An Amalgamated Novel IDS Model for Misbehaviour Detection using VeReMiNet

Abstract

Vehicle ad hoc networks (VANETs) are gaining popularity because they can potentially increase traffic efficiency with road safety in cooperative intelligent transport systems (C-ITS). Effective vehicle deployment requires the identification of nodes providing false data. Researchers have focused much of their attention on an important issue called misbehaviour detection for disseminating false data about vehicle information (position, time, speed and vehicle id) in VANETs. It can threaten the security and privacy of the network and cause some problems, including accidents, collisions, and traffic congestion. There are more works have been proposed based on machine learning methods. However, all of the machine learning-based detection models used currently are signature-based, necessitating previous knowledge of the attacks to identify them and also requiring numerous features, which increases the computational time needed to detect attacks and to select the high-performing features; this works a marine predator algorithm based on teaching-learning. The article provides a unique deep learning strategy to recognize the position falsification attacks in VANETs to find hidden patterns from the data for fighting against known and emerging threats. We evaluate our method using a publicly accessible VeReMi dataset, demonstrating that it achieves high detection accuracy of 99.55% for various attacks. According to our findings, our method can be useful for enhancing VANET security and assuring vehicular communications' reliable and secure functioning.