面向6G通感算深度融合的云-网-边智能协同  

Cloud-Network-Edge Intelligent Collaboration for the Deep Integration of 6G Communication, Sensing and Computing

作　　者：林妙预 李因新 赵力强[1,2] 王云峰 刘洋 LIN Miaoyu;LI Yinxin;ZHAO Liqiang;WANG Yunfeng;LIU Yang(School of Telecommunications Engineering,Xidian University,Xi'an 710071,China;Guangzhou Institute of Technology,Xidian University,Guangzhou 510100,China;China Telecom Research Institute,Beying 102209,China)

机构地区：[1]西安电子科技大学通信工程学院,陕西西安710071 [2]西安电子科技大学广州研究院,广东广州510100 [3]中国电信研究院,北京102209

出　　处：《移动通信》2023年第3期62-66,共5页Mobile Communications

基　　金：广东省重点领域研究开发计划(2020B01120003)。

摘　　要：通信-感知-计算(通感算)深度融合对6G网络端到端传输时延、可靠性等通信指标提出了极高的要求,因此提出一种云-网-边智能协同方案。首先,将云-网-边智能协同系统抽象为“三域四层三面”分布式体系结构,包括:计算域、感知域、通信域;基础设施层、虚拟化层、功能层、应用层;智能面、控制面、管理编排面。然后,通过云-网-边协同训练与推理,以支撑通感算智能协同。最后,搭建面向6G通感算深度融合的云-网-边智能协同实验平台。实验结果表明:云-网-边智能协同可提高系统可靠性,降低传输负载和业务时延。The deep integration of communication, sensing, and computing raises high requirements for the 6G network, such as end-to-end transmission delay and reliability. Therefore, a cloud-network-edge(CNE) intelligent collaboration scheme is proposed in this paper. Firstly, the CNE intelligent cooperative system is abstracted into a distributed architecture of “three domains, four layers and three facets”, including the computing domain, sensing domain, and communication domain;infrastructure layer, virtualization layer, function layer, and application layer;intelligent surface, control surface, and management orchestration surface. Then, through CNE collaborative training and reasoning, communication-sensing-computing intelligence collaboration can be enabled. Finally, a CNE intelligent collaborative experimental platform for deep integration of 6G communication, sensing, and computing is built. The experimental results show that CNE intelligent collaboration can improve system reliability and reduce transmission load and service delay.

关 键 词：通感算深度融合 云-网-边 智能协同 

分 类 号：TN919.1[电子电信—通信与信息系统]