5.3 DGPS Errors
As with point positioning, DGPS is subject to the errors discussed in Module 3. However, the system errors that impact point positioning, including satellite clock errors, orbit errors and atmospheric errors can be minimised or eliminated using DGPS.
Satellite clock errors
Satellite clock errors are resolved in a similar fashion to the technique in point positioning. The additional receiver observing the same satellites provides additional redundancy, and the pseudo range correction method allows for the calculation of time rather than range, which also minimises the clock errors.
Orbit errors
Using the pseudo range technique for DGPS reduces the impact of the orbit errors. Because the distance between the receivers is tiny compared to the distances to the satellites, errors measured by the base station will be almost exactly the same for the other receivers in the area.
Atmospheric errors
GNSS signals are delayed as they pass through the atmosphere due to the ionosphere and troposphere. While the satellite ephemeris and receiver have some basic atmospheric models to adjust for this delay, one of the most effective way to adjust observations for this type of error is through differential techniques (observing multiple frequencies is another, but we’ll get to that in Chapter 6).
Differential techniques assume that the atmospheric interference will be similar across an area, so the corrections determined at the base station can be applied to the rover. This means atmospheric errors are essentially removed from the positions at the rover.
Errors not resolved in DGPS
The errors that are not resolved or removed by DGPS are:
- DOP – GNSS receivers tolerance levels can be set to only accept data when DOP values are below certain levels
- Low elevations – similarly to DOP, GNSS receivers may have an elevation mask setting that can be adjusted depending on local obstructions.
- Obstructions and multipath – these are site dependent and impact DGPS observations. Users should consider choosing a different location if possible.
- Receiver noise – tolerance levels for SNR can be set in most receivers. Avoidance of electrical influence, such as power lines or similar can assist in reducing receiver noise issues.
- Spoofing – DGPS is still susceptible to spoofing as the majority of attacks are in the L1 band, however, the use of multiple receivers does reduce the risk marginally.
Human error – as with point positioning, our capacity to introduce error is infinite, however, can be mitigated by understanding of how GNSS operates, and also through development of quality systems to manage observations of positions.
Minimising errors in DGPS
The same techniques of error minimisation that apply to point positioning generally apply to DGPS, however, there are some variations.
Point averaging
GNSS errors can be significantly reduced by averaging individual point positions over time. The amount of time a point should be observed is dependent on the accuracy looking to be achieved, however, in DGPS 10-30 seconds is usual sufficient for 0.5m accuracy in open areas.
Redundant observations
Revisiting a point multiple times in point positioning generates redundant observations, however, in DGPS measuring a point multiple times cannot be considered as creating truly redundant observations, as the same baseline correction has been applied each time.
If there was an initial error in the baseline calculation, this error would carry through all measurements that used that baseline correction, so additional measurements on different days, from a different base station at a different time (to allow a different satellite configuration) or by different techniques are required to generate truly redundant measurements.
Measuring known points
While the base station is by definition on a known point, it can be useful to use the rover to observe known points periodically throughout an observation session as an additional check on the position information or real time corrections.