The principal problem of DGPS implementation for a large region is
cost, including maintenance costs due to the limited range of validity of the GPS
corrections. In general, the operational range of CDGPS cannot exceed 200 kilometers on
land and 700 kilometers at sea.
WADGPS uses a relatively sparse network of monitor stations to provide
a vector of error corrections comprising satellite ephemeris errors, optionally ionosphere
time delay errors, and satellite clock errors including SA. Many studies have shown the
potential of the WADGPS technique for precise GPS position over large areas4
The accuracy of WADGPS positioning will typically achieve 2-4 meters over 2000 kilometer
baselines using C/A code pseudo-range. In practice, the errors affecting the observed
pseudo-range can be separated into three parts in WADGPS as follows
Satellite component: the broadcast ephemeris is a predicted ephemeris
based on global tracking data. The contribution of SA creates an artificial degradation of
the broadcast ephemeris, which can lead to errors greater than 20m. The component of
ephemeris error present in any measured pseudo-range is dependent on the direction of the
receiver to the satellite. If the range from the remote to the monitor station is
sufficiently large, the difference of the respective component of ephemeris errors can be
very large. In WADGPS, a completely independent ephemeris determination is possible,
effectively replacing the broadcast ephemeris.
Atmospheric component: Both tropospheric and ionospheric delay can
vary significantly over as little as 25 km. The troposphere delay correction can be
modelled very precisely using a known model. However, ionospheric delays can only be
reduced by about 75% under the best conditions, and that factor Is a function of time and
latitude. Hence it is not desirable to use the delay correction of the monitor station
instead of that of the remote. In WADGPS, it is possible to establish a local atmosphere
delay model, and correct the delay to 2 meters.
Clock component: The clock component includes the satellite and
receiver clock. SA imposes a rapidly varying dither on top of the normal satellite clock.
In CDGPS, pseudo-range differential corrections also contain a term from the monitor
station receiver clock offset, which will be different for each monitor station. In
WADGPS, a master clock usually is selected among all monitor station receiver clocks, and
all clock terms to the single master clock may be computed. Hence, all pseudo-range
measurements to the different satellites are made to contain the same clock error, making
the determination of the user's position possible.
It is of course necessary to appreciate that no form of DGPS can give
protection from error due to multi path reception, blanking, and/or diffraction from
terrestrial objects. These give rise to the need for an effective backup for GPS. Most
navigators would be horrified at dependence on any single system of navigation. That
principle does not need to be sacrificed at the feet of GPS.
Derived Outputs
The actual accuracies obtained from GPS systems are also a function of
receiver quality. The number of channels, speed of processing, nature of processing i.e.
sequential, fast sequential or parallel, the algorithms employed for modelling propagation
error and for filtering, are Important variables in performance. In addition to this such
factors as vessel velocity, nature of siting, and selected settings for variables such as
elevation mask angle, number of satellites being acquired and speed of re-acquisition can
be important determinants in positional accuracy and reliability as far as the navigators
is concerned,
