Scholarlink

<p>Mobile Ad-Hoc Networks (MANETs) are inherently dynamic and infrastructure-less, making them susceptible to a range of security threats such as black hole and wormhole attacks. This research presents a novel and adaptive Intrusion Detection System (IDS) tailored for MANETs, combining Binary Relevance Strategy (BRS) and machine learning for efficient and accurate intrusion detection. The proposed framework incorporates feature extraction from NS-3 simulated network traffic, focusing on 21 crucial features to distinguish normal from malicious behaviors. It introduces a Fuzzy Logic System (FLS) to evaluate performance reliability, enabling adaptive sensitivity to network dynamics and resource constraints like energy and memory limitations. Unlike traditional IDS solutions, this system emphasizes real-time detection with reduced false positives and minimal computational overhead. The novelty of this work lies in its integration of BRS for multi-label classification, feature-optimized lightweight IDS design, and FLS-based adaptability, resulting in a scalable and energy-aware security solution. Comprehensive simulations using NS-3 validate the framework’s superiority in maintaining high detection accuracy and operational efficiency, making it suitable for deployment in critical MANET applications.</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 adaptive intrusion detection framework for mobile ad-hoc networks using feature extraction and NS3 simulation

  • 期刊论文

  • 发表日期：

    2025年10月1日

查看全文
  • Anand Verma,
  • Sunil Gautam

摘要

Mobile Ad-Hoc Networks (MANETs) are inherently dynamic and infrastructure-less, making them susceptible to a range of security threats such as black hole and wormhole attacks. This research presents a novel and adaptive Intrusion Detection System (IDS) tailored for MANETs, combining Binary Relevance Strategy (BRS) and machine learning for efficient and accurate intrusion detection. The proposed framework incorporates feature extraction from NS-3 simulated network traffic, focusing on 21 crucial features to distinguish normal from malicious behaviors. It introduces a Fuzzy Logic System (FLS) to evaluate performance reliability, enabling adaptive sensitivity to network dynamics and resource constraints like energy and memory limitations. Unlike traditional IDS solutions, this system emphasizes real-time detection with reduced false positives and minimal computational overhead. The novelty of this work lies in its integration of BRS for multi-label classification, feature-optimized lightweight IDS design, and FLS-based adaptability, resulting in a scalable and energy-aware security solution. Comprehensive simulations using NS-3 validate the framework’s superiority in maintaining high detection accuracy and operational efficiency, making it suitable for deployment in critical MANET applications.