Intent-driven UAV swarm network for emergency disaster relief

Facing the severe challenge of major natural disasters such as earthquakes and floods, which damage ground communication infrastructure and turn disaster-stricken areas into “information islands”, this paper proposes an intent-driven emergency networking architecture for Unmanned Aerial Vehicle (UAV) swarms. By integrating natural language processing and knowledge graph technologies, this architecture achieves intelligent representation and parsing of high-level rescue instructions at the intent layer, translating them into executable commands that are issued to the UAV swarm at the data layer. A multi-objective optimization problem is formulated to jointly optimize the dynamic deployment of the UAV swarm, aiming to minimize the energy consumption and transmission delay of both UAVs and ground users while maximizing network coverage. An Intent-Driven Artificial Neural Network (ID-ANN) self-optimization algorithm is designed, which employs K-means for region partitioning and introduces the Fermat point theory to optimize the matching between UAVs and users. Simulation results demonstrate that the proposed algorithm achieves an average improvement of 20.2% in network signal coverage, a 33.33% reduction in average reconfiguration delay, and a 12.72% increase in throughput, significantly outperforming baseline algorithms. This study provides theoretical support and a feasible technical pathway for the construction of space-based emergency communication systems, offering valuable insights for enhancing communication support capabilities in disaster emergency rescue operations.