Scholarlink

In this article, we propose a low-complexity method for identifying LoRa radios based on the signature of quadrature signals from the received baseband signal. The method utilizes computationally efficient analytical metrics derived from the Catch22 library and exploits short initial segments of the preamble in received LoRa frames. The results demonstrate that the proposed approach achieves a favorable trade-off between complexity and accuracy in device identification. With significantly lower computational demand than CNN-based spectrogram methods, it is well-suited for implementation in embedded systems. The method achieved accuracy levels of approximately 80%.

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

LoRa Device Identification: A Lightweight Alternative to CNN-Based Methods

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Wellington P. de Lima,
  • Marcelo E. Pellenz,
  • Marco A. S. Teixeira,
  • Antonio M. Alberti

摘要

In this article, we propose a low-complexity method for identifying LoRa radios based on the signature of quadrature signals from the received baseband signal. The method utilizes computationally efficient analytical metrics derived from the Catch22 library and exploits short initial segments of the preamble in received LoRa frames. The results demonstrate that the proposed approach achieves a favorable trade-off between complexity and accuracy in device identification. With significantly lower computational demand than CNN-based spectrogram methods, it is well-suited for implementation in embedded systems. The method achieved accuracy levels of approximately 80%.