Scholarlink

The escalating demands of modern wireless communication has increased the number of transmitters, and this has highlighted the necessity of security, availability, and data privacy for both military and civilian applications. Radio fingerprinting using Software Define Radio (SDR) (Akeela and Dezfouli in Comput Commun 128:106–125 in [1]) can be used to identify the transmitters which can give adequate decisive dividends in both conventional and anti-terrorism warfare. This paper presents a novel hybrid method for radio fingerprinting through RF classification. Further, it recommends the Federated learning (FL)-based self-learning model of Radio Fingerprinting for Defense forces which showcase 0.92 accuracy of our client model and 0.94 accuracy of universal model. The implementation encompasses critical aspects including spectrum sensing, signal classification, automatic parameter adaptation, countering spoofing attacks and fingerprint-based authentication, data privacy, and edge computing. Further, the solution is going to ensure all requisite parameters of security, availability, low latency and reduced power, processing, storage and bandwidth usage which is critical in a battle field of conventional warfare and anti-terrorism operations.

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

Advanced Federated Spectrum Busting: Federated Learning-Based Transmitter Identification System for Conventional Warfare and Counter Terrorism

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jyoti Ranjan Satapathy,
  • John Ayeelyan,
  • R. M. Bodade

摘要

The escalating demands of modern wireless communication has increased the number of transmitters, and this has highlighted the necessity of security, availability, and data privacy for both military and civilian applications. Radio fingerprinting using Software Define Radio (SDR) (Akeela and Dezfouli in Comput Commun 128:106–125 in [1]) can be used to identify the transmitters which can give adequate decisive dividends in both conventional and anti-terrorism warfare. This paper presents a novel hybrid method for radio fingerprinting through RF classification. Further, it recommends the Federated learning (FL)-based self-learning model of Radio Fingerprinting for Defense forces which showcase 0.92 accuracy of our client model and 0.94 accuracy of universal model. The implementation encompasses critical aspects including spectrum sensing, signal classification, automatic parameter adaptation, countering spoofing attacks and fingerprint-based authentication, data privacy, and edge computing. Further, the solution is going to ensure all requisite parameters of security, availability, low latency and reduced power, processing, storage and bandwidth usage which is critical in a battle field of conventional warfare and anti-terrorism operations.