移动通信2024，Vol.48Issue(7)：101-108,8.DOI:10.3969/j.issn.1006-1010.20240522-0002
算力网络驱动的语义通信
Computing Networks Enabled Semantic Communications
摘要
Abstract
Endogenous intelligence technology enhances network intelligence through the deep integration of network architecture and artificial intelligence technologies.As a core technology of 6G endogenous intelligent networks,semantic communication leverages AI to extract and compress semantic information from source data,significantly reducing transmission bandwidth consumption and improving communication reliability.However,the application of semantic communication requires substantial computation resources.This paper proposes a computing network-driven semantic communication system,which encompasses key technologies such as semantic sampling and semantic-channel joint coding driven by cloud-edge-device collaborative computing,as well as semantic-aware resource allocation and optimization.The system aims to provide sufficient computation resources for semantic processing and transmission,achieving deep integration of endogenous computing capability and communication.Experimental simulations demonstrate the advantages of the proposed computing network-driven semantic communication system.关键词
语义通信/算力网络/资源分配Key words
semantic communication/computing network/resource allocation分类
信息技术与安全科学引用本文复制引用
杨定熹,黄印桓,应憬恺,秦志金..算力网络驱动的语义通信[J].移动通信,2024,48(7):101-108,8.基金项目
国家重点研发计划"语义驱动的多模态网络全息通信关键技术"(2023YFB2904300) （2023YFB2904300）
国家自然科学基金项目"端到端语义信息传输方法","面向超高清视频的语义通信关键技术"(62293484,623B1012) （62293484,623B1012）
