边端协同关键技术、挑战及6G赋能展望*

doi:10.12267/j.issn.2096-5931.2026.10.005

信息通信技术与政策 ›› 2026, Vol. 52 ›› Issue (10): 40-46.doi: 10.12267/j.issn.2096-5931.2026.10.005

边端协同关键技术、挑战及6G赋能展望^*

Key technologies, challenges and prospects for 6G enabled edge-end collaboration

徐明枫, 李阳, 许伟华, 周吉喆, 江甲沫

中国信息通信研究院无线与移动通信研究所, 北京 100191

收稿日期:2026-08-26 出版日期:2026-10-25 发布日期:2026-10-28
通讯作者: 李阳
作者简介:
徐明枫,中国信息通信研究院无线与移动通信研究所工程师,主要研究方向为面向6G的无线AI、边端协同等;
许伟华,中国信息通信研究院无线与移动通信研究所工程师,主要研究方向为面向6G的无线AI、数字孪生和边端协同等;
周吉喆,中国信息通信研究院无线与移动通信研究所高级工程师,主要研究方向为6G网络、网络智能化技术和边端协同等;
江甲沫,中国信息通信研究院无线与移动通信研究所国际标准部主任,正高级工程师,主要研究方向为6G无线新技术及其测试验证等
基金资助:
^*国家科技重大专项项目(2025ZD1302800);北京市自然科学基金项目(L253004)

XU Mingfeng, LI Yang, XU Weihua, ZHOU Jizhe, JIANG Jiamo

Wireless and Mobile Communications Research Institude, China Academy of Information and Communications Technology, Beijing 100191, China

Received:2026-08-26 Online:2026-10-25 Published:2026-10-28
Contact: LI Yang

摘要/Abstract

摘要：

边端协同依托算力下沉、模型分层部署、终端协同调度和资源统一管控等优势,能够有效支撑算力受限的新一代智能终端完成复杂智能化协同作业。系统梳理了边端协同领域的研究现状,重点围绕模型拆分、终端模型轻量化、通算资源联合管控等技术,梳理该领域现存“瓶颈”和面临的挑战。在此基础上,结合算网融合、通感一体化、无线智能化等6G关键技术,展望6G赋能边端协同的演进方向,为边端协同后续研究提供参考。

关键词: 边端协同, 6G, 人工智能

Abstract:

Edge-end collaboration, leveraging the advantages of computing power sinking, layered model deployment, terminal cooperative scheduling and unified resource management, can effectively support new generation smart terminals with limited computing resources to accomplish complex intelligent collaborative tasks. This paper systematically reviews the research in the field of edge-end collaboration. It focuses on core technologies including model splitting, terminal-side model lightweighting, and joint communication-computing resource management, and summarizes the existing bottlenecks and research challenges in this domain. On this basis, combined with key 6G technologies such as computing-network integration, integrated sensing and communication, and wireless intelligence, the evolution directions of edge-end collaboration empowered by 6G are prospected, so as to provide references for followup research on edge-end collaboration.

Key words: edge-end collaboration, 6G, AI

中图分类号:

TN929.5

徐明枫, 李阳, 许伟华, 周吉喆, 江甲沫. 边端协同关键技术、挑战及6G赋能展望^*[J]. 信息通信技术与政策, 2026, 52(10): 40-46.

XU Mingfeng, LI Yang, XU Weihua, ZHOU Jizhe, JIANG Jiamo. Key technologies, challenges and prospects for 6G enabled edge-end collaboration[J]. Information and Communications Technology and Policy, 2026, 52(10): 40-46.

导出引用管理器 EndNote|Ris|BibTeX

链接本文:

