边缘智能计算的协同推理关键技术研究

张智涵; 张天魁; 党梅梅; 程伟强

doi:10.11959/j.issn.1000-0801.DXKX260099

您当前的位置：

首页 >

文章列表页 >

边缘智能计算的协同推理关键技术研究

更新时间：2026-05-19

- 边缘智能计算的协同推理关键技术研究
- Key Technologies for Collaborative Inference in Edge Intelligence Computing
- 电信科学 2026年
- 作者机构：
  
  1.北京邮电大学信息与通信工程学院，北京 100876
  2.中国信息通信研究院技术与标准研究所，北京 100191
  3.中国移动通信有限公司研究院，北京 100053
- 作者简介：
  
  张天魁，zhangtiankui@bupt.edu.cn
- 基金信息：
  
  国家自然科学基金资助项目(62371068);鹏城实验室科教基金会-中国移动科创基金项目(SQ2510300009)
- DOI：10.11959/j.issn.1000-0801.DXKX260099
  中图分类号： TP393
- 收稿：2026-02-10，
  
  修回：2026-04-27，
  
  录用：2026-05-18，
- 稿件说明：
移动端阅览
张智涵, 张天魁, 党梅梅, 等. 边缘智能计算的协同推理关键技术研究[J/OL]. 电信科学, 2026.

ZHANG Zhihan, ZHANG Tiankui, DANG Meimei, et al. Key Technologies for Collaborative Inference in Edge Intelligence Computing[J/OL]. Telecommunications Science, 2026.
张智涵, 张天魁, 党梅梅, 等. 边缘智能计算的协同推理关键技术研究[J/OL]. 电信科学, 2026. DOI： 10.11959/j.issn.1000-0801.DXKX260099.

ZHANG Zhihan, ZHANG Tiankui, DANG Meimei, et al. Key Technologies for Collaborative Inference in Edge Intelligence Computing[J/OL]. Telecommunications Science, 2026. DOI： 10.11959/j.issn.1000-0801.DXKX260099.

摘要

本文剖析了边缘智能计算技术在边缘计算网络中的实现逻辑，并围绕边缘协同推理任务的推理时延与推理质量需求，分别从判别式与生成式人工智能模型的特性出发，系统分析了边缘协同推理关键技术的实现方法。在此基础上，提出了一种面向生成式人工智能模型的边端协同推理系统设计实例，验证了其在推理时延与推理质量方面的优势，为边缘智能计算在边缘计算网络中高效部署与应用实现提供了可行的技术方案。

Abstract

This paper provided an in-depth analysis of the implementation logic of edge collaborative intelligent computing technologies in edge computing networks. Focusing on the latency and inference quality requirements of edge collaborative inference tasks

it systematically analyzed the characteristics of discriminative and generative artificial intelligence models

respectively

and summarized the core implementation approaches of key technologies for edge collaborative inference. On this basis

a device–edge collaborative inference system design tailored for generative artificial intelligence models was proposed

and its advantages in terms of inference latency and inference quality were validated. The proposed approach provided a technical reference for the efficient deployment and practical implementation of edge collaborative intelligent computing technologies.

关键词

Keywords

references

朱文武 . 多媒体智能计算若干研究进展 [J ] . 中国科学: 信息科学 , 2025 , 55 ( 09 ): 2153 - 2164 .

郑远鹏 , 张天魁 , 庞博 , 等 . 面向边缘智能的移动算力网络关键技术研究 [J ] . 邮电设计技术 , 2023 , ( 05 ): 88 - 92 .

WANG C X , YOU X , GAO X , et al . On the road to 6G: visions, requirements, key technologies, and testbeds [J ] . IEEE Communications Surveys & Tutorials , 2023 , 25 ( 2 ): 905 - 974 . DOI: 10.1109/COMST.2023.3249835 http://dx.doi.org/10.1109/COMST.2023.3249835 .

卢先领 , 李德康 . 面向大规模多接入边缘计算场景的任务卸载算法 [J ] . 电子与信息学报 , 2025 , 47 ( 1 ): 116 - 127 . DOI: 10.11999/JEIT240624 http://dx.doi.org/10.11999/JEIT240624 .

林鹏 , 黄新梁 , 宁兆龙 , 等 . 多指标意图驱动的无人机计算卸载与轨迹规划自适应优化策略 [J ] . 通信学报 , 2026 , 47 ( 3 ): 195 - 208 . DOI: 10.11959/j.issn.1000-436x.2026056 http://dx.doi.org/10.11959/j.issn.1000-436x.2026056 .

陈乐 , 马彰超 , 董芃 , 等 . 面向工业智能体互联网的“通信-控制”协同原子化重构机制 [J ] . 通信学报 , 2026 , 47 ( 3 ): 42 - 63 . DOI: 10.11959/j.issn.1000-436x.2026053 http://dx.doi.org/10.11959/j.issn.1000-436x.2026053 .

