Game-Theoretic Optimization of UAV Swarm and Relay Communication for Interference Resilience

This paper investigates the trade-off between relay communication and intra-swarm communication performance in UAV swarms operating under communication and navigation jamming environments, aiming to optimize the link between UAV swarms and the ground station. Specifically, during the UAV swarm's flight towards a predefined destination, the relay UAV selection, including external relay UAV and inter-swarm relay UAVs, and relay communication power allocation need to balance external relay communication, inter-swarm relay communication and intra-swarm communication performance. The external relay communication in this context refers to the communication between relay UAVs and the ground base station, with its power consumption also taken into account. Inter-swarm communication denotes the communication between external relay UAV and inter-swarm relay UAVs. Given that relay selection and power allocation in this problem exhibit non-convex multi-objective optimization properties, we propose using a game-theoretic framework to obtain a balanced equilibrium solution. The UAVs adjust their relay selection and power allocation strategies to maximize swarm and relay communication rates while minimizing relay power consumption, which has been proven to be an exact potential game, ensuring the existence of a Nash equilibrium (NE) strategy. To solve the equilibrium strategy, and considering the diversity of equilibrium solutions in the flight environment, as well as the continuity of spatial location, navigation, and communication performance parameters, we propose a solution algorithm based on a maximum entropy reinforcement learning (MERL) framework. Simulation results highlight the algorithm's superiority over tabular reinforcement learning and conventional strategies, achieving lower power consumption and reduced communication interruptions between the swarm and ground station in high-interference environments. This research provides valuable insights into UAV swarm collaborative positioning based on communication in navigation interruption environments, as UAV swarms rely on ground base stations for relative positioning information.