Anomaly information detection and fault tolerance control method for CAN-FD bus network.

With the increase in additional functions loaded on the vehicle, the CAN-FD bus gradually replaces the CAN bus to meet the high speed and high bandwidth required by the automotive industry. According to the information characteristics of the on-board CAN-FD bus network, we propose the method of horizontal-vertical information correlation. This method can be used to generate highly reliable arrays to store messages that are properly delivered, with fault tolerance control over the types of attacks detected. In the anomaly information detection module, when the receiver fails to pass the authentication, the attack types are classified according to different characteristics, and the abnormal information is fault tolerant and controlled. In the fault tolerance control module, the abnormal information is detected according to the prompts of the module, and the abnormal information is replaced and deleted. Thereby ensuring the fluency of the information transmission of the on-board bus network, the safety of the vehicle is guaranteed in real time. We tested 1500 in-vehicle bus messages, and the detection rate for injection, theft and flood attacks on on-board CAN-FD buses reached 95%, 99% and 97.4%, and the detection rate for replay attacks was 86.6%.