UAV Path Planning Based on Reinforcement Learning for Fair Resource Allocation in UAV-Relayed Cellular Networks

UAV-relayed cellular network is one of the promising applications of UAV systems. UAV can be used to increase the coverage of cellular networks or provide service to areas where infrastructure installation is difficult or impossible. However, unlike existing infrastructure-based cellular networks, the resources allocated to user terminals may be unbalanced due to the limited number of UAVs and change in coverage due to the movements of UAVs. To solve this problem, we propose a path planning that minimizes the unfairness using reinforcement learning. The UAV evaluates the local fairness according to the information of user terminal within the communication range of the UAV, then it determines the appropriate path to increase the global fairness.