Certain types of satellite failures, such as broadcast ephemeris messages that do not correspond to true satellite locations, can be difficult to detect by a Local Area Augmentation System (LAAS) ground facility (LGF) because satellites are observed from reference antennas that are very close together. Although ephemeris failures large enough to threaten LAAS user integrity should be very rare, a combination of monitors in the LGF is needed. This paper analyzes the combined effectiveness of several monitoring techniques and proposes solutions for all phases of LAAS operations.
The first solution is the combination of the GPS Operational Control Segment (OCS) and monitors included in the existing Category I LGF prototypes. Built-in LGF checks confirm that broadcast range and range-rate corrections have reasonable values and also compare computed satellite positions based on the current and previous ephemeris messages as well as the current almanac message. OCS monitoring by itself is insufficient for LAAS because OCS alerts are not guaranteed to meet the minimum time-to-alert requirement of 6 seconds for precision approaches. However, the combination of OCS and existing LGF monitoring is sufficient to meet the allocated per-approach probability of an undetected ephemeris error for Category I operations.
For Category II and III operations, the integrity requirement is 200 times tighter than for Category I; thus the monitors described above may be insufficient. This paper introduces new algorithms for validating ephemerides in three dimensions that can provide the needed improvement. The Differential Pseudorange Residual method (DPR) is used to detect ephemeris errors parallel to the LGF satellite line of sight, and the Double Phase Difference with Ambiguity Search method (DPDAS) is used to detect ephemeris errors perpendicular to the line of sight (and thus very difficult to observe). Unlike RAIM, these two methods need only one other already-approved GPS satellite in view. These two methods can detect all ephemeris failures in three dimensions. This paper develops these algorithms in detail and presents experimental results from the Stanford LAAS Integrity Monitor Testbed (IMT).