摘要
在回顾历代无线接入网络架构演进、5G系统引入人工智能的进展和现状的基础上,指出基于AI内生的方式是未来网络架构设计的必由之路。未来6G时代,无线接入网络将向空天地一体化、用户超密集、为更多工业互联网场景服务等方向发展,目前业界已经提出了将服务化、智能和开放相关技术作为6G无线网络的具体使能技术,以应对进一步提升用户体验、降低网络能耗、拥抱更多新兴技术的挑战。为此,提出了一种基于AI内生的无线接入网络架构,包括智能面、控制面、用户面三个方面的功能,通过引入智能流、智能体的方法,实现了AI内生,并阐述了这个架构如何满足以用户为中心的无线组网需求。最后,提出未来基于AI内生的无线网络设计还需要应对ICDOT融合、通信感知算力融合等方面的挑战。
Reviewing the progress of wireless radio access network architectures and AI introduction in 5G system,AI-Native Radio Access Network(RAN)is an inexorable tread for RAN architecture evolution.In future 6G era,wireless access network should be applied to satellite-terrestrial integrated network,ultra-dense network,and diverse IIoT networks.Enabling technologies for 6G wireless network should provide service-based,intelligent,and open characteristics,so as to meet the challenges of improving user experience,reducing energy consumption,and accommodating upcoming new technologies.Therefore,AI-Native RAN Architecture which contains intelligence plane,control plane and user plane is prompted.The functions of the three planes,combined with intelligence streams and intelligence entities,will achieve AI-Native in the 6G RAN.Furthermore,how the AI-Native RAN architecture achieve the user centric network was elaborated.Finally,the design of AI-Native RAN should face more challenges including integration of ICDOT,integration of sensing,communication and computation,etc.
作者
谌丽
艾明
孙韶辉
CHEN Li;AI Ming;SUN Shaohui(CICT Mobile Communication Technology Co.,Ltd.,Beijing 100083,China)
出处
《无线电通信技术》
2022年第4期574-582,共9页
Radio Communications Technology
基金
国家重点研发计划(2020YFB1806803)。
关键词
无线接入网络架构
AI内生
用户为中心
频谱效率
radio access network architecture
AI native
user-centric
spectrum efficiency
