基于强化学习的双簇头无人机蜂群通信抗干扰算法  

Anti-Jamming Algorithm for Dual-Cluster Head UAV Swarm Communications Based on Reinforcement Learning

作　　者：杨斌善 魏琛 夏晓晨 许魁 叶子绿 耿存奕 廖程建 YANG Binshan;WEI Chen;XIA Xiaochen;XU Kui;YE Zilv;GENG Cunyi;LIAO Chengjian(The Army Engineering University of PLA,Nanjing 210007,China)

机构地区：[1]陆军工程大学,江苏南京210007

出　　处：《移动通信》2024年第9期94-101,共8页Mobile Communications

基　　金：国家自然科学基金“通信定位一体去蜂窝大规模MIMO智能传输方法研究”,“面向低空空域的无定形大规模MIMO一体化感知与通信方法研究”(62071485,62271503);江苏省基础研究计划“天地融合卫星移动通信组网理论与技术”(BK20192002);江苏省自然科学基金“基于无定形网络的低空空域融合感知与通信方法研究”(BK20231485)。

摘　　要：无人机蜂群是指由多个无人机组成的集群,在复杂环境中协同完成任务。但无人机通信易受到电磁干扰,使得部分无人机失去中继转发功能,针对这一情况,提出了一种强化学习辅助的双簇头无人机蜂群通信抗干扰算法(DCHRLC)。通过提出的双簇头HEED算法对无人机进行重新分簇,在簇内选举产生双簇头,在簇间多跳路由中,通过强化学习对簇头节点的下一跳选择进行优化。仿真结果表明,相比于传统HEED算法,采用双簇头HEED算法情况下网络寿命、能耗均衡等方面具有明显的增益。Swarms of unmanned aerial vehicles(UAVs)are designed to collaborate and accomplish tasks in complex environments.However,UAV communication is highly vulnerable to electromagnetic interference,which can cause some UAVs to lose their relay forwarding capabilities.To address this issue,a reinforcement learning-based dual-cluster head anti-jamming communication algorithm for UAV swarms is proposed.The UAVs are re-clustered using the proposed dual-cluster head HEED algorithm,where two cluster heads are elected within each cluster.In inter-cluster multi-hop routing,reinforcement learning is employed to optimize the next-hop selection of the cluster head nodes.Simulation results demonstrate that,compared to the traditional HEED algorithm,the dual-cluster head HEED algorithm significantly enhances network lifetime and improves energy consumption balance.

关 键 词：无人机 双簇头 抗干扰 强化学习 

分 类 号：TN929.5[电子电信—通信与信息系统]