http://ictp.caict.ac.cn/CN/10.12267/j.issn.2096-5931.2026.10.005

http://ictp.caict.ac.cn/CN/Y2026/V52/I10/40

图/表 3

参考文献 32

[1]	IMT-2030(6G)推进组. 6G移动算力网络需求、关键技术及架构研究[R]. 2025-11.
[2]	中国信息通信研究院. 新一代智能终端蓝皮书:从“人工智能+终端”到人工智能终端:No.202520[R]. 2026-03.
[3]	黄文康, 姜宁, 贺家乐, 等. 面向边端协同推理的模型分割与无线资源分配方法研究[J]. 移动通信, 2026, 50(8): 41-55.
[4]	徐明枫, 李阳, 许伟华, 等. AI空口传输技术赋能边端协同的思考[J]. 移动通信, 2026, 50(8): 23-30.
[5]	华为终端有限公司. AI终端白皮书[R]. 2025.
[6]	王志勤, 周吉喆, 韩凯峰. 端边协同的6G内生AI网络[J]. 中兴通讯技术, 2025, 31(4): 29-33.
[7]	Mao Y, You C, Zhang J, et al. A survey on mobile edge computing: the communication perspective[J]. IEEE Communications Surveys & Tutorials, 2017, 19(4): 2322-2358.
[8]	Yao Y, Hou J, Wu G, et al. SecoInfer: secure DNN end-edge collaborative inference framework optimizing privacy and latency[J]. ACM Transactions on Sensor Networks, 2024, 20(6): 1-29.
[9]	Taufique Z, Miele A, Lijeberg P, et al. Adaptive workload distribution for accuracy-aware DNN inference on collaborative edge platforms[C]// 2024 29th Asia and South Pacific Design Automation Conference. Incheon, South Korea: IEEE, 2024: 109-114.
[10]	Sun H, Chen X, Qian Z, et al. BIRP: batch-aware inference workload redistribution and parallel scheme for edge collaboration[C]// 52nd International Conference on Parallel Processing. Salt Lake City, USA: ACM, 2023: 72-81.
[11]	Bin K, Park J, Park C, et al. CoActo: coactive neural network inference offloading with fine-grained and concurrent execution[C]// 22nd Annual International Conference on Mobile Systems, Applications and Services. Tokyo, Japan: ACM, 2024: 412-424.
[12]	Hu C, Bao W, Wang D, et al. Dynamic adaptive DNN surgery for inference acceleration on the edge[C]// IEEE INFOCOM 2019-IEEE Conference on Computer Communications. Paris, France: IEEE, 2019: 1423-1431.
[13]	Chen D, Dong Q, He Q, et al. Deep reinforcement learning for collaborative inference in local-edge computing[C]// 2024 International Academic Conference on Edge Computing, Parallel and Distributed Computing. Xi’an, China: IEEE, 2024: 138-142.
[14]	Li Y, Liu Z, Kou Z, et al. Real-time adaptive partition and resource allocation for multi-user end-cloud inference collaboration in mobile environment[J]. IEEE Transactions on Mobile Computing, 2024, 23(12): 13076-13094. doi: 10.1109/TMC.2024.3430103 URL
[15]	Liu Z, Tian M, Dong M, et al. Mobility-aware edge inference based on model partition and service migration[J]. IEEE Transactions on Mobile Computing, 2024, 23(10): 9437-9450. doi: 10.1109/TMC.2024.3366186 URL
[16]	Fan W, Chen Z, Hao Z, et al. Joint task offloading and resource allocation for quality-aware edge-assisted machine learning task inference[J]. IEEE Transactions on Vehicular Technology, 2023, 72(5): 6739-6752. doi: 10.1109/TVT.2023.3235520 URL
[17]	Huang J, Wu J, Wu Y, et al. Joint request offloading and resource allocation for long-term utility optimization in collaborative edge inference with time-coupled resources[J]. IEEE Transactions on Network Science and Engineering, 2025, 12(4): 2622-2639. doi: 10.1109/TNSE.2025.3551148 URL
[18]	Liu X, Song Y, Li X, et al. Efficient partitioning vision transformer on edge devices for distributed inference[C]// 2025 IEEE 45th International Conference on Distributed Computing Systems (ICDCS). Glasgow, UK: IEEE, 2025: 286-296.
[19]	Liu X, Xue N, Bao R, et al. CSGO: generalized optimization for cold start in wireless collaborative edge LLM systems[J]. Journal of Communications and Information Networks, 2025, 10(4): 340-351. doi: 10.23919/JCIN.2025.11357495 URL
[20]	Wang X, Yu F, Dou Z, et al. SkipNet: learning dynamic routing in convolutional networks[C]// European Conference on Computer Vision (ECCV). Munich, Germany: Cham, 2018: 409-424.
[21]	Han S, Mao H, Dally W J. Deep compression: compressing deep neural networks with pruning, trained quantization and Huffman coding[PP/OL]. arXiv(2016-02-15)[2026-04-30]. https://arxiv.org/abs/1510.00149.
[22]	Hu P, Peng X, Zhu H, et al. OPQ: compressing deep neural networks with one-shot pruning-quantization[C]// AAAI Conference on Artificial Intelligence. New Orleans, USA: AAAI Press, 2021, 35(9): 7780-7788.
[23]	刘强, 钟荣锋. 基于多任务学习与知识蒸馏的工业大模型轻量化部署优化[J]. 无线互联科技, 2026, 23(14): 87-90+99.
[24]	丁男, 王佳佳, 冀承慧, 等. 边-端协作下基于早期退出机制的深度神经网络动态自适应分区[J]. 电子与信息学报, 2025, 47(10): 4005-4017.
[25]	Johnsen T K, Harshbarger I, Xia Z, et al. NaviSplit: dynamic multi-branch split DNNs for efficient distributed autonomous navigation[C]// 2024 IEEE 25th International Symposium on a World of Wireless, Mobile and Multimedia Networks (WoWMoM). Perth, Australia: IEEE, 2024: 196-201.
[26]	於志文, 陈柯帆, 刘思聪, 等. 群智驱动的多终端轻量化智能计算模型协同演化框架[J]. 微电子学与计算机, 2025, 42(10): 9-27.
[27]	Tian Y, Zhang Z, Yang Y, et al. An edge-cloud collaboration framework for generative AI service provision with synergetic big cloud model and small edge models[J]. IEEE Network, 2024, 38(5): 37-46. doi: 10.1109/MNET.2024.3420755 URL
[28]	Liu Z, Tian M, Dong M, et al. MoEI: mobility-aware edge inference based on model partition and service migration[J]. IEEE Transactions on Mobile Computing, 2024, 23(10): 9437-9450. doi: 10.1109/TMC.2024.3366186 URL
[29]	李嘉盛, 姜宁, 贺家乐, 等. 基于边端协同的多用户通信计算资源联合管理方法研究[J]. 移动通信, 2025, 49(3): 78-85.
[30]	金石, 刘凡. 通感算智融合技术与应用的发展思考[J]. 信息通信技术, 2026, 20(2): 4-6.
[31]	赵亚鹏, 冯大权, 程义超. 通感算一体化:关键技术与架构演进[J]. 移动通信, 2026, 50(6): 13-17+27.
[32]	李阳, 崔厚虓, 蔡晓雄, 等. 6G物理层AI赋能技术的思考[J]. 移动通信, 2024, 48(8): 25-29.

