首页|期刊导航|移动通信|AI空口传输技术赋能边端协同的思考

AI空口传输技术赋能边端协同的思考OA

Thoughts on Empowering Edge-End Collaboration via AI-Enabled Air-Interface Transmission Technology

中文摘要英文摘要

随着终端智能化的不断演进,业务需求也从传统的通信服务向通算一体化服务升级,边端协同依托就近算力调度和低时延交互等优势,成为保障终端智能业务高效运行、优化用户服务体验的关键支撑技术.首先介绍了边端协同的关键技术和AI空口传输关键技术;然后针对边端协同中数据传输易受噪声干扰导致协同低效的问题,研究了AI智能空口传输技术赋能边端协同的可行性和有效性.仿真结果表明,在边缘侧部署AI解调模型,能够有效降低端侧上传数据特征图在低信噪比传输环境下的误码率,从而提升边端协同中边侧执行后续推理的性能,适配边端协同场景下的可靠数据传输需求,为后续边端协同系统设计提供参考.

With the continuous evolution of terminal intelligence,service requirements are upgrading from conventional communication services to integrated communication-computing services.Benefiting from proximity computing resource scheduling and low-latency data interaction,edge-end collaboration has become a pivotal enabling technology for guaranteeing efficient execution of terminal intelligent services and improving user experience.This paper first introduces key technologies of edge-end collaboration and artificial intelligence(AI)-enabled air-interface transmission.Then,to address the problem of inefficient collaboration incurred by data transmission susceptible to noise and interference in edge-device collaboration,this paper investigates the feasibility and effectiveness of empowering edge-end collaboration via AI-driven intelligent air interface transmission technologies.Simulation results verify that deploying an AI demodulation model at the edge node can substantially reduce the bit error rate of feature maps uploaded from terminals under low signal-to-noise ratio transmission environments.Accordingly,the inference performance at the edge is improved,satisfying the reliable transmission requirements in edge-end collaborative scenarios and offering instructive references for the future design of edge-end collaborative systems.

徐明枫;李阳;许伟华;周吉喆

中国信息通信研究院,北京 100191中国信息通信研究院,北京 100191中国信息通信研究院,北京 100191中国信息通信研究院,北京 100191

信息技术与安全科学

6GAI空口传输边端协同

6Gartificial intelligenceair-interface transmissionedge-end collaboration

《移动通信》 2026 (8)

23-30,8

北京市自然科学基金项目"面向6G多形态终端差异化能力需求的智能协同关键技术研究与验证"(L253004) 本研究受中国信息通信研究院青年课题基金资助(QN-2026-021).

10.3969/j.issn.1006-1010.20260605-0001

评论