Sensor Fusion and GNSS Augmentation Services for Safe Train Positioning - Accuracy and Integrity Performance Evaluations

Abstract

European rail actors are exploring innovative solutions for ETCS to improve the competitiveness of railway transportation services. One potential technology being studied is GNSS (Global Navigation Satellite System) and SBAS (Satellite Base Augmentation System), which can enhance safe train positioning and make it profitable for railways. However, the main challenge lies in providing safety demonstration to certify these new systems. Airbus D&S works on designing new EGNOS features for Rail aligned with rail safety-of-life application needs. The proposed solution, developed in CLUG’s projects [3], involves using sensor fusion and integrity algorithms with GNSS and EGNOS based on extended Kalman filter called LOC-OB (LOCalisation On-Board unit). This combination allows real-time position and velocity determination with a high level of integrity required for rail Safety-of-Life operations. The focus of this study is evaluating how these new services impact LOC-OB's integrity performance by determining achievable safe Confidence Intervals (CIs). The paper presents performance predictions considering features offered by future EGNOS add-ons such as time correlations, satellite velocity corrections, clock drift corrections, and integrity indicators. Improved models and algorithms are implemented in Airbus’ SALSA-for-Rail performance prediction tool. It demonstrates that time correlation significantly inflates estimated CIs and emphasizes the importance of accounting for this inflation in CI computation algorithm to ensure safety compliance.