近日，福建省新发布6个项目为2025年福建省工业互联网技术创新重点攻关及产业化项目。至今，福建 省累计发布10个工业互联网技术创新重点攻关及产业化项目。名单显示，2025年福建省工业互联网技术 创新重点攻关及产业化项目包括：福建佰孟医学科技有限公司牵头实施的"融合AI视觉与分子诊断的禽 蛋安全品质智能检测平台研制"、福建火炬电子科技股份有限公司牵头实施的"基于人工智能和可信数据 空间的网络黑客主动甄别与联防封阻服务平台"、泉州信息工程学院牵头实施的"模块化智能柔性生产综 合管理系统研发及产业化"、联通（福建）产业互联网有限公司牵头实施的"基于纺纱生产全场景智能预 测与优化解决方案-纺纱智能体"、福建辅布司纺织有限公司牵头实施的"经编花边图案智能生成系统研 发及产业化"、福建龙钢新型材料有限公司牵头实施的"福建龙钢智慧能源管理平台"。 ...

Core Viewpoint - The "Application-Oriented Deterministic Data Bus Technology Blue Paper" focuses on the integration needs of "network-computation-control" in the 2.0 phase of the industrial internet, proposing a deterministic data bus architecture and technical system aimed at solving the challenges of data integration and collaboration across the entire industrial chain, thereby promoting the intelligent transformation of the manufacturing industry [1][21]. Group 1: Positioning and Significance - The industrial internet 2.0 phase requires autonomous intelligence across the design-manufacturing-supply "three chains," necessitating a "network-computation-control" integrated technology system to support the integration of data, information, and business across the entire chain [1][21]. - The current deterministic capabilities of the industrial internet are primarily ensured at the network layer, which does not adequately cover end-to-end application needs, thus necessitating the extension of deterministic capabilities from the network layer to the application layer [1][21]. - The application-oriented deterministic data bus serves as a key component of this system, acting as a link between networks, computing resources, and application businesses, facilitating cross-domain data integration and sharing, and driving the application of artificial intelligence throughout the industrial lifecycle [1][21]. Group 2: Architecture and Features - The data bus architecture consists of two layers: functional architecture and deployment architecture [2][34]. - Functional architecture includes four core modules: data service interface, data model, QoS management, and data distribution [2][34]. - Deployment architecture is installed in terminal and computing devices, providing a unified interface upwards and adapting to the underlying network downwards [2][34]. - Key features include: - "Network-computation-control" collaborative scheduling to ensure end-to-end determinism [2][41]. - Unified service interface that simplifies application development and migration [3][43]. - Compatibility with heterogeneous networks, allowing seamless access without user awareness of underlying differences [3][47]. - A unified data space based on standardized models and coding to facilitate IT/OT interoperability [4][48]. Group 3: Supporting Technologies - The data bus technology system is centered around data distribution and models, supported by multi-level technologies [5][50]. - Network-computation orchestration for global planning of network and computing resources [5][51]. - Standardized models and coding using OPC UA and semantic technologies to address interoperability issues among heterogeneous systems [6][50]. - Data distribution protocols supporting various communication needs [7][50]. - High-speed data channels to enhance processing efficiency and reduce latency [7][50]. - Adaptation to underlying networks through protocol conversion or direct adaptation [7][50]. Group 4: Application Scenarios - The data bus demonstrates multidimensional application potential in industrial scenarios: - Distributed computing for AI large models, supporting cross-device data collection and collaborative training [8][28]. - Edge-end collaborative detection, ensuring stable image transmission and optimizing real-time interaction between edge computing and terminal devices [9][28]. - Collaborative planning for design and production processes, breaking down data silos and enhancing data interaction efficiency [10][28]. - Large-scale multi-robot collaboration, supporting edge-cloud collaborative control and real-time communication [11][28]. Group 5: Future Outlook - The development of the data bus is still in its early stages, requiring industry consensus to unify interfaces and model standards [12][21]. - In the short term, focus can be placed on full-chain collaboration, multi-robot cooperation, and AI large model scenarios, gradually expanding from industry-specific applications to cross-industry applications [12][21].