Anand M, Muthurajkumar S
Department of Computer Technology, Madras Institute of Technology, Anna University, Chennai, India.
Sci Rep. 2025 Jul 24;15(1):26952. doi: 10.1038/s41598-025-09633-4.
Internet of Vehicles consists of vehicular nodes that communicate with each other for making intelligent transportation systems, where cyber physical attacks are increasing continuously. Intrusion Detection System (IDS) is able to provide a better security solution for minimizing such cyber physical attacks. Many existing IDSs developed using classification algorithms fail to provide the expected intrusion detection accuracy and they exhibit higher false positive rates. Hence, an efficient Feature Selection Algorithm named Weightage and Ranking Based Feature Selection Algorithm and a Bagging based Fuzzy Convolutional Neural Network classification algorithm with Adam optimizer are proposed in this article which are used to identify the attacks more effectively using bagging with fuzzy inference in the deep convolutional neural network classifier. The proposed system was tested using benchmark and network trace datasets and proved that the proposed IDS enhances the detection accuracy and reduces the false positive rate.
车联网由车辆节点组成，这些节点相互通信以构建智能交通系统，而其中的信息物理攻击正持续增加。入侵检测系统（IDS）能够提供更好的安全解决方案，以尽量减少此类信息物理攻击。许多使用分类算法开发的现有入侵检测系统未能提供预期的入侵检测准确率，并且它们呈现出较高的误报率。因此，本文提出了一种名为基于权重和排名的特征选择算法的高效特征选择算法，以及一种带有Adam优化器的基于Bagging的模糊卷积神经网络分类算法，该算法用于在深度卷积神经网络分类器中通过模糊推理的Bagging更有效地识别攻击。所提出的系统使用基准数据集和网络跟踪数据集进行了测试，结果证明所提出的入侵检测系统提高了检测准确率并降低了误报率。
