Joint Optimization of Multi-UAV Topology Control, Offloading and Path Planning in Air-Space Edge Computing Network

In the event of an emergency, such as a post-disaster response scenario, uploading information to the edge cloud for processing can often result in considerable delays. In such contexts, the utilization of unmanned aerial vehicle (UAV)-assisted edge computing emerges as a viable solution to minimize the time taken for information processing. In this paper, we aim to establish a comprehensive data acquisition system, encompassing UAV swarms, ground users, and satellites. Specifically, during emergencies, satellites will be instrumental in planning the trajectory of the UAV swarms based on the observed disaster scene, including signal points to be collected and potential obstacles. This approach has the potential to significantly reduce rescue time and mitigate losses. Given the intricate nature of the proposed problem, which is influenced by numerous factors, a straightforward convex optimization approach is insufficient. Instead, we adopt a distributed approach to tackle the problem in two distinct parts. Firstly, we establish a fixed topology for the drones, and resolve the associated challenges through the application of a suitable control protocol. Secondly, we focus on optimizing the overall system’s energy loss. This is achieved by reformulating the problem as a Markov decision-making process (MDP) and subsequently addressing it through the application of a deep-Q-network (DQN).