Multiple Faults Isolation For Multi-Constellation GNSS Positioning through Incremental Expansion of Consistent Measurements

Yan, P., Hu, Y., Wen, W., Hsu, L. T.

IEEE Sensors Journal (2025)

journal Q1

Fault isolation first builds a minimum, fault-free 'basic set' from ordered studentized residuals, then incrementally admits measurements that pass a no-fault hypothesis test — measurements that never fit are isolated as faults.

Key idea. Multi-constellation GNSS supplies many redundant measurements, but also a higher chance of several simultaneous faults (multipath / NLOS). Instead of searching every measurement subset, the method first constructs a small minimum basic set assumed fault-free — selected via ordered studentized residuals — then incrementally expands it, admitting only measurements that pass a no-fault hypothesis test. Measurements that never fit are isolated as faults.

Why it matters. This turns multiple-fault isolation from a combinatorial search into a tractable, incremental test, improving the reliability of multi-constellation positioning in challenging urban environments. (IEEE Sensors Journal, 2025, Q1.)