Health Assessment Method for Controller Area Network in Braking Control System of Metro Train

摘要

Controller area network (CAN) has been widely used in many control subsystems of metro trains because of its high real-time performance and low cost. In the braking control system, CAN bus is used to connect multiple electronic brake control units (EBCU) and transmit braking instructions. Once the network fails, it will endanger the safety of driving. In order to warn early faults and detect network performance degradation in time, the common faults of CAN bus are carefully investigated, and a health assessment method based on self-organizing map network (SOM) is proposed to monitor network performance and detect early network anomaly. The characteristic parameters of the physical layer are extracted, and the SOM model representing the normal patterns of the CAN devices is trained by the normal dataset. Performance degradation is indicated by the distance between the test sample and the SOM best matching unit (BMU). Experimental results show the validity of the proposed method for anomaly detection of CAN bus.