基于频谱环境网络化认知的无人机网络资源高效利用与优化

With the rapid development of unmanned aerial vehicle (UAV) industry, the spectrum resources licensed to UAV networks will be exhausted in the near future. Meanwhile, UAV networks are vulnerable due to malicious jamming, which may cause that UAVs are captured or damaged, and significant information in UAVs is stolen. To address the dual challenges of spectrum shortage and malicious jamming faced by UAV networks, it is a promising technical means to reuse idle spectrum and combat malicious jamming through networked cognition. Therefore, this project studies the high-efficiency resource utilization and optimization problem for UAV networks based on networked cognition of spectrum environments, and it mainly includes: For the precise spectrum cognition problem of UAV networks moving with high speed, the traditional cognition of spectrum opportunities is expanded to the cognition of spectrum environments, and the networked cognition methods are proposed. For the high-efficiency resource utilization problem of UAV networks without malicious jamming, three kinds of typical UAV movement patterns are summarized and the corresponding approaches to jointly optimize UAV trajectories and network resources are proposed. For the high-efficiency resource utilization problem of UAV networks with malicious jamming, the networked anti-jamming approaches are proposed, including robust transmission, jamming avoidance, and jamming deception. This project fully explores the high-speed moving characteristics of UAV networks, and makes innovations in networked cognition, resource utilization, anti-jamming, etc. Thus, this research work has theoretical and practical significance to guide the design of UAV networks.

随着无人机产业的蓬勃发展，无人机网络的频谱资源将会面临枯竭。同时，无人机网络容易遭受恶意干扰，导致无人机被捕获、损毁及重要信息丢失。针对无人机网络面临的频谱短缺和恶意干扰双重挑战，通过认知再利用周围空闲频谱以及对抗恶意干扰是一种重要的技术解决手段，因此本项目研究基于频谱环境网络化认知的无人机网络资源高效利用与优化问题。主要包括：针对高速运动的无人机网络精确认知频谱问题，拓展传统“频谱机会”认知为“频谱环境”认知，并提出“网络化”认知方法；针对无恶意干扰情况下资源高效利用问题，归纳出三种典型无人机运动模式，提出相应的无人机轨迹和网络资源联合优化方法；针对恶意干扰情况下资源高效利用问题，提出包括鲁棒传输、干扰避让及干扰诱骗的“网络化”抗干扰方法。本项目充分挖掘无人机网络的高速运动特性，在“网络化”认知、资源利用及抗干扰等方面形成创新，研究成果对于无人机网络设计具有理论及实际指导意义。

随着无人机产业的蓬勃发展，无人机通信网络频谱资源将会面临枯竭。同时，无人机网络容易遭受恶意干扰，导致其性能急剧下降，甚至发生损毁、坠机等危险。针对无人机网络面临的频谱短缺和恶意干扰双重挑战，通过认知再利用周围空闲频谱以及对抗恶意干扰是一种重要的技术解决手段。因此，本项目针对基于频谱环境网络化认知的无人机网络资源高效利用与优化问题展开研究，旨在提升认知无人机网络的资源利用效率和抗干扰能力。.主要工作包括：首先，拓展传统“频谱机会”认知为“频谱环境”认知，并提出“网络化”认知方法，利用无监督聚类和相位差特征提取等方法，进行频谱机会、频谱拥有网络和干扰节点位置网络化认知和频谱环境网络化多维协同认知。进一步，基于网络化认知结果，针对三种典型无人机运动模式，利用多优先级排队理论和强化学习等方法，提出了无人机轨迹和频谱、功率等网络资源联合优化模型和方案。最后，针对恶意干扰场景，提出了基于无速率编码等方案的数据鲁棒传输方法。以此为基础，逐步研究并提出了基于人工势场法的干扰避让策略，基于恶意干扰功率诱骗以及基于Raptor码随机接入干扰诱骗的无人机网络抗干扰策略，从而增强了无人机抗干扰能力，提升了网络容量。通过4年攻关研究，项目发表SCI/EI论文32篇，其中中科院一区/二区SCI论文12篇，通信顶级会议6篇，获授权国家发明专利4项；针对高速运动的无人机网络精确认知频谱问题，提出了基于相位差特征提取等方法的无人机网络频谱环境精准网络化认知方法；针对无恶意干扰情况下资源高效利用问题，提出了任务驱动下无人机网络频谱资源高效利用方法；针对恶意干扰情况下资源高效利用问题，提出了陌生环境下无人机网络恶意干扰的对抗方法。本项目区别于传统基于频谱机会的地面认知网络研究，在“网络化”认知、资源利用及抗干扰等方面形成创新，对于无人机网络的资源利用效率提升、抗干扰能力增强具有重要的理论及实际指导意义。

内容获取失败，请点击重试