Performance of Stand-Alone GPS - Understanding GPS: Principles and Applications - page 370

more than four satellites in the position solution. We note here that most receivers
that employ an overdetermined solution do not truly use all satellites in view; they
limit their tracking generally to no more than eight.
From a system management point of view, the core performance metric just dis-
cussed provides a reasonable way to evaluate global performance on a daily basis.
Still, every user group has its own unique needs and ways to employ GPS, under a
wide variety of physical and environmental conditions. The effect of making one
simple change to the accuracy metric (e.g., a switch from a dual- to a single-fre-
quency receiver) can have a profound effect on the results of the metric. An example
of such a contrast may be seen by examining data from June 3, 2000. Dual-fre-
quency performance for the selected day was nominal and reasonably uniform
across the globe [53]. The selected day, however, occurred at the height of the solar
cycle, and performance for single-frequency receivers varied considerably from that
experienced by dual-frequency receivers. Single-frequency accuracy across the globe
exhibited significant variation compared to dual-frequency performance, with verti-
cal 95% errors sampled over 24 hours reaching as high as 55m. Contrast this perfor-
mance with a maximum 95% vertical error over 24 hours of slightly greater than
8m for the same day.
GPS error distributions are driven by several factors. For a dual-frequency posi-
tion solution, the geometry of the position solution is generally the primary factor in
determining how errors are distributed. For a single-frequency position solution,
however, URE has a significant correlation with elevation angle and tends to corre-
late across all satellites in view as the elevation angle increases. This correlation is
driven by the error in the ionosphere single-frequency model. The result is to
increase the ratio of local vertical to horizontal error in the position solution. One
important result of this behavior is that performance predictions of three-dimen-
sional position error computed by multiplying URE and the PDOP will not give a
valid result for single-frequency position solutions. An example of a global compari-
son of vertical/horizontal error ratios for single- and dual-frequency receivers is
presented in Figure 7.33.
Regardless of the metric employed, GPS has continued to exceed expectations
based on original specifications since it began operations. Several factors contribute
to the long-term trend of excellent GPS performance. These factors are:

• On-orbit frequency standard performance: Clocks are approaching an
order-of-magnitude improvement in frequency stability compared to original
performance specifications.

Constellation size: Original satellite life expectancy was conservative, yielding
a constellation that has averaged 25-29 satellites over the past decade.

•

Satellite availability: The majority of operational satellites have proven more
reliable than their original specifications, resulting in an average of more than
25 healthy satellites being available at any time over the past decade.

•

Consistent constellation operations: U.S. Air Force Space Command has
maintained a long-term degree of consistency in its conduct of constellation
navigation uploads and satellite maintenance, and has gradually tuned the
GPS MCS to take full advantage of improving satellite performance.

•