AI-Based Adaptive Beamforming for Wireless Networks : Mitigating NLOS and 3D Beamforming Challenges for Enhanced Coverage

초록

Beam management (BM) is a critical component of 5G and emerging 6G wireless networks, as emphasized by 3rd generation partnership project (3GPP) standards, enabling efficient signal alignment, enhanced coverage, and reliable communication in dynamic environments. It optimizes beamforming to maximize signal-to-interference-plus-noise ratio (SINR), reduces interference, and ensures robust signal transmission in scenarios like urban areas, high-speed mobility, and indoor environments. However, challenges like non-line of sight (NLOS) propagation in dense urban areas, rapid direction of arrival (DOA) changes in high-mobility scenarios, and propagation loss at mmWave frequencies degrade signal reliability. This research proposes 3D beamforming adaptive network (3DBAN) for NLOS environments, leveraging federated learning (FL), generative adversarial networks (GANs), and reinforcement learning (RL) to dynamically predict angle of arrival (AOA) and angle of departure (AOD), optimize beamforming weights for 3D coverage, and adapt to evolving conditions. By integrating real-time and synthetic data, AI-driven systems enable precise beam steering, interference nulling, and robust performance, ensuring scalability and adaptability in resource-constrained scenarios.