Scholarlink

<p>Vehicular Ad Hoc Networks (VANETs) are known to facilitate real-time peer-to-peer communications between automobiles and roadside infrastructure to support smart transportation networks. However, their distributed nature and fast mobility renders VANET to be highly vulnerable to malice routing procedures, including black hole and gray hole attacks. This study constructs an adaptive and intelligent model to identify and prevent routing abnormalities in VANETs both in the laboratory and in real-life situations with the help of deep learning and optimization algorithms. This framework, Dynamic Vortex Search-tuned Intelligent Long Short-Term Memory Network (DVS-ILSTM-Net), combines the learning ability of long short-term memory (LSTM) with a dynamic vortex search (DVS) hyperparameter optimization method The proposed system is trained on synthetic datasets with driving patterns that mimic realistic vehicular environments, with distributions of packet interchange rates, packet exchange request time and response time, and mobility patterns. Preprocessing of the data involves temporal normalization and Z-score filtering of noise, followed by non-linear feature extraction using Kernel-PCA for dimensionality reduction. Experimental results prove that DVS-ILSTM-Net consistently achieves over 98.62% ± 0.28 accuracy, and SD across key evaluation metrics and a statistical evaluation outperforming conventional machine learning and deep learning methods. The model effectively identifies complicated temporal patterns of malicious routing behaviours while providing fast detection with low latency. As a result, DVS-ILSTM-Net is highly applicable in different environments with VANET, enhancing security, reliability, and real-time decision making in next-generation intelligent transportation systems.</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

Real-Time Detection and Mitigation of Routing Attacks in VANETs Using a Novel Deep Learning Framework

  • 期刊论文

  • 发表日期：

    2026年2月7日

查看全文
  • Shashikant Patil,
  • Madhur Taneja,
  • N. Raghu,
  • Biswaranjan Swain,
  • N. Hanuman Reddy,
  • Sheetal

摘要

Vehicular Ad Hoc Networks (VANETs) are known to facilitate real-time peer-to-peer communications between automobiles and roadside infrastructure to support smart transportation networks. However, their distributed nature and fast mobility renders VANET to be highly vulnerable to malice routing procedures, including black hole and gray hole attacks. This study constructs an adaptive and intelligent model to identify and prevent routing abnormalities in VANETs both in the laboratory and in real-life situations with the help of deep learning and optimization algorithms. This framework, Dynamic Vortex Search-tuned Intelligent Long Short-Term Memory Network (DVS-ILSTM-Net), combines the learning ability of long short-term memory (LSTM) with a dynamic vortex search (DVS) hyperparameter optimization method The proposed system is trained on synthetic datasets with driving patterns that mimic realistic vehicular environments, with distributions of packet interchange rates, packet exchange request time and response time, and mobility patterns. Preprocessing of the data involves temporal normalization and Z-score filtering of noise, followed by non-linear feature extraction using Kernel-PCA for dimensionality reduction. Experimental results prove that DVS-ILSTM-Net consistently achieves over 98.62% ± 0.28 accuracy, and SD across key evaluation metrics and a statistical evaluation outperforming conventional machine learning and deep learning methods. The model effectively identifies complicated temporal patterns of malicious routing behaviours while providing fast detection with low latency. As a result, DVS-ILSTM-Net is highly applicable in different environments with VANET, enhancing security, reliability, and real-time decision making in next-generation intelligent transportation systems.