边端协同关键技术、挑战及6G赋能展望^*

Key technologies, challenges and prospects for 6G enabled edge-end collaboration

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 3

参考文献 32

相关文章 15

编辑推荐

Metrics

本文评价

[1]	王燕敏, 冯伟, 董超, 任保全, 段向阳. 面向6G无人作业的“感-传-算-控”闭环:模型、度量与网络设计^*[J]. 信息通信技术与政策, 2026, 52(9): 2-8.
[2]	刘光毅, 贺智敏, 王新尧, 蔡青, 王凯悦, 陈天骄, 金婧. 6G通信-计算-智能融合赋能具身智能应用[J]. 信息通信技术与政策, 2026, 52(9): 29-37.
[3]	刘军, 董霁, 王威. 基于风险管控视角的终端智能体安全防护与治理路径研究[J]. 信息通信技术与政策, 2026, 52(8): 18-22.
[4]	聂山棚, 吴荻, 杜敦伟. 面向输电安全的AI终端异物检测技术与差异化部署策略研究[J]. 信息通信技术与政策, 2026, 52(8): 59-66.
[5]	董广智, 吴寒冰, 王瑞, 由佳, 王景尧. 智慧教育场景大模型应用安全风险与分层全域防护机制研究[J]. 信息通信技术与政策, 2026, 52(8): 67-73.
[6]	郭晓彤, 谢润彬, 吴迪, 夏鹏. 面向离散行业智能工厂的工业大模型技术路径与体系架构研究[J]. 信息通信技术与政策, 2026, 52(8): 91-96.
[7]	姜春宇, 艾博焕, 郭彦美, 崔一妍, 尹正. 面向AI的数据工程体系构建:基于AI-ready DataOps的范式转型[J]. 信息通信技术与政策, 2026, 52(6): 86-96.
[8]	白玉真, 杨靖世, 袁博, 王思佳, 贾轩, 王思源. 企业高质量数据集建设运营路径研究[J]. 信息通信技术与政策, 2026, 52(5): 10-15.
[9]	康绍莉, 王可, 秦海超, 冯云九, 王丹, 杨祺欣. 面向6G的低空智能网联系统用例研究[J]. 信息通信技术与政策, 2026, 52(4): 2-14.
[10]	毕然, 李枫秋, 葛坚, 黄雍涛. 智能终端AI计算技术的路径与挑战[J]. 信息通信技术与政策, 2026, 52(3): 14-21.
[11]	程莹, 刘志鹏, 张萌. 终端智能体发展路径、潜在影响与治理对策研究[J]. 信息通信技术与政策, 2026, 52(3): 2-7.
[12]	赵伟博, 桑柳, 陈锐豪, 苏越, 马飞. 人工智能时代高性能网络发展现状与趋势[J]. 信息通信技术与政策, 2026, 52(2): 30-35.
[13]	杨一禾, 马飞, 李哲, 董晓慧. 算力互联网赋能生成式AI视频的内容安全研究[J]. 信息通信技术与政策, 2026, 52(2): 69-77.
[14]	郭振铭, 陈力, 陈锋. 6G时代下AI赋能的无线定位技术^*[J]. 信息通信技术与政策, 2026, 52(10): 16-25.
[15]	张晶, 王焕丽, 崔璨, 黎卓芳, 杜刘通. 6G网络与AI双向赋能标准化进展研究^*[J]. 信息通信技术与政策, 2026, 52(10): 2-8.