AN AI-POWERED SECURITY SYSTEM FOR CAN BUS ATTACKS IDENTIFICATION IN ELECTRIC AUTOMOBILES

Abstract

The network connection within the car, the "Controller Area Network" (CAN) bus serves as an alternative protocol for electric automobiles. Tragically, the lack of a data authentication technique in the CAN bus protocol makes it susceptible to several types of assaults, making it easier for attackers to infiltrate the network. The CAN dataset is collected and the collected datasets attains for preprocessing stage using z-score normalization. For feature extraction, a restricted boltzmann machine (RBM) is used to extract the data. Next, our proposed method (MCFO-DANN) is used to identify and mitigate CAN Bus attacks in electric vehicles. Evaluation against other CAN bus anomaly detection methods demonstrates the superiority of MCFO-DANN, exhibiting higher accuracy. This proactive security solution fortifies electric vehicles against cyber-threats, ensuring real-time monitoring and response, thereby preserving the integrity and safety of the CAN Bus network in electric automobiles.

Keywords

• anomaly detection
• can bus security
• attacks identification
• ai-powered
• electric automobiles
• autonomous vehicles.