TY - JOUR
T1 - Bayesian fault-tolerant position estimator and integrity risk bound for GNSS navigation
AU - Chan, Fang Cheng
AU - Joerger, Mathieu
AU - Khanafseh, Samer
AU - Pervan, Boris
N1 - Copyright:
Copyright 2014 Elsevier B.V., All rights reserved.
PY - 2014/9
Y1 - 2014/9
N2 - The advent of multiple Global Navigation Satellite System (GNSS) constellations will result in a considerable increase in the number of satellites for positioning worldwide. This substantial improvement in measurement redundancy has the potential to radically advance receiver autonomous integrity monitoring (RAIM) performance. However, regardless of the number of satellites, the performance of existing RAIM methods is sensitive to the assumed prior probabilities of individual fault hypotheses. In this paper, a new method is developed using Bayes' theorem to generate upper bounds on posterior probabilities of individual fault hypotheses given current user measurements. These bounds are used in a Bayesian fault-tolerant position estimator (FTE) that minimizes integrity risk. The detection test statistic is a measurement-based integrity risk bound, which is directly compared with a pre-specified risk requirement. The associated challenge of quantifying continuity risk is resolved using a bounding approach, which is also detailed in this work. The new Bayesian FTE method is shown to be more robust to uncertainty in prior probability of fault occurrence than existing RAIM methods.
AB - The advent of multiple Global Navigation Satellite System (GNSS) constellations will result in a considerable increase in the number of satellites for positioning worldwide. This substantial improvement in measurement redundancy has the potential to radically advance receiver autonomous integrity monitoring (RAIM) performance. However, regardless of the number of satellites, the performance of existing RAIM methods is sensitive to the assumed prior probabilities of individual fault hypotheses. In this paper, a new method is developed using Bayes' theorem to generate upper bounds on posterior probabilities of individual fault hypotheses given current user measurements. These bounds are used in a Bayesian fault-tolerant position estimator (FTE) that minimizes integrity risk. The detection test statistic is a measurement-based integrity risk bound, which is directly compared with a pre-specified risk requirement. The associated challenge of quantifying continuity risk is resolved using a bounding approach, which is also detailed in this work. The new Bayesian FTE method is shown to be more robust to uncertainty in prior probability of fault occurrence than existing RAIM methods.
KW - Bayesian probability
KW - Fault-tolerance position estimation
KW - Global Navigation Satellite Systems (GNSS)
KW - Navigation integrity
KW - RAIM
UR - http://www.scopus.com/inward/record.url?scp=84905589970&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84905589970&partnerID=8YFLogxK
U2 - 10.1017/S0373463314000241
DO - 10.1017/S0373463314000241
M3 - Article
AN - SCOPUS:84905589970
VL - 67
SP - 753
EP - 775
JO - Journal of Navigation
JF - Journal of Navigation
SN - 0373-4633
IS - 5
ER -