面向多业务多速率的空间异构链路汇聚控制方法研究

Benefiting from the rapid development of inter-satellite link technology and the increasing processing power on the satellite, each space-based system has gradually evolved into a networked direction and tends to develop into an integrated spatial information network. However, the current space-based systems have significant differences in service QoS requirements, transmission rates, and switching modes, which pose challenges to efficient integration of heterogeneous networks. This project intends to address the issue of existing convergent node scheduling algorithms in terms of service heterogeneity and link heterogeneity in terms of low scheduling efficiency and poor service quality assurance capabilities, and conducts research on spatial heterogeneous link aggregation control methods for multi-service and multi-rate. Focus on breakthroughs in key technologies such as space business intelligence identification, link rate dynamic forecasting, and hierarchical scheduling of uplink services to fundamentally solve the problem that the existing global spatial information network has limited global coverage and weak network collaborative application capabilities, thereby greatly promoting spatial information network service capabilities.

受益于星间链路技术的长足发展和星上处理能力的日益增强，各天基系统已逐步向网络化方向演进，并趋于发展成为天地一体的空间信息网络。但是，当前各天基系统在业务QoS需求、传输速率、交换方式等方面存在显著差异，对各异构网络的高效融合提出挑战。本项目拟针对现有汇聚节点调度算法在业务异质和链路异构条件下调度效率低、服务质量保障能力差的问题，开展面向多业务多速率的空间异构链路汇聚控制方法研究，重点突破空间业务智能识别、链路速率动态预测、上行业务分级调度等关键技术，从根本上解决现有空间信息网络全域覆盖能力有限、网络协同应用能力弱的问题，从而推动空间信息网络服务能力的大幅提升。

受益于星间链路技术的长足发展和星上处理能力的日益增强，各天基系统已逐步向网络化方向演进，并趋于发展成为天地一体的空间信息网络。但是，当前各天基系统在业务QoS需求、传输速率、交换方式等方面存在显著差异，对各异构链路的高效汇聚提出了新的挑战。本项目对面向多业务多速率的空间异构链路汇聚控制方法展开研究，突破了空间信息网络任务智能感知与端-端可靠性分析方法、联合注意力机制与图神经网络的空间信息网络流量智能预测方法、基于大偏差理论的异构链路汇聚节点最小缓存的边界计算模型、面向多业务多链路的汇聚节点缓存队列管理与调度方法等关键技术，实现了多业务多链路条件下空间信息网络任务的精确感知、流量的精准预测、缓存的合理设计和队列的高效调度，为解决空间信息网络异构链路汇聚控制难、网络协同应用能力弱等问题提供了有力支撑。