A Guide To Robust Navigation Filters Under Multi-Fault Environments

Kurzfassung

Global Navigation Satellite Systems (GNSS) positioning is severely compromised in urban environments by multipath effects and non-line-of-sight (NLOS) reception, which introduce non-Gaussian measurements errors and generate multiple simultaneous outliers. These challenging conditions significantly degrade the performance of conventional Precise Point Positioning with Ambiguity Resolution (PPP-AR). Multiple Hypotheses Solution Separation (MHSS)-based schemes are considered the standard approach for Fault Detection and Exclusion (FDE). However, they become computationally infeasible when applied to real-time, multi-constellation, multi-frequency GNSS systems due to their combinatorial complexity, particularly under complex urban scenarios. This paper builds upon a Robust Kalman Filter (R-KF), which integrates M-estimators into the KF update step to mitigate the impact of faulty or contaminated observations without the need of enumerating fault hypotheses. This paper provides practical guide to implementing a MHSS scheme and compares its performance against the R-KF to showcase its potential as a alternative method for resilient PPP-AR under multi-fault conditions. Simulation results under controlled fault injections scenarios show that the robust filter achieves performance comparable to MHSS while offering significant advantages in computational efficiency and scalability for future real-time deployment.