Fig. 2

Architecture of the proposed algorithm for fault-robust localization and integrity monitoring. The algorithm incorporates GNSS and odometry measurements to infer the probability distribution of the vehicle state as a set of weighted particles in a particle filter. Robustness to faults in GNSS measurements is achieved using the GMM weighting, measurement voting and vote pooling steps. By analyzing the estimated probability distribution and GNSS measurements, the framework derives metrics of precision and probability of positioning failure, enabling continuous detection of positioning failure to trigger alarm across all time steps