Preliminary Integrity Assessment for GPS/GLONASS RAIM with Multiple Faults
Eugene Bang, Carl Milner and Christophe Macabiau, Ecole Nationale de l'Aviation Civile (ENAC), France; Philippse Estival, DSNA, France
With the full deployment of the Russian Global Orbiting Navigation Satellite System (GLONASS), an increased number of redundant Global Navigation Satellite System (GNSS) measurements are available, which has recently drawn interest in the feasibility of GPS/GLONASS Receiver Autonomous Integrity Monitoring (RAIM). Accordingly, the design of a rigorous integrity test methodology for GPS/GLONASS receiver has been needed in developing the GPS/GLONASS Minimum Operational Performance Standards (MOPS) for GPS/GLONASS L1-only airborne equipment . These standards and test procedures must be validated in order to show that they protect the user with respect to the higher level requirements of integrity and continuity relating to safety. Thus a preliminary integrity performance assessment should be taken by applying key test procedures under relevant standards and assumptions. The resulting performance could be then used as a baseline to design the appropriate integrity test methodology. If the baseline performance does not meet the integrity requirements, new test procedures for the system with multiple faults must be developed.
In this paper, we propose a multiple hypothesis-based approach to test the performance of the conventional RAIM fault detection and the corresponding integrity bound against multiple failures. In particular, we investigate the resistance of GPS/GLONASS RAIM against multiple faults based on the proposed method [1-2]. We examine several failure modes: GPS dual faults, GLONASS double faults, a single GPS fault and a single GLONASS fault, GPS constellation failure and GLONASS constellation fault, and some combinations of those fault modes. For this purpose, we first determine the worst case fault, which is the most difficult to detect whilst leading to a potential positioning failure. This corresponds, for each failure, to identifying both the worst fault direction and magnitude. Next, we evaluate the maximum probability of missed detection (PMD) under each fault by finding the worst orientation of the position error distribution under the relevant worst case fault. In this step, two types of horizontal protection bound are accounted for: one based on the previously proposed multi-bias RAIM protection method  and the other one based on the preliminary FDE algorithm for GPS/GLONASS MOPS . We compare the resulting PMD to the requirement for a single failure, as in lines with the proposed GPS/GLONASS MOPS test procedures . This is to check if the protection bounds based on the current RAIM could protect the users against multiple failures. Also, we compute total probability of hazardous misleading information (PHMI) by accounting for all possible fault hypothesis and compare it with the integrity risk requirement of 10-7. In this work, all the evaluations are based on a single day and eight days which correspond to a multiple of the periods of GPS and GLONASS constellation.
Based on the PMD and PHMI analysis performed, this work proposes some essential considerations on the fault detection and exclusion (FDE) algorithm such as protection bound formulation, and the integrity test procedures and assumptions for GPS/GLONASS RAIM. Furthermore, we discuss the potential of the newly proposed test method in identifying the integrity test methodology for Advance RAIM (ARAIM) receiver.
This paper presents an integrity test method for GPS/GLONASS RAIM receiver and carries out integrity risk evaluations with several failure modes based on the newly proposed method. It is shown that the PMD ranges from approximately 10-6 to 10-1 throughout the world and the maximum PMD of roughly 10-1 occurs under constellation fault conditions. It indicates that the fault detection performance of the current residual based RAIM method for a pair of failures would not meet the PMD requirement of 10-4, especially under constellation faults. We also show that the current RAIM fault detection could provide worse performance for multiple correlated failures than for a single failure. In the same way, PHMI analysis for the proposed multiple failure modes is conducted. In particular, it is shown that GPS/GLONASS protection outlined in this work appears to be safe even for multiple failures when the multi-hypothesis approach is applied. This study would help to design new integrity monitoring test procedures for GPS/GLONASS L1-only airborne equipment, and may eventually lead to the identification of standards and integrity test procedures for ARAIM.
 RTCA Special Committee (SC) 159, “Minimum Operational Performance Standards for GPS/GLONASS (FDMA + antenna) L1-only Airborne Equipment,” draft version (as of 2017).
 Angus, J. E., "RAIM with Multiple Faults", NAVIGATION, Journal of The Institute of Navigation, Vol. 53, No. 4, Winter 2006-2007, pp. 249-257.