Retracted: An Effective Anomaly Detection Model for Securing Communications in Electric Vehicles

The high penetration of electric vehicles of different types in power grid states the future competitive and active role of this significant technology in the modern human life. Along with the many benefits of these devices, their high dependency on the inter-vehicle communication makes them vulnerable to possible penetrations for affecting both the vehicle and the grid. This paper proposes an effective cyber resistive model to detect and stop cyberattacks in the vehicle. To this end, any malicious behavior in the controller area network (CAN) bus is detected by a hybrid smart structure composed of wavelet decomposition method and modified support vector machine. Owing to the high complication and severe nonlinearity of the message frame frequency in the CAN protocol, an optimization tool based on crow search algorithm (CSA) is introduced to adjust the anomaly detection model parameters, optimally. A three-phase modification method is developed to add to the CSA population diversity for escape to get trapped in local optima and increasing the convergence rate. The real dataset recorded from an electric vehicle brand along with some simulated cyberattacks advocate the high security and efficacy of the proposed model.