Frequency-Enhanced Dual-Domain Attention Network for Automotive Intrusion Detection System
Weiming Ren, Yongyi Chen, Dan Zhang · 2025
For the security of autonomous vehicles in the Internet of Vehicles (IoV), the intrusion detection system (IDS) is developed to detect malicious behavior and identify potential threats. Deep learning (DL)-based IDS for controller area network (CAN) bus have shown excellent performance. However, most existing DL-based IDSs only classify data into normal and attack categories, which is not conducive to subsequent datatargeted remediation. To more granularly detect and classify cyber-attacks, this paper proposes a frequency-enhanced dual-domain attention network for automotive IDS (FDAN-based IDS). FDAN-based IDS can classify the types of cyber-attacks more accurately enhancing frequency features and generating attention weights from the channel and spatial domains, respectively. To evaluate the performance of FDAN-based IDS, experiments were conducted on the Car-Hacking dataset used for most research. FDAN-based IDS has good classification performance in a wide range of cyber-attacks, which is superior to other similar algorithms.