Robust PPP-AR Using M-Estimators for Multi-Fault Scenarios

Urban GNSS positioning is severely affected by multipath propagation and non-line-of-sight (NLOS) reception, which give rise to non-Gaussian measurement errors and multiple simultaneous outliers. These effects challenge conventional Precise Point Positioning with Ambiguity Resolution (PPP-AR) techniques, whose estimation performance degrades significantly under such conditions. Existing fault detection and exclusion methods, particularly those based on multi-hypothesis solution separation (MHSS), become computationally infeasible when applied to multi-constellation, multi-frequency GNSS due to their combinatorial complexity. In this paper, we propose a robust filtering framework for PPP-AR that incorporates M-estimators into the Kalman filter update step to mitigate the impact of faulty or contaminated observations without needing to enumerate fault hypotheses. Our method improves the reliability of the float solution in the presence of outliers while remaining scalable to modern GNSS configurations. Simulation results under fault injection scenarios demonstrate that the robust filter achieves performance comparable to an ideal fault-free estimator, effectively preventing divergence and enabling consistent navigation even under degraded conditions.