Autonomous-Positioning Information Aided Train Integrity Detection Risk Analysis Method

Autonomous-positioning information error and failure can lead to a safety risk in train integrity detection. In this study, the problem of train integrity detection, which determines whether a train remains intact in terms of relative position, velocity and acceleration, was investigated. A Bayesian risk-based train integrity detection risk analysis model was proposed to determine both the false-alarm rate and failure rate of autonomous-positioning information based train integrity detection. In this model, both the misdetection probability and failure probability of each autonomous-positioning information were applied to find the misdetection rate and false-alarm rate of train integrity detection. Then, a simulation based on field data was conducted to verify the train integrity detection risk analysis method. The results show that the risk of false alarms is 10-1.8, which is larger than the risk of misdetection (10-5.5) in the complete train scenario. In contrast, in the train break scenario, the risk of misdetection is 10-1.8 when cruising and 1 when accelerating, which are larger than the risk of false alarms (10-4.5 when cruising and 0 when accelerating). The results conform to the train integrity states, indicating that the risk analysis model has sufficient capability to calculate the risk of autonomous-positioning information aided train integrity detection.