Scholarlink

Currently, the rapid growth of mobile elements (ME) brings a great challenge to the management of MEs in Space-Air-Ground integrated network (SA GIN), such as the recognition of MEs, fault analysis, target tracking, etc. Traditional schemes of ME management are dependent on human experience, time­consuming and with limited perfomance. As a result, this paper proposes a novel framework for ME management based on federal learning (FL) and knowledge graph (KG). Firstly, a FL-based network architecture is proposed for the accurate recognition of MEs in SAGIN, where the size of input data is changeable and data privacy is considered by integrating FL module. Secondly, the KG is applied to ME management and the KG construction procedures of MEs are discussed based on the observed historical data. Finally, the KG-driven management frame­work of MEs is exhibited and the possible research directions for the KG-assisted ME management are summarized, including unknown attribute reasoning of MEs, fault analysis of MEs and the tracking of MEs. This paper provides a novel vision with the management of ME from the perspective of FL and KG, and may bring new impetus to the management of MEs.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

The Management of Mobile Elements Based on Federal Learning and Knowledge Graph in Space-Air-Ground Integrated Network

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Junsheng Mu,
  • Yulin Liu,
  • Huaiwen Zhang,
  • Zexuan Jing,
  • Ning Gao,
  • Zhiyue Zhang,
  • Yaqiong Liu,
  • Xiaojun Jing

摘要

Currently, the rapid growth of mobile elements (ME) brings a great challenge to the management of MEs in Space-Air-Ground integrated network (SA GIN), such as the recognition of MEs, fault analysis, target tracking, etc. Traditional schemes of ME management are dependent on human experience, time­consuming and with limited perfomance. As a result, this paper proposes a novel framework for ME management based on federal learning (FL) and knowledge graph (KG). Firstly, a FL-based network architecture is proposed for the accurate recognition of MEs in SAGIN, where the size of input data is changeable and data privacy is considered by integrating FL module. Secondly, the KG is applied to ME management and the KG construction procedures of MEs are discussed based on the observed historical data. Finally, the KG-driven management frame­work of MEs is exhibited and the possible research directions for the KG-assisted ME management are summarized, including unknown attribute reasoning of MEs, fault analysis of MEs and the tracking of MEs. This paper provides a novel vision with the management of ME from the perspective of FL and KG, and may bring new impetus to the management of MEs.