GNSS/INS Based Risk Assessment in Train Localization

Risk assessment of global navigation satellite system (GNSS) used in train localization services is essential to fulfilling safety requirements. In this work, GNSS and inertial navigation system (INS) are integrated to build train localization units. The fault detection and identification is realized by the analysis of sensor fusion data, and the horizontal protect level (HPL) is calculated. According to horizontal position error (HPE), horizontal alert limit (HAL) and time to alert (TTA), the states of a train localization unit can be identified. Then the hazardous events caused by risky states are analyzed, the probability of wrong-side failure and the hazard rate are calculated. Finally, the risk assessment method is tested with field data. It is shown that when the false alarm rate and miss detection rate are 1 × 10⁻⁷/h, and the HAL is 20 m, the hazard rate of the train localization unit are calculated as 9.14 × 10⁻⁷/h in open area and 1.52 × 10⁻⁴ /h in constrained environment, respectively. When the railway line requirements on the risk indexes become low, i.e., the false alarm rate, miss detection rate and HAL are all increased, the hazard rate will be reduced. As a result, the hazard rate will be 0 in constrained environment when the false alarm rate and miss detection rate are all 1 × 10⁻⁵/h, and HAL is 100 m. Therefore, it is necessary to cosider requirement difference for railway lines when implementing the risk assessment.