Scholarlink

<p>Smart Logistic Systems (SLS) enhance operational efficiency and visibility through the Internet of Things (IoT) and edge computing yet they are increasingly vulnerable to cyber threats targeting edge networks. This paper introduces a robust and automated Intrusion Detection System (IDS) designed for edge-centric SLS environments. Our proposed method, the Self-Attention-based Bi-directional Long Short-Term Memory (SAT-B-LSTM) model, integrates a Bi-directional Long Short-Term Memory (BiLSTM) network with a self-attention mechanism to dynamically prioritize critical segments of the input sequence utilizing the trainable hidden states from both forward and backward layers of the BiLSTM. To improve model performance, we use a chi-square feature selection technique before training that effectively reduces the dimensionality of the feature set. This approach enhances detection accuracy while maintaining the model’s lightweight characteristics by minimizing additional parameters. We outline a comprehensive, automated deployment strategy that proactively addresses various potential intrusion scenarios. Our extensive experiments including ablation studies and evaluations across multiple metrics were conducted on large-scale public datasets, namely CICIoT2023 and CICDDOS2019. The results demonstrate that SAT-B-LSTM achieves superior detection accuracy compared to existing models. This positions the proposed SAT-B-LSTM as an effective solution for enhancing the security of smart logistics systems and operations.</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

An intelligent light-weight intrusion detection system for edge-centric smart logistics system

  • 期刊论文

  • 发表日期：

    2025年9月29日

查看全文
  • Miftah Bedru Jamal,
  • Muhammad Adil,
  • Danish Javeed,
  • Yaping Zhao

摘要

Smart Logistic Systems (SLS) enhance operational efficiency and visibility through the Internet of Things (IoT) and edge computing yet they are increasingly vulnerable to cyber threats targeting edge networks. This paper introduces a robust and automated Intrusion Detection System (IDS) designed for edge-centric SLS environments. Our proposed method, the Self-Attention-based Bi-directional Long Short-Term Memory (SAT-B-LSTM) model, integrates a Bi-directional Long Short-Term Memory (BiLSTM) network with a self-attention mechanism to dynamically prioritize critical segments of the input sequence utilizing the trainable hidden states from both forward and backward layers of the BiLSTM. To improve model performance, we use a chi-square feature selection technique before training that effectively reduces the dimensionality of the feature set. This approach enhances detection accuracy while maintaining the model’s lightweight characteristics by minimizing additional parameters. We outline a comprehensive, automated deployment strategy that proactively addresses various potential intrusion scenarios. Our extensive experiments including ablation studies and evaluations across multiple metrics were conducted on large-scale public datasets, namely CICIoT2023 and CICDDOS2019. The results demonstrate that SAT-B-LSTM achieves superior detection accuracy compared to existing models. This positions the proposed SAT-B-LSTM as an effective solution for enhancing the security of smart logistics systems and operations.