面向卫星无人机协同网络的NOMA辅助流水线并行推理研究

杨萌琪; 吴义曼; 周任飞; 吴鹏; 潘鹏; 郑长亮

doi:10.11959/j.issn.1000-0801.DXKX260294

Chinese

您当前的位置：

首页 >

文章列表页 >

面向卫星无人机协同网络的NOMA辅助流水线并行推理研究

专题:低空智联网 | 更新时间：2026-06-23

- 面向卫星无人机协同网络的NOMA辅助流水线并行推理研究
- NOMA-Assisted Pipeline Parallel Inference for Satellite-UAV Collaborative Networks
- 电信科学 2026年
- 作者机构：
  
  1.杭州电子科技大学通信工程学院，浙江杭州 310018
  2.中国人民解放军31007部队，北京 100079
  3.中国人民解放军军事科学院国防科技创新研究院，北京 100071
- 作者简介：
  
  吴鹏，hitwupeng@outlook.com
- 基金信息：
  
  国家自然科学基金项目(62401612)
- DOI：10.11959/j.issn.1000-0801.DXKX260294
  中图分类号： TN927
- 收稿：2026-05-09，
  
  修回：2026-05-31，
  
  录用：2026-06-23，
- 稿件说明：
移动端阅览
杨萌琪, 吴义曼, 周任飞, 等. 面向卫星无人机协同网络的NOMA辅助流水线并行推理研究[J/OL]. 电信科学, 2026.

YANG Mengqi, WU Yiman, ZHOU Renfei, et al. NOMA-Assisted Pipeline Parallel Inference for Satellite-UAV Collaborative Networks[J/OL]. Telecommunications Science, 2026.
杨萌琪, 吴义曼, 周任飞, 等. 面向卫星无人机协同网络的NOMA辅助流水线并行推理研究[J/OL]. 电信科学, 2026. DOI： 10.11959/j.issn.1000-0801.DXKX260294.

YANG Mengqi, WU Yiman, ZHOU Renfei, et al. NOMA-Assisted Pipeline Parallel Inference for Satellite-UAV Collaborative Networks[J/OL]. Telecommunications Science, 2026. DOI： 10.11959/j.issn.1000-0801.DXKX260294.

摘要

低轨卫星与无人机协同推理为实时智能感知任务提供了新的解决途径。针对现有协同计算方法中任务卸载或非流水线模型切分易造成通信与计算相互等待、资源利用效率较低的问题，本文提出一种结合模型切分与流水线并行的卫星无人机协同推理方案。该方案对深度神经网络进行切分，其中头部和尾部部署于无人机侧，计算密集的中间子网络部署于卫星侧，并通过中间特征传递实现完整推理功能。在此基础上，所提方案引入微批次流水线并行机制与非正交多址接入（non-orthogonal multiple access，NOMA）技术，使不同微批次的计算和传输阶段并行执行。进一步，以端到端协同推理时延最小化为目标，联合优化了神经网络部署结构、微批次划分粒度及上下行传输功率。仿真结果表明，所提方案能够有效降低协同推理时延，提升系统资源利用率。

Abstract

Collaborative inference between low-Earth-orbit (LEO) satellites and unmanned aerial vehicles (UAVs) provides a new solution for real-time intelligent sensing tasks. Existing collaborative computing methods based on task offloading or non-pipelined model partitioning tend to cause mutual waiting between communication and computation

resulting in low resource efficiency. To address this problem

a satellite-UAV collaborative inference scheme combining model splitting and pipeline parallelism was proposed. In this scheme

the deep neural network was partitioned

with the head and tail subnetworks deployed on the UAVs and the computation-intensive intermediate subnetwork deployed on the satellite. Complete inference was achieved through intermediate feature transmission. On this basis

the micro-batch pipeline parallel mechanism and non-orthogonal transmission were employed

allowing the computation and transmission stages of different micro-batches to be executed in parallel. Furthermore

with the objective of minimizing the end-to-end collaborative inference delay

the neural network deployment structure

micro-batch partitioning granularity

and uplink and downlink transmission power were jointly optimized. Simulation results showed that the proposed scheme could effectively reduce the collaborative inference delay and improve system resource utilization.

关键词

Keywords

references

BAENA E , TESTOLINA P , POLESE M , et al . Space-O-RAN: enabling intelligent, open, and interoperable non-terrestrial networks in 6G [J ] . IEEE Communications Magazine , 2026 , 64 ( 2 ): 112 - 118 .

FONTANESI G , ORTIZ GOMEZ F G , LAGUNAS E , et al . Artificial intelligence for satellite communication: a survey [J ] . IEEE Communications Surveys & Tutorials , 2026 , 28 : 1381 - 1435 .

FAN G , WU P , YANG M , et al . Internet of satellites (IoS) for intelligent satellite cluster: applications, methods, and challenges [J ] . Engineering , 2025 , 54 ( 11 ): 155 - 170 .

WEN D , ZHANG S , ZHU G , et al . UAV-assisted edge inference with integrated sensing, communication, and computation [J ] . IEEE Transactions on Wireless Communications , 2026 , 25 : 13336 - 13350 .

