Collaborative Anti-Jamming for Multi-UAV Communications Based on Deep Reinforcement Learning

This paper investigates the issue of anti-jamming spectrum sharing in multi-UAV communication systems using deep reinforcement learning. To address the complex spectrum environment and dynamic jamming issues in the system, we propose a multi-UAV distributed collaborative anti-jamming method based on the SAC algorithm. By introducing an entropy regularity term, the algorithm is enabled to adaptively find an effective compromise between exploration and exploitation. The method is decentralized, eliminating the need for information exchange, enabling UAVs to independently explore the spectrum environment and adapt to dynamic jamming changes. Simulation experiments demonstrate that compared with traditional methods, this method enhances the anti-jamming performance and spectrum utilization efficiency of the system, while also improving the adaptability and communication robustness of multi-UAV communication systems in complex jamming environments.