Scholarlink

<p>In the Internet of Things (IoT) context, traditional static security models such as role-based or attribute-based access controls prove inadequate due to IoT connections’ dynamic and opportunistic nature. To address these challenges, we propose a trust evaluation methodology by incorporating game theory with Perfect Bayesian Equilibrium (PBE) and a smoothing function (SF). Our enhanced PBE-ES model dynamically adjusts the trustworthiness of IoT nodes based on observation and Bayesian updates, incorporating node power, distance, and latency to improve trust evaluation accuracy. Our experimental study demonstrates significant improvements in detecting malicious nodes in a network, achieving up to 99% accuracy across various scenarios with different noise levels and network sizes. Additionally, we used the PBE-SF algorithm to generate synthetic data for training classification models like Logistic Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), and Neural Networks (NN). Random Forest (RF) performed best, achieving an RMSE of 0.01 in evaluating the utility of IoT nodes, which improves the accuracy of detecting malicious nodes. Our results demonstrate the PBE-ES model’s effectiveness in enhancing trust evaluation and addressing security challenges in IoT ecosystems.</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

Enhanced trust in IoT environments: utilizing perfect Bayesian equilibrium, exponential smoothing, and machine learning

  • 期刊论文

  • 发表日期：

    2025年8月30日

查看全文
  • Himan Namdari,
  • Victor Morales Avalos,
  • Amal Alshehri,
  • Cihan Tunc,
  • Ram Dantu

摘要

In the Internet of Things (IoT) context, traditional static security models such as role-based or attribute-based access controls prove inadequate due to IoT connections’ dynamic and opportunistic nature. To address these challenges, we propose a trust evaluation methodology by incorporating game theory with Perfect Bayesian Equilibrium (PBE) and a smoothing function (SF). Our enhanced PBE-ES model dynamically adjusts the trustworthiness of IoT nodes based on observation and Bayesian updates, incorporating node power, distance, and latency to improve trust evaluation accuracy. Our experimental study demonstrates significant improvements in detecting malicious nodes in a network, achieving up to 99% accuracy across various scenarios with different noise levels and network sizes. Additionally, we used the PBE-SF algorithm to generate synthetic data for training classification models like Logistic Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), and Neural Networks (NN). Random Forest (RF) performed best, achieving an RMSE of 0.01 in evaluating the utility of IoT nodes, which improves the accuracy of detecting malicious nodes. Our results demonstrate the PBE-ES model’s effectiveness in enhancing trust evaluation and addressing security challenges in IoT ecosystems.