Soaring Beyond Rotors: Decentralized Coordination of Fixed-Wing UAVs for Energy-Efficient Wireless Coverage

Fixed-wing unmanned aerial vehicles (UAVs) use wing-based lift generation and aerodynamic principles to achieve higher performance than multi-rotor systems, allowing for long-term wireless network coverage. However, their constant motion requires intelligent trajectory planning to balance continued coverage, energy efficiency, and collision avoidance in a dynamic environment. A generalized framework for cooperative functioning and decentralized UAV swarm communication across mesh-based wireless networks is presented in this work. Building on this, the system maximizes energy-efficient flight paths while preserving dynamic network coverage by training a multi-agent reinforcement learning (MARL) policy. Cluster heads and high-priority zones for adaptive redirection are identified by UAVs exchanging local observations to create real-time cost maps. The framework is designed to support high-frequency LOS communication systems and is tested in real-world scenarios with fixed-wing UAVs via WiFi and assessed using NS-3 simulations. By providing a scalable solution to the energy-coverage-connectivity trade-off, the suggested technology provides persistent airborne networks for uses including surveillance, emergency response, and rural connectivity.