on-Board Edge DNN Inference Strategies for LEO Satellite Networks

被引：0
|
作者：
Xie R. [1 ]
Yang Y. [1 ]
Tang Q. [1 ]
Chen Q. [2 ]
Xiang X. [2 ]
机构：
[1] State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications, Beijing
[2] Qian Xuesen Laboratory of Space Technology, China Academy of Space Technology, Beijing
来源：
Beijing Youdian Daxue Xuebao/Journal of Beijing University of Posts and Telecommunications | 2023年 / 46卷 / 02期
关键词：
distributed deep neural network inference; low-earth-orbit satellite; task offloading;
D O I：
10.13190/j.jbupt.2022-239
中图分类号：
学科分类号：
摘要：
To equip a low-orbit satellite network with adaptive edge inference ability, the deep neural network(DNN) model is first established based on the directed acyclic graph. Then, the distributed DNN inference problem in a low-earth-orbit (LEO) satellite network is studied. Next, a quantum evolution algorithm based on joint optimization of excitation function and processing delay is designed to realize the optimal decision of sampling rate setting and task offloading. Finally, the simulation results show that the performance of the quantum evolution algorithm based on excitation function and processing delay is better than that of traditional methods. © 2023 Beijing University of Posts and Telecommunications. All rights reserved.
引用
收藏
页码：57 / 63and103
相关论文
共 19 条
  • [1] LENG T X, QIAN F B, HU W P., Satellite image classification based on artificial intelligence in-depth learning, Forest Inventory and Planning, 46, 1, pp. 1-4, (2021)
  • [2] LI C C, ZHANG Y S, SUN C H., Computation offloading for satellite-based edge computing in LEO satellite network, Radio Communications Technology, 48, 3, pp. 401-407, (2022)
  • [3] XIE R C, TANG Q Q, QIAO S, Et al., When serverless computing meets edge computing: architecture, challenges, and open issues, IEEE Wireless Communications, 28, 5, pp. 126-133, (2021)
  • [4] TANG Q Q, XIE R C, LIU X, Et al., MEC enabled satellite-terrestrial network: architecture, key technique and challenge, Journal on Communications, 41, 4, pp. 162-181, (2020)
  • [5] YU X Y, WANG K, DENG Z L, Et al., Research on satellite network architecture based on MEC, Information and Communications Technology and Policy, 8, pp. 22-27, (2018)
  • [6] XIE R C, TANG Q Q, WANG Q N, Et al., Satellite-terrestrial integrated edge computing networks: architecture, challenges, and open issues, IEEE Network, 34, 3, pp. 224-231, (2020)
  • [7] CHEN Q, GIAMBENE G, YANG L, Et al., Analysis of inter-satellite link paths for LEO mega-constellation networks, IEEE Transactions on Vehicular Technology, 70, 3, pp. 2743-2755, (2021)
  • [8] DING R, CHEN T T, LIU L, Et al., 5G integrated satellite communication systems: architectures, air interface, and standardization, 2020 International Conference on Wireless Communications and Signal Processing, pp. 702-707, (2020)
  • [9] GENG S M., A study on low earth orbit satellite mobile communication channel’s models, (2003)
  • [10] LI J, LIANG W F, LI Y C, Et al., Delay-aware DNN inference throughput maximization in edge computing via jointly exploring partitioning and parallelism, 2021 IEEE 46th Conference on Local Computer Networks (LCN), pp. 193-200, (2021)
12