PROCEEDINGS OF THE 2014 INTERNATIONAL TECHNICAL MEETING OF THE INSTITUTE OF NAVIGATION
The paper focuses on multipath analyses for business aircraft as the current GBAS RTCA multipath model is based on data collected for mainline aircraft. The key objective is to characterize airframe multipath errors and validate that they meet the RTCA model requirement for a typical business aircraft installations, using real data collected during flight as well as simulations to extend the results for more aircraft. In order to collect data for business aircraft multipath analysis, Honeywell flight test aircraft King Air 200 was equipped with three GNSS receivers. The recording equipment included two aviation grade receivers (one with 0.2 chips correlator spacing, the other with adjustable 0.1 or 0.2 chips) and one dual-frequency receiver as a reference. The flight plan covered typical manoeuvres that are encountered in normal operations. There is a particular need for repeated approaches in order to evaluate the multipath for regions where GAST D is applicable. In first campaign, there were a total of 30 flights with all three receivers connected to one dual-frequency antenna. Of those, 15 were flown with the 0.1 chip correlator and 15 with the 0.2 chip correlator for the receiver with adjustable spacing. In the second part, there were a total of 19 flights with the avionics receivers connected to a different single-frequency antenna, then the truthing system. On average, each flight lasted about two hours. The data were analyzed separately in different phases (ground, air, on approach) because multipath effects can be very different when the aircraft is on the ground or airborne. The approach phase which is a subset of the air phase was extracted and analyzed separately. In order to estimate the multipath error, it is necessary to eliminate all the other contributions of the total error. Since multipath is one of the smaller contributions, its isolation requires accurate cancelation of other effects. One of the commonly used techniques is the code-minus-carrier (CMC) in which the common error terms cancel out leaving only the ionospheric divergence, carrier ambiguity, multipath and thermal noise effects. The divergence-free CMC is more suitable for the estimation of code noise and multipath in the L1 measurement because it does not contain the L2 pseudorange error which is not of interest in this analysis. In order to eliminate thermal noise, GNSS simulator was used, to measure CMC versus carrier-to-noise. The multipath plus noise values are compared to the multipath + AAD B model (Airborne Accucy Designator B; minimum and maximum signal levels), and the estimated multipath values are compared to the multipath model. On the ground (during taxi and while stationary), the pseudorange error is the highest because of reflections from the ground and airport structures. The air phase exhibits values which are about half of those seen on the ground. However, they are slightly above the RTCA model in some elevation bins. For the approach phase, the pseudorange error is the lowest and just below the GAST D requirement across all elevations. A legacy multipath diffraction model was tailored for business aircraft. The diffraction model is generated based on shape and size of actual aircraft. Its performance was modified to fit the results of collected data. The analyses have shown that for small aircraft, such as King Air, there is no benefit for multipath mitigation in the use of 0.1 correlator instead of 0.2 correlator in the airborne receiver. The measured results show the performance to be just below the requirement on the approach, i.e. the multipath model currently in the SARPS is applicable as well to a small business aircraft and not only mainline aircraft for which it was originally tailored. This can be extended to other business aircraft. Preliminary analysis using diffractor model has been performed on larger business aircraft that will be used during validation flight tests and indicate results that are conforming with the current requirements.