天地一体化信息网络 ›› 2022, Vol. 3 ›› Issue (2): 72-80.doi: 10.11959/j.issn.2096-8930.2022023

所属专题: 专题:卫星互联网空间载荷

• 专题:卫星互联网空间载荷 • 上一篇    下一篇

卫星互联网业务智能识别分类算法与仿真

崔涛1, 任智源2, 黎军1, 谭庆贵1, 李静玲1, 梁薇1   

  1. 1 中国空间技术研究院西安分院空间微波技术重点实验室,陕西 西安 710100
    2 西安电子科技大学综合业务网理论及关键技术国家重点实验室,陕西 西安 710071
  • 修回日期:2022-04-22 出版日期:2022-06-20 发布日期:2022-06-01
  • 作者简介:崔涛(1984-),中国空间技术研究院西安分院高级工程师,主要研究方向为卫星网络处理与交换
    任智源(1983-),西安电子科技大学教授,博士生导师,主要研究方向为星载在轨边缘计算
    黎军(1975-),中国空间技术研究院西安分院研究员,主要研究方向为卫星高速数据处理与传输
    谭庆贵(1975-),中国空间技术研究院西安分院研究员,主要研究方向为微波光子、卫星光通信和光学相控阵天线技术
    李静玲(1984-),中国空间技术研究院西安分院高级工程师,主要研究方向为卫星高速数据处理与传输
    梁薇(1986-),中国空间技术研究院西安分院高级工程师,主要研究方向为卫星高速数据处理与传输
  • 基金资助:
    国家重点实验室稳定支持基金(HTKJ2020KL504010);国家重点实验室稳定支持基金(HTKJ2021KL504006)

Intelligent Identifi cation and Classifi cation Algorithm and Simulation of Satellite Internet Business

Tao CUI1, Zhiyuan REN2, Jun LI1, Qinggui TAN1, Jingling LI1, Wei LIANG1   

  1. 1 National Key Laboratory of Science and Technology on Space Microwave, China Academy of Space Technology (Xi'an), Xi'an 710100, China
    2 State Key Laboratory of Integrated Services Networks, Xidian University, Xi'an 710071, China
  • Revised:2022-04-22 Online:2022-06-20 Published:2022-06-01
  • Supported by:
    The Sustained Supported Foundation by National Key Laboratory(HTKJ2020KL504010);The Sustained Supported Foundation by National Key Laboratory(HTKJ2021KL504006)

摘要:

任务需求的不断增长以及接入业务种类的不断增加,对卫星互联网的多业务协同传输能力和网络资源全局调度能力提出新的挑战,需要对其承载的业务进行区分服务从而进行资源的最优化调配,满足卫星互联网多业务服务质量保障要求。提出一种面向卫星动态网络的业务智能识别分类算法,通过分析卫星间的连通关系,建立时空稳态图,确定动态卫星网络拓扑,构建卫星仿真网络;在此基础上对5种卫星业务数据进行采样从而形成训练数据集,提出一种基于灰度图的业务智能识别方法,在卫星动态场景下将抓包获取的数据包转换为灰度图作为算法的输入进行训练与识别。结果表明随着迭代次数的增加,识别准确率随之提高,可在97%以上,验证了所提算法的有效性,为卫星互联网空间载荷的发展提供了技术支撑。

关键词: 卫星互联网, 动态网络, 业务智能识别, 服务质量

Abstract:

With the continuous growth of task requirements and the increasing types of access services, new challenges are raised to the multi-service collaborative transmission capability and the global scheduling capability of network resources of satellite Internet.It is necessary to diff erentiate the services carried by it to optimize resource allocation to meet the service quality assurance requirements of satellite internet multi-service.A business intelligence identifi cation and classifi cation algorithm for satellite dynamic network was proposed.By analyzed the connectivity between satellites, a spatio-temporal steady state diagram was established, the dynamic satellite network topology was determined, and a satellite simulation network was constructed.Sampling to form a training data set, a business intelligence recognition method based on grayscale image was proposed, and the captured data packet was converted into a grayscale image as the input of the algorithm for training and identifi cation under the satellite dynamic scene.The results showed that with the increase of the number of iterations, the recognition accuracy increases, and the recognition accuracy was 97% or more, which verifi ed the eff ectiveness of the proposed algorithm and provided technical support for the development of satellite internet space payloads.

Key words: satellite internet, dynamic network, business intelligence identifi cation, service quality

中图分类号: 

No Suggested Reading articles found!