高动态星地融合多级边缘网络协同控制技术研究

This project is aimed at the deep integration of 5G wireless networks and satellite communication networks. Based on the space-time characteristics of the wireless communication network, this project studies the service space-time model in high-dynamic satellite-terrestrial networks to characterize the continuity, suddenness, and supportive characteristics of the traffic. In addition, the high-dynamic network organization and prediction model of the satellite-integrated network is proposed. With these models, the architecture of multi-edge integrated satellite-terrestrial networks is proposed, and the network service control system is studied, mining network organization with the multi-edge control group and cooperative adaption mechanism for high dynamic integrated satellite-terrestrial networks. Moreover, intelligent decision-making and collaborative control mechanism are studied in integrated satellite-terrestrial networks, along with the approaches of intelligent backhaul and collaborative transmission and control technology..The project is intended to solve three scientific problems: the “uncertainty” of traffic space-time fluctuation and high-speed satellite movements, the “isolation” of satellite and terrestrial network control schemes, and the “mismatch” of traffic requirements and network resources. The project is expected to be achieved to solve the contradiction between the high dynamic of traffic requirements and the static resource utilization, the contradiction between the independent control of satellite-terrestrial networks and the development of network integration, and explore the fundamental breakthrough to optimize the architecture of the integrated satellite-terrestrial network. These theoretical and technical innovations of coordinated transmission and control innovations can provide a research foundation for engineering applications in future construction and application of wireless integrated satellite-terrestrial networks.

本项目面向5G无线网络与卫星通信网络深度融合的重大需求，研究高动态星地融合网络环境中的业务空时模型，对业务的连续性、突发性、保障性需求特征，以及星地融合网络高动态组网及预测模型进行建模；提出星地融合多级边缘网络体系架构，对星地网络服务控制体系进行深入研究，挖掘面向高动态星地融合网络的多级边缘控制群组网及协作适配机理；同时研究星地融合网络智能决策与协同控制机制，实现基于高动态业务认知的智能回传方法及协同传输与控制技术。本项目拟解决三个科学问题：业务空时变化与卫星高速运动的“不确定性”问题、卫星与地面网络控制模式的“孤立性”问题、业务需求与网络资源的“失配性”问题，打破业务高动态性和资源静态利用之间的矛盾、卫星地面网络独立控制与网络融合发展之间的矛盾，探索从根本上突破星地网络构建和协同传输与控制的理论及技术瓶颈，为构建下一代卫星地面深入融合的无线通信网络工程应用提供研究基础。

星地融合通信系统受到高动态空时业务及网络变化，以及星地资源独立管控的严重制约，严重影响了系统的频谱效率、覆盖效率以及能量效率。本课题针对星地网络融合管控的根本问题，探索提出星地融合多级边缘网络体系架构及控制机理，将大数据、边缘计算和网络协作控制技术引入高动态卫星地面融合网络中，以解决业务空时动态变化与卫星高速动态运动的“不确定性”问题、卫星与地面网络控制模式的“孤立性”问题和业务需求与网络资源高效利用的“失配性”。首先，针对蜂窝网络采集得到的实测数据，利用卷积神经网络模型提取出基站业务流量在时间、空间维度上的相关性，对实际网络业务流量进行预测；并将实测数据的特点和卫星网络特征相结合，将星地融合网络重点场景分为广域数据智能精准感知、星上计算密集型业务快速处理、广域宽带视频业务文件高效分发三大类，并为研究提供模型和需求支撑。然后，针对星地多级边缘网络场景，提出星地控制群机制，设计出边缘—边缘协同模式和边缘—云集群协同模式，实现拥塞控制、计算卸载和具有鲁棒性的资源主动调度策略。最后，搭建了基于3GPP TR38.821、TR38.811等标准的OPNET系统级星地融合仿真平台，实现了大规模卫星互联网组网与星间链路仿真平台，可支持控制信令与传输数据解耦在NTN无线接入网部署场景下的架构流程、关键技术和整体解决方案，并提出了星地融合多链路传输方法、组播单播协作传输机制和服务功能链资源分配与映射方法，依托仿真平台进行了仿真评估和技术验证。依托本课题，共发表期刊论文4篇，会议论文1篇，出版专著1部，授权国内发明专利4项，3项专利得到成果转化，在运营商和企业中得到应用，获得麒麟科学技术奖和WCSP十周年十篇最佳会议论文奖。依托本课题，培养了博士生1名，硕士生6名。