Scholarlink

<p>The rapid adoption of mini-Unmanned Aerial Vehicles (mini-UAVs) has brought significant security challenges, particularly due to their capacity for payload delivery and unauthorized surveillance. This paper introduces a real-time detection and identification system for mini-UAVs, leveraging radio frequency (RF) signal features and a customized Multi-Layer Perceptron (MLP) model to address these threats. Our approach utilizes a Software-Defined Radio (SDR) platform to capture and analyze RF signals exchanged between mini-UAVs and their control stations. By extracting key RF parameters—such as center frequency, Peak-to-Average Power Ratio (PAPR), spectral flatness, and spectral entropy—we effectively differentiate UAV communications from other RF signals. The core of our system is a tailored MLP model, optimized with a novel SinLU activation function and the Lion optimizer, which enhances the model’s ability to capture complex non-linear relationships. Trained on a comprehensive dataset of extracted RF features, the proposed model demonstrates high accuracy in identifying UAV signals. The system’s effectiveness is validated through rigorous experiments using SDR-recorded RC signals, confirming its applicability in real-world scenarios. This research makes a significant contribution to the ongoing efforts to enhance UAV detection capabilities through advanced RF signal analysis and machine learning techniques.</p>

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

Real-Time Detection and Identification of mini-UAVs Using RF Signal Features and a Customized MLP Model

  • 期刊论文

  • 发表日期：

    2025年3月28日

查看全文
  • Tijeni Delleji,
  • Feten Slimeni,
  • Anis Ayadi,
  • Rafik Salhi,
  • Ahmed Siala

摘要

The rapid adoption of mini-Unmanned Aerial Vehicles (mini-UAVs) has brought significant security challenges, particularly due to their capacity for payload delivery and unauthorized surveillance. This paper introduces a real-time detection and identification system for mini-UAVs, leveraging radio frequency (RF) signal features and a customized Multi-Layer Perceptron (MLP) model to address these threats. Our approach utilizes a Software-Defined Radio (SDR) platform to capture and analyze RF signals exchanged between mini-UAVs and their control stations. By extracting key RF parameters—such as center frequency, Peak-to-Average Power Ratio (PAPR), spectral flatness, and spectral entropy—we effectively differentiate UAV communications from other RF signals. The core of our system is a tailored MLP model, optimized with a novel SinLU activation function and the Lion optimizer, which enhances the model’s ability to capture complex non-linear relationships. Trained on a comprehensive dataset of extracted RF features, the proposed model demonstrates high accuracy in identifying UAV signals. The system’s effectiveness is validated through rigorous experiments using SDR-recorded RC signals, confirming its applicability in real-world scenarios. This research makes a significant contribution to the ongoing efforts to enhance UAV detection capabilities through advanced RF signal analysis and machine learning techniques.