WANG Y , YANG C , LAN S , et al . End-edge-cloud collaborative computing for deep learning: a comprehensive survey [J ] . IEEE Communications Surveys & Tutorials , 2024 , 26 ( 4 ): 2647 - 2683 . DOI: 10.1109/COMST.2024.3393230 http://dx.doi.org/10.1109/COMST.2024.3393230 .

牛涛 . 面向边缘智能的计算加速研究 [D ] . 北京 : 北京邮电大学 , 2024 . DOI: 10.26969/d.cnki.gbydu.2024.000146 http://dx.doi.org/10.26969/d.cnki.gbydu.2024.000146 .

傅文军 , 谭伟 , 胡露航 . 从判别式人工智能到生成式人工智能的演进逻辑及场景策略研究 [J ] . 中国仪器仪表 , 2024 ( 10 ): 17 - 21 .

BANH L , STROBEL G . Generative artificial intelligence [J ] . Electronic Markets , 2023 , 33 ( 1 ): 63 . DOI: 10.1007/s12525-023-00680-1 http://dx.doi.org/10.1007/s12525-023-00680-1 .

DAI P , HAN B , LI K , et al . Joint optimization of device placement and model partitioning for cooperative DNN inference in heterogeneous edge computing [J ] . IEEE Transactions on Mobile Computing , 2025 , 24 ( 1 ): 210 - 226 . DOI: 10.1109/TMC.2024.3457793 http://dx.doi.org/10.1109/TMC.2024.3457793 .

XIAO X , ZHANG J , WANG W , et al . DNN-driven compressive offloading for edge-assisted semantic video segmentation [C ] // IEEE INFOCOM 2022-IEEE Conference on Computer Communications . London, United Kingdom : IEEE , 2022 : 1888 - 1897 .

HAO Z , XU G , LUO Y , et al . Multi-agent collaborative inference via DNN decoupling: intermediate feature compression and edge learning [J ] . IEEE Transactions on Mobile Computing , 2023 , 22 ( 10 ): 6041 - 6055 .

PACHECO R G , SHIFRIN M , COUTO R S , et al . AdaEE: adaptive early-exit DNN inference through multi-armed bandits [C ] // ICC 2023-IEEE International Conference on Communications . Rome, Italy : IEEE , 2023 : 3726 - 3731 .

ZHONG R , MU X , ZHANG Y , et al . Mobile edge generation: a new era to 6G [J ] . IEEE Network , 2024 , 38 ( 5 ): 47 - 55 . DOI: 10.1109/MNET.2024.3420240 http://dx.doi.org/10.1109/MNET.2024.3420240 .

LUO H , et al . Toward edge general intelligence with multiple-large language models (Multi-LLM): architecture, trust, and orchestration [J ] . IEEE Transactions on Cognitive Communications and Networking , 2025 , 11 ( 6 ): 3563 - 3585 .

HO J , JAIN A , ABBEEL P . Denoising diffusion probabilistic models [C ] // Proceedings of the 34th International Conference on Neural Information Processing Systems . Red Hook, NY, USA : Curran Associates Inc. , 2020 : 6840 - 6851 .

GAO S , YANG P , KONG Y , et al . Characterizing and scheduling of diffusion process for text-to-image generation in edge networks [J ] . IEEE Transactions on Mobile Computing , 2025 , 24 ( 10 ): 11137 - 11150 . DOI: 10.1109/TMC.2025.3574065 http://dx.doi.org/10.1109/TMC.2025.3574065 .

HO J , SAHARIA C , CHAN W , et al . Cascaded diffusion models for high fidelity image generation [J ] . Journal of Machine Learning Research , 2022 , 23 ( 1 ): 1 - 33 .

PERNIAS P , RAMPAS D , RICHTER M L , et al . Wuerstchen: an efficient architecture for large-scale text-to-image diffusion models [C ] // Proceedings of the 2024 International Conference on Learning Representations . Vienna, Austria : OpenReview.net , 2024 .

李松 , 王新荣 , 王博文 , 等 . 基于随机网络演算的车联网边缘计算多跳任务卸载性能分析 [J ] . 电子与信息学报 , 2023 , 45 ( 7 ): 2459 - 2466 .

CUI P , HAN S , LI L , et al . Roundtrip interaction delay analysis of immersive communications: a stochastic network calculus perspective [J ] . IEEE Transactions on Wireless Communications , 2025 , 24 ( 3 ): 2188 - 2202 . DOI: 10.1109/TWC.2024.3518589 http://dx.doi.org/10.1109/TWC.2024.3518589 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

面向5G-Advanced与6G的智能无线电接入网：关键标准技术与未来演进

数字孪生技术赋能6G网络

OSI-RM视角下的移动通信网络AI技术应用综述

通感融算赋能的空天地一体化网络架构及关键技术

云化的智算中心万卡集群创新与实践