A Multilevel Hybrid Deep Learning Technique for Detection of Attacks in IoV for Marine Application

Abstract

The Internet of Vehicles (IoV) is a subset of the larger Internet of Things (IoT) program that enables inter-vehicle communication and smart vehicle connectivity. Customers have shown a lot of interest in smart automobiles because the inception of IoV technology. Unfortunately, several security and privacy issues have arisen as a consequence of the IoV’s fast expansion, and these issues have the potential to trigger disasters. The integration of marine engines in the IoV enhances performance, safety, and enabling efficient performance tracking and predictive maintenance. This technology enables seamless communication between vehicle systems, optimizing fuel efficiency and operational reliability. Various researchers have reported models for Intrusion Detection System (IDS) in IoT networks that are based on Deep Learning (DL) with the goals of reducing smart vehicle in marine environment and identifying harmful assaults in vehicular networks. Several types of assaults on marine IoV networks can be identified using the new multilevel hybrid DL architecture suggested in this research. Bi-Directional Long Short-Term Memory (Bi-LSTM), Dense, and Gated Recurrent Units (GRUs) form the basis of the recommended model. Examined using the CI-CIDS 2017 dataset is the suggested model’s performance. The investigational findings show that the suggested method reaches a level of attack detection A_accuracy of 94.07%, which is high. Along with F1-score, additional performance metrics like R_recall and P_preecisionconfirm that the suggested framework works better than its competitors. Robust testing shows that the hybrid DL model can accurately and precisely identify a variety of assaults, including distributed denial of service (DDoS), spoofing, and man-in-the-middle attacks. To ensure the future of smart transportation is safe, this technology could enhance the dependability and security of networks that are used by autonomous vehicles.