HUANG J , LIAN L , WEN D , et al . Dynamic UAV-assisted cooperative edge AI inference [J ] . IEEE Transactions on Wireless Communications , 2025 , 24 ( 1 ): 615 - 628 .

LIN Y , FENG W , ZHOU T , et al . Integrating satellites and mobile edge computing for 6G wide-area edge intelligence: minimal structures and systematic thinking [J ] . IEEE Network , 2023 , 37 ( 2 ): 14 - 21 .

YUAN S , PENG M , SUN Y . Satellite-terrestrial integrated fog networks: architecture, technologies, and challenges [J ] . IEEE Wireless Communications , 2025 , 32 ( 4 ): 208 - 215 .

FAN W , MENG Q , WANG G , et al . Satellite edge intelligence: DRL-based resource management for task inference in LEO-based satellite-ground collaborative networks [J ] . IEEE Transactions on Mobile Computing , 2025 , 24 ( 10 ): 10710 - 10728 .

YAO S , LIN Y , WANG M , et al . LEOEdge: a satellite-ground cooperation platform for the AI inference in large LEO constellation [J ] . IEEE Journal on Selected Areas in Communications , 2025 , 43 ( 1 ): 36 - 50 .

WANG Y , ZHAO K , ZHANG X , et al . Towards space intelligence: adaptive scheduling of satellite-ground collaborative model inference with space edge computing [C ] // Proceedings of IEEE INFOCOM 2024 Workshops . Piscataway : IEEE Press , 2024 : 1 - 6 .

QIAO Y , TENG S , LUO J , et al . On-orbit DNN distributed inference for remote sensing images in satellite Internet of Things [J ] . IEEE Internet of Things Journal , 2025 , 12 ( 5 ): 5687 - 5703 .

YIN B , FANG X , WANG X . Robust group target awareness inference in multi-UAV-enabled ISCC networks based on split deep reinforcement learning [J ] . IEEE Transactions on Wireless Communications , 2025 , 24 ( 11 ): 9478 - 9492 .

CHEN Y , ZHANG Q , XING R , et al . SLICE: energy-efficient satellite-ground co-inference via layer-wise scheduling optimization [J ] . IEEE Transactions on Services Computing , 2025 , 18 ( 4 ): 2388 - 2402 .

ZHAO H , HUANG S , XU B , et al . Hastening co-inference in space-air-ground integrated networks: a submodular optimization based task offloading approach [J ] . IEEE Transactions on Vehicular Technology , 2026 : 1 - 14 .

XU Z , ZHANG P , LI C , et al . A collaborative inference algorithm in low-earth-orbit satellite network for unmanned aerial vehicle [J ] . Drones , 2023 , 7 ( 9 ): 575 .

FAN S , RONG Y , MENG C , et al . DAPPLE: a pipelined data parallel approach for training large models [C ] // Proceedings of the 26th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming . New York : ACM , 2021 : 431 - 445 .

JEON B , WU M , CAO S , et al . GraphPipe: improving performance and scalability of DNN training with graph pipeline parallelism [C ] // Proceedings of the 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems , Volume 1 . New York : ACM , 2025: 557 - 571 .

XU G , LE Z , CHEN Y , et al . AutoCCL: automated collective communication tuning for accelerating distributed and parallel DNN training [C ] // Proceedings of the 22nd USENIX Symposium on Networked Systems Design and Implementation . Berkeley : USENIX Association , 2025 : 667 - 683 .

WANG L , DONG Y , ZHANG L , et al . Wireless model splitting for communication-efficient personalized federated learning with pipeline parallelism [C ] // Proceedings of the 2023 IEEE 24th International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) . Piscataway : IEEE Press , 2023 : 421 - 425 .

GAO L , LIU J , XU H , et al . Accelerating end-cloud collaborative inference via near bubble-free pipeline optimization [C ] // Proceedings of IEEE INFOCOM 2025 . Piscataway : IEEE Press , 2025 : 1 - 10 .

CHEN Y , JIA X , WANG H . Hierarchical offloading optimization for collaborative DNN inference in satellite edge computing networks [J ] . IEEE Transactions on Mobile Computing , 2026 : 1 - 16 .

WU C , GUO C , YIN Z . A distributed-computing architecture for enabling in-orbit satellite adaptation to evolving AI models [J ] . IEEE Internet of Things Journal , 2026 , 13 ( 2 ): 3513 - 3528 .

ZHANG S , WU W , WU S , et al . Collaborative LLM inference over LEO satellite networks: model splitting and pipeline parallelism [C ] // Proceedings of the 2025 Seventeenth International Conference on Wireless Communications and Signal Processing (WCSP) . Piscataway : IEEE Press , 2025 : 1 - 6 .

WU C , ZHENG P , JIA Y , et al . PipeSIA: an efficient pipeline-parallel distributed space inference approach for LEO constellations [C ] // Proceedings of IEEE International Conference on Communications (ICC) . Piscataway : IEEE Press , 2025 : 6976 - 6981 .

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

基于NOMA的双无人船海上通信资源分配与功率优化

基于时频变换与动态图注意力的多模态序列推荐方法

基于CPO算法的FANET路由优化方法

基于长短期记忆深度Q网络的卫星网络多业务路由算法

IRS辅助MIMO系统的低复杂度自适应相位优化算法