6G networks.