Scholarlink

<p>In this paper, a combination of pre-trained Convolutional Neural Network (CNN) models was employed to identify and classify various types of jamming interference in GNSS signals. The selected models include DenseNet, ResNet50, and VGG16 which were chosen for their proven ability to capture complex visual patterns. To enhance performance, two hybrid architectures (Hybrid I and Hybrid II) were designed to leverage the strengths of different models. The models were trained using advanced optimization strategies and fine-tuning of hyperparameters. Notably, the classification was conducted across 18 distinct jamming categories, and Hybrid I achieved the highest training accuracy of 99%, demonstrating superior performance across evaluation metrics such as accuracy, F1-score, precision, and recall. This paper shows that combining powerful CNN architectures with fine-tuned settings provides an effective approach to identifying complex jamming patterns in GNSS signals, even in challenging multi-class classification scenarios. </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

Hybrid CNN architectures for detecting and classification of GNSS jamming attacks

  • 期刊论文

  • 发表日期：

    2025年5月20日

查看全文
  • Maryam M. N. Zanjani,
  • Seyed Hassan Sedighy

摘要

In this paper, a combination of pre-trained Convolutional Neural Network (CNN) models was employed to identify and classify various types of jamming interference in GNSS signals. The selected models include DenseNet, ResNet50, and VGG16 which were chosen for their proven ability to capture complex visual patterns. To enhance performance, two hybrid architectures (Hybrid I and Hybrid II) were designed to leverage the strengths of different models. The models were trained using advanced optimization strategies and fine-tuning of hyperparameters. Notably, the classification was conducted across 18 distinct jamming categories, and Hybrid I achieved the highest training accuracy of 99%, demonstrating superior performance across evaluation metrics such as accuracy, F1-score, precision, and recall. This paper shows that combining powerful CNN architectures with fine-tuned settings provides an effective approach to identifying complex jamming patterns in GNSS signals, even in challenging multi-class classification scenarios.