RAIM Detector and Estimator Design to Minimize the Integrity Risk

摘要

Future multi-constellation GNSS open the possibility tofulfill stringent navigation integrity requirementsspecified in safety-critical applications using receiverautonomous integrity monitoring (RAIM). In this paper,both the RAIM detector and its estimator are analyzed todevelop a new algorithm. In the first part of this paper,the detector is selected by rigorously comparing two ofthe most widely implemented methods. In particular, thepaper reveals fundamental differences between solutionseparation (SS) and residual-based (RB) RAIM. SSprovides higher fault-detection performance than RBRAIM because the SS test statistic is tailored to the faulthypothesis, and to the state of interest. To prove theseresults in presence of multi-measurement faults, whichoccur in multi-constellation GNSS, analytical expressionsof the worst-case fault direction are derived for both SSand RB RAIM. In the second part of the paper, a nonleast-squares estimator is designed to reduce the integrityrisk at the cost of lower accuracy performance forapplications where integrity requirements are moredemanding than accuracy requirements. The newestimator is numerically determined by solving anintegrity risk minimization problem that includes multiplesimultaneous fault hypotheses. Performance analysesshow a substantial drop in integrity risk using the new RAIM algorithm as compared to a SS method that uses aleast-squares estimator. In parallel, the decrease inaccuracy performance is quantified. Combinedavailability of accuracy and integrity is evaluated at anexample location for a GPS/Galileo navigation system.