Scholarlink

Each electronic transmitter emits a distinct pattern, allowing for the potential identification of emitters such as radios and radars based on their waveform pattern and modulation. The evolution of Software-Defined Radio (SDR) technology has significantly enhanced the capability to intercept transmissions. Leveraging artificial intelligence, one can analyze the IQ data and waterfall diagram associated with each transmission to validate transmitter identification. This, in turn can be exploited in many civil and military applications and specially in Electronic warfare domain. This paper outlines two methods using Radio Frequency (RF) and image classification for transmitter identification. The paper further introduces a new approach that improves accuracy in data analysis. It combines feature matching from IQ data using 1D and narrow 2D CNN models with YOLO-based image classification for waveforms and waterfall diagrams. This hybrid model addresses the limitations of both RF characteristics and waveform image classification, resulting in higher accuracy.

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

Blind Demodulation and Radio Finger Printing Using CNN for Electronic Warfare Systems

  • 会议论文

  • 发表日期：

    2025年

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

摘要

Each electronic transmitter emits a distinct pattern, allowing for the potential identification of emitters such as radios and radars based on their waveform pattern and modulation. The evolution of Software-Defined Radio (SDR) technology has significantly enhanced the capability to intercept transmissions. Leveraging artificial intelligence, one can analyze the IQ data and waterfall diagram associated with each transmission to validate transmitter identification. This, in turn can be exploited in many civil and military applications and specially in Electronic warfare domain. This paper outlines two methods using Radio Frequency (RF) and image classification for transmitter identification. The paper further introduces a new approach that improves accuracy in data analysis. It combines feature matching from IQ data using 1D and narrow 2D CNN models with YOLO-based image classification for waveforms and waterfall diagrams. This hybrid model addresses the limitations of both RF characteristics and waveform image classification, resulting in higher accuracy.