Deployment Optimization of Tethered Drone-Assisted Integrated Access and Backhaul Networks
Yongqiang Zhang, Mustafa A. Kishk, Mohamed‐Slim Alouini
Abstract
Millimeter-wave (mmWave) integrated access and backhaul (IAB) has recently received considerable interest for its advantage in reducing the expenses related to the deployment of fiber optics, such as the Terragraph proposed by Meta’s Connectivity Lab. Terragraph networks aim to provide high-speed internet access to dense urban environments. However, due to the vulnerability to blockages and high path loss associated with mmWave frequencies, the proper deployment planning of mmWave networks is required to achieve the desired service quality. By obtaining a stable power supply through its tether connected to the ground, tethered unmanned aerial vehicle (UAV)-mounted base station (BS) can provide reliable communication service with the sacrifice of limited mobility. In this paper, we investigate the potential of incorporating tethered UAVs into Terragraph-like networks. To this end, we propose a novel deep reinforcement learning (DRL) framework that aims to minimize the overall deployment cost by optimizing the number of required UAVs and terrestrial BSs (TBSs), the hovering positions of deployed UAVs, and the multi-hop backhauling topology. Unlike the conventional DRL frameworks that focus on maximizing the expected cumulative or average reward, we formulate the proposed framework based on the max-Bellman optimality equation in order to maximize the maximum reward. Numerical results reveal that the proposed algorithm is able to yield significant reduction in terms of deployment cost. We also use case studies from cities in Asia, Europe, and North America to verify the practical applicability of the proposed framework.