面向多场景端到端性能可控的新型工业互联网体系架构

工业互联网是推动新一代信息技术与制造业深度融合的基石，其规模化推广和应用，需进一步强化智能工厂内现场级网络对生产和管理的核心支撑作用，提升网络端到端性能可控性，提高服务承载能力。如何在网络环境下实现感知分布性、传输适变性、管控协同性成为新的研究挑战。本项目围绕“按需互联的协同机理和调控机制”这一关键科学问题，采用“需求分析-功能定义-实施部署”的总体思路，提出包含现场级IIoT、边缘数据中台、工业云平台的新型工业互联网架构。攻克感传融合的动态组网和按需传输、多域资源协同优化与智能决策、基于数字孪生的网络管控等关键技术，形成“感-传-算-控”一体化技术方案。构建飞机装配制造工业互联网平台并开展多场景应用验证，确保动态大规模网络端到端延迟≤10ms，可靠性满足工业现场控制级要求。预期成果将为高端装备制造等典型行业多场景协同提供参考架构和设计方案。

The Industrial Internet is the cornerstone for the deep integration of a new generation of information technologies and manufacturing. Its large-scale promotions and applications require further strengthening of the core supporting role of the field-level network in the smart factory for production and management, and improving the end-to-end performance controllability of the network, promoting service carrying capacity. The problem of how to realize distributive sensing, adaptive transmission, and the cooperative management and control in the network environment has become a new research challenge. This project focuses on the key scientific issue of "on-demand interconnection coordination scheme and network regulation mechanism". Adopting the main idea of "demand analysis-function definition-implementation and deployment", we propose an Industrial Internet framework comprising field-level IIoT, edge data center and industrial cloud platform..This project aims to break through key technologies, such as dynamic networking and on-demand transmission based on the integration of sensory and transmission, collaborative optimization of multi-domain resources and intelligent decision-making, and digital twin-assisted network management and control to construct an integrated technical support system of "sensing-communication-computation-control". An Industrial Internet platform will be built for aircraft assembly manufacturing. Multi-scenario applications verification will be conducted to ensure that the end-to-end delay of the dynamic large-scale network is less or equal to 10ms, and the reliability meets the requirements of industrial field control level. The expected results will provide a reference architecture and design scheme for multi-scene collaboration in typical industries such as high-end equipment manufacturing.