Scholarlink

<p>The area of wireless technologies witnessed a progressive advancement that has eventually assisted Cognitive Radio Networks (CRN). CRN deploys accessibility to spectrum dynamically to mitigate scarcity problem of spectrum. However, this leads to an eventual threats that exploits the open nature and decentralization of CRNs. With increase of Artificial Intelligence (AI) in networking technologies, it is noted that existing security system towards safeguarding intrusion in CRN mainly utilizes static rules and seriously lacks adaptability. This causes delayed threat response and increased outliers which makes existing solution as highly inadequate towards mitigating such lethal threats in CRN. Therefore, the proposed study introduces a novel framework harnessing machine learning where the communication system in CRN is secured by fusing proactive response system, decision-making with trust awareness, statistical extraction of feature, and anomaly detection using entropy. The adoption of the hybrid design integrates prediction using machine learning with evaluation of dynamic trust as well as mitigation strategies in real-time. An extensive analysis is carried out to find that proposed model accomplishes the highest detection rate of 95.7% in average when exposed to various types of lethal threats in contrast to existing machine learning-based models.</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

Proactive security in cognitive radio network: an AI-driven trust and anomaly detection approach

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月13日

查看全文
  • Sudha Yellappa,
  • Suneetha Yellappa,
  • Salma Firdose,
  • S. Reshma,
  • Suman Jayakumar

摘要

The area of wireless technologies witnessed a progressive advancement that has eventually assisted Cognitive Radio Networks (CRN). CRN deploys accessibility to spectrum dynamically to mitigate scarcity problem of spectrum. However, this leads to an eventual threats that exploits the open nature and decentralization of CRNs. With increase of Artificial Intelligence (AI) in networking technologies, it is noted that existing security system towards safeguarding intrusion in CRN mainly utilizes static rules and seriously lacks adaptability. This causes delayed threat response and increased outliers which makes existing solution as highly inadequate towards mitigating such lethal threats in CRN. Therefore, the proposed study introduces a novel framework harnessing machine learning where the communication system in CRN is secured by fusing proactive response system, decision-making with trust awareness, statistical extraction of feature, and anomaly detection using entropy. The adoption of the hybrid design integrates prediction using machine learning with evaluation of dynamic trust as well as mitigation strategies in real-time. An extensive analysis is carried out to find that proposed model accomplishes the highest detection rate of 95.7% in average when exposed to various types of lethal threats in contrast to existing machine learning-based models.