Mitigating tropospheric propagation delay errors in precise airborne GPS navigation

The high spatial and temporal variability of the troposphere is well known, as is its effect-through propagation delays-on GPS positioning. This effect can be particularly problematical in airborne kinematic differential positioning where the altitude difference between reference station and aircraft is typically quite large. The use of zenith delay models and mapping functions at ground stations is fairly well understood, however their use for processing data collected on board aircraft is less well understood. Previous tests have indicated that some of the models often used for navigation purposes (e.g. Altshuler, NATO and the proposed WAAS model) perform poorly compared to those generally used for static positioning. These tests were not done under kinematic conditions however, but as comparisons with ray tracing through radiosonde data. This paper outlines the work recently done at UNB on testing the reliability of tropospheric models in precise airborne GPS navigation. Particular attention has been paid to the performance of the currently proposed WAAS model. The data used to test the models is from an adverse weather flight dynamics experiment undertaken off Newfoundland, Canada, in March 1995. The paper includes an analysis of the GPS flight data to determine the influence of different tropospheric models on the reliability and accuracy of the solutions.