Deployment of unmanned aerial vehicles (UAVs) as aerial base stations can deliver a fast and flexible solution for serving varying traffic demand. In order to adequately benefit of UAVs deployment, their efficient placement is of utmost importance, and requires to intelligently adapt to the environment changes. In this paper, we propose a learning-based mechanism for the three-dimensional deployment of UAVs assisting terrestrial cellular networks in the downlink. The problem is modeled as a noncooperative game among UAVs in satisfaction form. To solve the game, we utilize a low complexity algorithm, in which unsatisfied UAVs update their locations based on a learning algorithm. Simulation results reveal that the proposed UAV placement algorithm yields significant performance gains up to about 52% and 74% in terms of throughput and the number of dropped users, respectively, compared to an optimized baseline algorithm.
The integration of space and air components considering satellites and unmanned aerial vehicles (UAVs) into terrestrial networks in a space-terrestrial integrated network (STIN) has been envisioned as a promising solution to enhancing the terrestrial networks in terms of fairness, performance, and network resilience. However, employing UAVs introduces some key challenges, among which backhaul connectivity, resource management, and efficient three-dimensional (3D) trajectory designs of UAVs are very crucial. In this paper, low-Earth orbit (LEO) satellites are employed to alleviate the backhaul connectivity issues with UAV networks, where we address the problem of jointly determining backhaul-aware 3D trajectories of UAVs, resource management, and associations between users, satellites and base stations (BSs) in an STIN while satisfying ground users' quality-of-experience requirements and provisioning fairness concerning users' data rates. The proposed approach maximizes a novel objective function with joint consideration for BS's load and fairness, which can be categorized as a non-deterministic polynomial time hard (NPhard) problem. To tackle this issue, we leverage a reinforcement learning framework, in which our problem is modeled as a multi-armed bandit problem. Accordingly, BSs learn the environment and its dynamics and then make decisions under an upper confidence bound based method. Simulation results show that our proposed approach outperforms the benchmark methods in terms of fairness, throughput, and load.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.