Intelligent Beam and RACH Assignment in mmWave Communications

Abstract

Millimeter-wave (mmWave) communication enables high data rates and large bandwidth for next-generation wireless networks, making it a promising solution for industrial environments. However, mmWave links are highly sensitive to blockage and environmental geometry, resulting in strong spatial variations in signal-to-interference-plus-noise ratio (SINR) and challenging beam management. This paper proposes a digital twin-assisted reinforcement learning (RL) framework for beam management and resource allocation in mmWave industrial networks. Raytracing simulations within the DT are used to generate locationdependent SINR information, which is incorporated into the RL state space to guide beam and resource allocation selection decisions. Results show the proposed DT framework outperforms other state-of-the-art solutions in terms of data rate, success probability, and convergence time.