Scholarlink

<p>Intelligent Transportation System (ITS) facilitates cooperation among vehicles and other entities by exchanging Vehicle-to-Everything (V2X) messages. This information exchange is leveraged by safety and traffic applications, enhancing situational awareness and enabling safer, efficient and sustainable transportation. However, the open interconnected nature of ITS makes it vulnerable to both internal and external attacks from dishonest nodes. This paper proposes a novel trust model (MP-TMD) to detect multiple types of misbehaviours by assigning trust to ITS entities based on multidimensional plausibility and consistency checks on the semantics of the received information. MP-TMD incorporates adaptive weighted neighbour opinions from a filtered set of recommenders to reduce the impact of bogus recommendations. Multiple simulations were conducted to compare the performance of MP-TMD with an increasing number of misbehaving nodes under different attack models. Experiments were also conducted to compare the performance with deterministic and machine-learning models. Simulation results show an improvement of 13.1, 12.27, and 1.3% in F1 score compared to the deterministic algorithm depending on the type of attack. The results also reveal that the proposed model outperforms SVM and MLP, achieving similar performance to LSTM with precision, recall and F1 scores of 97.1, 87.24, and 91.72%, respectively, while executing 10, 12 and 41&#xa0;× faster.</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

MP-TMD: A Multidimensional Plausibility-driven Cooperative Trust Model for Multiple Misbehaviour Detection in Intelligent Transportation Systems

  • 期刊论文

  • 发表日期：

    2025年1月21日

查看全文
  • Faisal Rasheed Lone,
  • Harsh K. Verma

摘要

Intelligent Transportation System (ITS) facilitates cooperation among vehicles and other entities by exchanging Vehicle-to-Everything (V2X) messages. This information exchange is leveraged by safety and traffic applications, enhancing situational awareness and enabling safer, efficient and sustainable transportation. However, the open interconnected nature of ITS makes it vulnerable to both internal and external attacks from dishonest nodes. This paper proposes a novel trust model (MP-TMD) to detect multiple types of misbehaviours by assigning trust to ITS entities based on multidimensional plausibility and consistency checks on the semantics of the received information. MP-TMD incorporates adaptive weighted neighbour opinions from a filtered set of recommenders to reduce the impact of bogus recommendations. Multiple simulations were conducted to compare the performance of MP-TMD with an increasing number of misbehaving nodes under different attack models. Experiments were also conducted to compare the performance with deterministic and machine-learning models. Simulation results show an improvement of 13.1, 12.27, and 1.3% in F1 score compared to the deterministic algorithm depending on the type of attack. The results also reveal that the proposed model outperforms SVM and MLP, achieving similar performance to LSTM with precision, recall and F1 scores of 97.1, 87.24, and 91.72%, respectively, while executing 10, 12 and 41 × faster.