Usha G, Karthikeyan H, Gautam Kumar, Pachauri Nikhil
Department of Networking and Communications, SRMIST, Kattankulathur, Chennai, 603203, India.
Department of Computing Technologies, SRMIST, Kattankulathur, Chennai, 603203, India.
Sci Rep. 2025 Jul 1;15(1):20597. doi: 10.1038/s41598-025-06719-x.
An intelligent transportation system consists of a variety of applications that analyze and exchange information to reduce traffic, enhance traffic management, lessen the impact on the environment, and boost the advantages of transportation for both business users and the general public. Moreover, Intelligent Transportation Systems is different from the standard vehicular ad hoc network design since it functions in a highly dynamic environment brought on by the quick mobility between the nodes in short connection times. These traits make various threats, weaknesses, and denial-of-service assaults possible. The protection of the intelligent transportation system from attacks and continual maintenance is crucial. In this research work, a Distributed Denial of Service attack detection scheme is proposed to protect the Intelligent Transportation System ecosystem, making use of the Adaptive Neuro-Fuzzy Inference System. By resolving the flaws in the DDoS attack detection methods that are currently in use, the security needs of the Intelligent Transportation Systems ecosystem are taken into account. The learning approach of artificial neural networks and the fuzzy logic model is integrated into the Fuzzy System. Based on the experimental results, the proposed model achieved 94.3% accuracy, outperforming traditional classifiers such as Support Vector Machine, Random Forest, Extreme Gradient Boosting, and Convolutional Neural Network. The system demonstrated low false positive rates and high detection reliability, ensuring suitability for real-world Intelligent Transportation Systems security. The proposed scheme attained better results in terms of accuracy, precision, recall, and F1 score.
智能交通系统由各种应用程序组成，这些应用程序分析和交换信息，以减少交通流量、加强交通管理、减轻对环境的影响，并提高交通对商业用户和公众的优势。此外，智能交通系统不同于标准的车载自组织网络设计，因为它在节点之间短连接时间内快速移动所带来的高度动态环境中运行。这些特性使得各种威胁、弱点和拒绝服务攻击成为可能。保护智能交通系统免受攻击并持续维护至关重要。在这项研究工作中，提出了一种分布式拒绝服务攻击检测方案，利用自适应神经模糊推理系统来保护智能交通系统生态系统。通过解决当前使用的分布式拒绝服务攻击检测方法中的缺陷，考虑了智能交通系统生态系统的安全需求。将人工神经网络的学习方法和模糊逻辑模型集成到模糊系统中。基于实验结果，所提出的模型达到了94.3%的准确率，优于支持向量机、随机森林、极端梯度提升和卷积神经网络等传统分类器。该系统显示出低误报率和高检测可靠性，确保适用于现实世界的智能交通系统安全。所提出的方案在准确率、精确率、召回率和F1分数方面取得了更好的结果。
