A new mathematical model for estimating and correcting the residual neutral wet and hydrostatic delay of radio waves along aircraft trajectories, calibrated and tested with data from numerical weather models and weather balloons, as well as actual aircraft GPS data, has shown, so far, good promise for its use with or without meteorological information. Further testing, with different kinds of non-GPS "truth" data, is planned for the near future. This model is the sum of four error states, each multiplied by a specially chosen function of height. It represents the residual delay that remains uncorrected, due to imperfections in the refraction correction of GPS data during pre-processing (e.g., with a simple model based on a standard atmosphere). The residual zenith delay estimated with the new model is transformed into slant delay along the path of the radio signals, with the Niell's mapping functions. The new model is intended for use in precise (sub-decimeter), kinematic, recursive solutions, in very-long-baseline differential and point-positioning navigation, whether carried out in post-processing or in real time. The recursive, point-positioning, kinematic solutions whose results are shown here, include error states for the rover's position and zenith delay, and for the floated biases of the ionosphere-free linear combination of the L1 and L2 carrier phase (or Lc). In differential mode, the solutions should also include as unknowns the base-stations' residual zenith delays, modeled as simple random walks, one for each fixed site. The results of the preliminary tests described here, some based on hours of airborne GPS data, suggest that this model is suitable for GNSS kinematic navigation with sub-decimeter precision.
展开▼
