6G无线接入网通算智融合关键技术与标准化思考

发布时间:2026-03-13 作者:解宇瑄,李响,李婷,孙奇

摘要:生成式人工智能等AI技术正推动6G网络从连接管道向智能服务平台加速演进,通信、计算与智能的深度融合已成为无线接入网(RAN)发展的关键方向。针对6G RAN通算智融合的需求与挑战,提出了由基础设施层、网络功能层与编排服务层三层组成的通算智融合架构体系。该体系突破了以单一连接为核心的传统范式:在基础设施层,实现异构智算硬件资源的统一管理与边缘基站算力的强实时高效能调度;在网络功能层,支持原生多维数据精细管控与高效传输、AI计算及AI模型全生命周期管理等核心功能;在编排服务层,构建面向差异化业务的通算智联合编排与服务开放机制,从而实现通算智多维资源的高效协同与多元服务的定制化保障。在此基础上,进一步探讨了6G无线通算智融合的核心使能技术和标准化路径,旨在通过重构RAN的基础能力与服务形态,为实现6G泛在智联的愿景提供坚实的技术支撑与系统基座。

关键词:6G;无线接入网;通算智融合;标准化

 

Abstract: To address the requirements and challenges of this convergence in 6G RAN, this paper proposes a three-layer integrated architecture consisting of an infrastructure layer, a network function layer, and an orchestration and service layer. This architecture moves beyond the traditional connectivity-centric model. At the infrastructure layer, it enables unified management of heterogeneous intelligent computing hardware and real-time, high-efficiency scheduling of computing resources at edge base stations. The network function layer natively supports fine-grained multi‑dimensional data management and efficient transmission, AI computation, and full lifecycle management of AI models. At the orchestration and service layer, it establishes a joint orchestration and service exposure mechanism tailored to diverse service requirements. In this way, the architecture achieves efficient coordination of multi‑dimensional resources and customized provisioning of differentiated services. The paper further explores key enabling technologies and potential standardization pathways for communication‑computing‑intelligence convergence in 6G. By reshaping the foundational capabilities and service paradigms of the RAN, solid technical support and a systematic foundation are provided for realizing the vision of ubiquitous intelligent connectivity in 6G.

Keywords: 6G; RAN; communication-computing-intelligence convergence; standardization