MADRL Based Joint Resource Allocation and 3D CoordinateOptimization of UAVs in Air-Ground Integrated IoT Networks

With the development of IoT networks, more and more types of devices are connected to the Internet, and different devices generate different tasks.Therefore, a reasonable resource allocation method with limited resources is crucial.This paper, we introduce Unmanned Aerial Vehicles (UAVs) as an airborne base station (BS) to provide users with more flexible deployable resources.All BSs carry bandwidth resources and computing resources, and all BSs can dynamically perform network slicing depending on the type of service requested by the user.Therefore, K-means++ Algorithm is used to optimize the 2D coordinates of the UAV, while the MATD3 Algorithm is used to jointly optimize the altitude of the UAV and resource slice fraction.Simulation results show that the proposed Algorithm has better packet arrival rate as well as lower latency.