Detailed Analysis of RAIM Performance for ADS-B Separation Error

Curtis A. Shively and Daniel G. O'Laughlin

Abstract:	FAA, industry and the user community are currently developing performance standards for GPS-based Automatic Dependent Surveillance Broadcast (ADS-B). Of concern is the possibility of significant error in apparent aircraft separation when the ADS-B position reports are based on slightly different sets of satellites but a fault on a common satellite goes undetected by RAIM algorithms in both aircraft. This paper provides an indepth analysis of the performance of solution separation RAIM in the above circumstance. The analysis characterizes the conditional risk that given a fault occurs, it goes undetected by both RAIM algorithms and the total separation error (including a fault-free component) exceeds a given level. The dependence of this conditional separation error risk on the correlation between the two RAIM decision statistics is derived in detail. Examples of problematic satellite geometries (rare, but use could be permitted) are evaluated. Some of these geometries have positive RAIM correlation and thus, joint RAIM missed detection (Pmd) performance worse than if the RAIM decision statistics were totally independent (negligible correlation). Some other geometries have negative correlation and therefore joint RAIM Pmd performance better than if independent. While the resulting conditional separation risk probabilities are of interest, no performance requirement or standards currently exist to which conditional risk may be compared. Therefore, a previously developed concept of total risk is evaluated to compare ADS-B performance to current surveillance radar. This total risk considers both fault-free and faulted conditional risk and the prior probability of each circumstance. The application of interest is surveillance to achieve the 3.0 nmi separation standard in the terminal area. Total ADS-B risk is evaluated based on two different prospective values of the RAIM Horizontal Alert Limit (HAL) which corresponds to the radius of containment (Rc) associated with ADS-B position reports. The values of HAL are 0.5 nmi and 0.6 nmi, both of which may be associated with an ADS-B navigation integrity category (NIC) of 6. For HAL = 0.6 nmi, total risk is found to be slightly worse than radar for risk probabilities smaller than about 10-5 if the RAIM correlation is positive or negligible. For HAL = 0.6 and negative correlation or for all correlation cases with HAL = 0.5 nmi the ADS-B total separation error risk is always at least as good as radar. The analysis also includes an example of receivers designed assuming SA on (even though it is not). For receivers designed assuming SA on, total ADS-B separation risk is as good as radar, even for HAL = 0.6 nmi.
Published in:	Proceedings of the 2009 International Technical Meeting of The Institute of Navigation January 26 - 28, 2009 Disney's Paradise Pier Hotel Anaheim, CA
Pages:	203 - 220
Cite this article:	Updated citation: Published in NAVIGATION: Journal of the Institute of Navigation
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In