Integrity monitoring is a crucial concern in Global Navigation Satellite System (GNSS) based positioning for railway transportation. The accurate detection and exclusion of the fault in GNSS measurements will greatly enhance the stability of the positioning performance for the safety critical application of GNSS. However, the existing Receiver Autonomous Integrity Monitoring (RAIM) method may fail in detecting and excluding the incipient fault due to the constraint of a Gaussian assumption. In this paper, a novel Fault Detection and Exclusion (FDE) approach with non Gaussian assumption is proposed to eliminate the effect from the incipient fault. Hatch filter is adopted to eliminate gross errors and smooth the observation but retain the incipient characteristic of deviated GNSS residual. A sliding window-based strategy is introduced to extract the empirical non Gaussian fault-free distribution using kernel density-based distribution fitting. Based on that, a global/local integrated test method is proposed to realize FDE, where the Kolmogorov-Smirnov test and Chi-square test are involved to detect the fault(s), and the improved Efficient Leave One Block Out (ELOBO) strategy is adopted to realize fault isolation. Results of fault injection tests under the GNSS-based train positioning scenario established with the field data demonstrate the performance of the proposal. In the comparison with the Gaussian-domain FDE method, the proposed approach realizes an enhanced sensitivity and exclusion capability to both the low level step fault and the incipient ramp fault. The fault exclusion capability ensures a stable positioning precision level, which is significant in GNSS-based train positioning for specific railway applications.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Detection and Exclusion of Incipient Fault for GNSS-based Train Positioning under Non-Gaussian Assumption


    Contributors:
    Yang, Xuan (author) / Liu, Jiang (author) / Cai, Bai-Gen (author) / Wang, Jian (author) / Lu, Debiao (author)


    Publication date :

    2022-06-01


    Size :

    2959689 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English