Misbehavior Detection in the Internet of Vehicles using Machine Learning based Situation Identification

To improve communications network efficiency, situation identification is essential. The requirement for real-time scenario identification of vehicular communications is growing more pressing provided a certain vehicle movement makes vehicle interaction circumstances vibrant and that there are significant variations in signal propagation, vehicle speeds, and vehicular concentrations among possible situations. On the other hand, it has been demonstrated that cryptography methods are less effective at preventing insider attacks that originate inside the automotive communication network. A misbehaviour detection system has been found to be more effective in identifying and preventing insider attacks. This research study presents a misbehaviour detection system based on machine learning and is learned utilizing datasets produced by numerical simulations in a genuine vehicular distributed system. Vehicle communication systems may choose the proper spatial multiplexing and system integration by correctly detecting the current conditions in order to maintain an efficient and reliable operating condition. A back-propagation neural network (BPNN) is trained, and a scenario identification model is developed, using datasets that comprise channel properties acquired from channel measurements in various scenarios. Additionally, the proposed identification model's best performance-enabling model configuration technique is analysed and offered. The identification accuracy is then verified using validation information from the relevant circumstances.