How accurate is my GPS?
See the press release from the Clinton Administration regarding GPS Accuracy !
The physical laws governing the theory on the GPS and navigational issues are exact and should, in the ideal world, leave us with exact navigational calculations and positioning. The GPS relies on the integration of many aspects in order to eventually calculate the exact or true position at any point in time. However - we live in an imperfect world and the GPS, as smart as it is, is also not perfect. Hence the inherent inaccuracy of GPS positioning. To understand accuracy factors when using GPS receivers for positioning, we should understand the different errors and their combined (non-linear) impact on positioning accuracy. Error analysis can be performed by a so-called ",truth",-model, but again, is only as good as the ",truth",-model itself. However, for general Traveling purposes, we should not be to worried about GPS errors and inaccuracies. How close to an oasis do you need to get in the desert before you actually can see it ?
For all practical purposes we can combine all error effects and rely on the following accuracies:
1. Military purposes - Accurate to 1 meter (not available for commercial or other use)
2. Civilian GPS receivers - Horizontally : between 7 and 20 meters (Have a look at our Error histogram)
- Vertically: between 60 and 100 meters
3. Civilian DGPS** receivers - Horizontally between 2 and 7 meters
- Vertically between 60 and 100 meters
** DGPS = Differential GPS (with differential beacon receiver) A differential receiver can be fitted to most
GARMIN GPS receivers, including some handheld's.
4. Civilian survey class differential instruments
(with acknowledgement to Anton Reynecke (Land Surveyor - Thanks Anton !)
Civilian survey class differential instruments gives accuracies in the order of one to two centimetre for horizontal, and three to four for vertical, depending on the distance from the base (fixed) station. The centimetre level accuracy is for anything up to the order of 20 km from the base, and degrades slightly for longer distances.
These accurate receivers were available and operative even with Selective availability still active since some 12 (I think) years ago.
These type of instruments are rather pricey, in the order of R 300 000.00 for a rover and a base station.
GPS errors can be classified according to the sources of the errors:
1. Environmental induced errors
These errors arise from inaccuracies in the modeling of the environment, ionospheric conditions and changes in ionospheric conditions, inability to determine or predict the direction and magnitude of the gravitational vectors at any given time at a given location, inaccuracies in modeling the earth's shape etc. Ionospheric interference on GPS signals are counteracted by using dual frequencies to completely remove the impact of the ionospheric interferences in military receivers. Unfortunately, only one frequency is available for civilian GPS receivers and they have to rely on mathematical models to estimate the distortion impact of free electrons on the satellite signals. Of the inherent inaccuracy of 15 meters for civilian GPS receivers without SA, about 5-10 meters are due to the influence of ionospheric interferences, and unless that improves, we can do nothing else to increase the accuracy of one-frequency, civilian GPS receivers.
Satellite geometry or constellation, is another source of errors or inaccuracy. DOP (dilution of precision) is the term used to describe the impact of the satellite geometry on the accuracy of GPS receiver readings. DOP has a number of components: horizontal, vertical, position, time and geometric. When visible satellites are in a straight line and relatively close to each other, DOP will be at its greatest. GARMIN GPS receivers calculates the different DOP figures from all visible satellites and uses the ",best", positioned satellites based on the LPDOP, (lowest position DOP) A calculated DOP value of more than 6 might lead to a GPS receiver not locking onto the satellites due to the large error that might result. The best thing to do in such a case is to wait a few minutes and the try again.
Reflected satellite signals (i.e.. from mountains) might lead to a multi-path signal (that is more than one path of one satellite's signal) which cannot be corrected by normal civilian GPS receivers and readings based on multi-path signals will therefore not be accurate.
2. Computational errors
All our GPS components rely on digital computer systems to perform all calculations. Rounding errors, Overflow errors, limited stack space etc. all contributes little by little to the inherent accuracy capability of our GPS receivers. When performing velocity calculations, we can use State-estimation techniques, relating to the motion of the sensor (i.e.. acceleration), to perform measurements with the resultant impact on accuracy.
3. Alignment errors
GPS sensors should all, theoretically, be perfect aligned with their ",assumed", directions. Again, differences in actual alignment and ",assumed", alignments lead to small errors in the ",chain of calculations", and influence the accuracy of our GPS receiver coordinate read-outs.
4. Instrumentation induced errors
From the GPS ground control stations to the GPS Satellite constellation to our GPS receivers, uses sensors to sense various variables. These sensors may sense variables that do not equal the actual physical quantity it is measuring. Impurities in sensor material, random noise sensed, sensor noise (especially at high frequencies), inaccurate zero-settings or bias, sampling rate, scale factors etc. may all influence the accurate sensing of physical sensed quantities and lead to instrument induced errors.
The above errors lead to the so-called inherent accuracy limitation on commercially available GPS receivers which is currently in the order of 15m. Again nothing to get really worried about when navigating to that oasis in the desert.
5. Human induced errors.
Very obviously but sometimes forgotten - are your map-datum on your GPS corresponding to the Map datum that was used for determining the original Waypoint ?? Large errors might result from using two different map-datums to reference the same location.
For a view on the impact on accuracy while marking waypoints on the move - look here
Seth Tyler's most recent article is about homer gps, he presents articles and information on gps equipment sales, gps for trucks, and related topics...