Abstract
In a post-disaster area, Unmanned Aerial Vehicles (UAVs) are considered one of the most effective ways to provide emergency wireless communication services, especially when the wireless infrastructure becomes malfunctioned due to the extensive damage. In this paper, we investigate the UAV-based emergency wireless communication network for a post-disaster area, where the UAV acts as a flying Base Station (BS) to provide wireless connectivity from the sky. UAV should collect as much data as possible from ground users in the affected area. Considering the malfunction of power supplies in the post-disaster area, the available energy for ground users is very limited. Moreover, UAV operates with an onboard battery with a limited capacity. Aiming to maximize the uplink throughput by maximizing the number of visited ground users during the flight round, the UAV trajectory optimization problem is formulated under the concern of dual limited available energy (i.e., limited ground user and UAV energy capacities). Considering that both energy terms are dynamic and cumulative over time, this opti-mization problem becomes hard to be solved using conventional optimization methods. Therefore, a multi-armed bandit (MAB)-based algorithm controlled with dual limited energy capacities is proposed to tackle this problem. The simulation results show that the proposed algorithm could solve the optimization problem and maximize the achievable throughput under these energy constraints.