Task-Oriented Multi-Hop Backhaul in Flying Ad-Hoc Network

In this paper, we study the task-oriented multi-hop backhaul problem in a flying ad-hoc network (FANET). Specifically, UAVs perform multi-class tasks and forward the heterogeneous real-time traffic towards the ground base station (GBS). The problem is to bridge the task demand and end-to-end wireless resource supply concurrently from both sides of the network, and then maximize the overall backhaul benefits. We propose a cross-layer design jointly optimizing the link resource allocation, channel allocation, and power control strategies. The non-convex, coupled, hierarchical, and multi-staged features perplex this problem. To address these challenges, the multi-leaders one-follower Stackelberg game framework is introduced to analyze the mutual influence. By leveraging log-transform, convex approximation, and exact potential game techniques, we prove the existence of Stackelberg equilibrium (SE) and propose a sequential fixing successive convex approximation-best response (SFSCA - BR) based hierarchical iterative algorithm. Simulation results verify the effectiveness.