Intrusion Detection for Intelligent Vehicle CAN Bus Based on Tsetlin Machine

With the development in Intelligent Connected Vehicle(ICV) technology, the Controller Area Network (CAN) bus responsible for vehicle safety and real-time communication has become a primary target for cyberattacks, which poses a threat to automotive operational safety. We propose the first time a low-power, lightweight, and highly interpretable intrusion detection system (IDS) for CAN bus based on Tsetlin Machine (TM). Experimental results demonstrate that our method achieves an overall detection accuracy of 99.79 % on the Car-Hacking dataset. During the detection process, our method outperforms traditional LSTM and DCNN approaches by more than 3 times in inference speed, while consuming only 1 W of power on the Raspberry Pi 4B.