基于多域认知的空天信息网络智能拓扑构建机制基础研究

The contradiction between the requirement of high availability of aerospace information networks and the abominable environment in which they operate is becoming increasingly conspicuous. Dynamic topology control and optimization are important means to improve the ability of networking..The intelligent topology of aerospace information network is constructed on the basis of multi-dimensional cognition in this research which comprises three parts. First, the relationship between three cognitive domains including the environment of aerospace, network and service are analyzed, as well as their distribution law in the network. Service requirement and network support ability are clustered, abstracted and matched. The situation of network is predicted and evaluated through four steps which are cognitive information assimilating, mining, integrating and processing. Thus, the theoretic connections of the topology and the effective use of network resources are established. Second, the proposed topology control scheme takes network connectivity and system capacity into comprehensive consideration. Besides, the scheme combines automatic configuration and adaptive control with variable granularity, which can extremely approach optimal performance.Third, different adaptive topology optimization mechanism is studied according to diverse services. Dynamical adaptive topologies are generated according to all kinds of services respectively with consideration of multi-service demand differences and group user behavior convergence..With cooperation of the above three aspects, a topology control scheme which realizes closed loop control that consists of network sensing, cognizing and controlling is designed. The scheme can provide scientific reference for construction of an intelligent topology for aerospace information networks.

空天信息网需求与所处恶劣环境的矛盾日益突出，拓扑动态优化控制是提高网络能力的重要手段。.本项目以多域认知为基础，构建空天信息网络智能拓扑，研究内容包括：（1）研究空天环境、网络环境、业务环境三个认知域的关系及其在网络中的分布规律，对业务需求和网络提供的基础保障能力进行聚类、抽象和组配，通过认知信息的摄取、挖掘、融合、推理四个阶段，实行网络态势的预测、评估，建立拓扑结构与网络资源高效利用之间的理论关系；（2）研究兼顾网络连通与系统容量的拓扑优化控制机制，建立可变粒度拓扑自主优化和适变控制的联合设计方法，给出逼近其性能极限的拓扑控制方案；（3）针对多样化业务需求研究差异化自适应拓扑优化机理，主动利用多业务之间的需求差异性和用户群体行为的趋同性，分别建立针对多样化业务的动态自适应拓扑结构。三个方面相互协同，形成一套感知、认知、控制紧密耦合的拓扑控制机制，为构建空天信息网智能拓扑提供理论支持。

空天信息网所处的环境更加复杂多变，对网络结构产生巨大影响，应用的高可用性需求与所处恶劣环境及自身弱点的矛盾日益突出。因此，研究拓扑优化控制技术，通过动态行为来保持网络结构的相对稳定和可靠，是保证网络高效、不间断运行，提高空天信息系统损毁恢复和应急反应能力的重要手段。本项目包括三个紧密耦合、相互协同的研究内容：研究空天信息网多域认知机理，对网络状态及业务应用需求进行测算评估，建立拓扑结构与各种网络资源高效利用之间的理论关系与评价方法；研究兼顾网络连通与系统容量的拓扑协同优化控制机制，给出逼近其性能极限的拓扑优化配置方案；针对多样化业务需求研究差异化自适应拓扑优化机理，主动利用业务的差异性、以及节点群体趋同性等特征，构建业务自适应网络拓扑。目标是提高基于复杂多变环境下空天信息网的网络结构稳定性、可靠性、损毁恢复能力和应急反应能力而定，主要研究空天信息网智能网络拓扑优化、调整、控制机制关键问题的理论分析和方法研究，课题拟在空天网络高效优化、抗毁自愈、容错通信、拓扑协同优化等方面取得原创性理论成果和关键技术突破，为实际应用提供理论基础和技术支撑。