Quote:
I'm not sure I understand. I suspect you are saying you often disagree with the suggested routing a Navigation system is providing.

If you are suggesting that a GPS has provided position data that was off by more than its reported accuracy limits (due to satellite geometry) I would be more concerned.


GPS is most certainly a useful tool, but GPS errors are not understood well enough by most folks to risk their own lives with. Most folks assume you just stick the battery in and punch in your destination and away they go. GPS systems have only really become popular in the mainstream since manufacturers dispensed with the monochrome screens that just showed WGS84 polar co-ordinates with no maps to display and didn't ask for such setup parameters such as a map datum and magnetic/grid/true north deviations or declinations etc.

The relationship between the GPS polar coordinated and the electronic map and the algorithms which compute routes and methods to get back on the computed routes if that route is deviated from. A road type GPS system will rarely tell you to stop and tell you turn around because it has detected an incorrect turn off. i.e. there might be 2 right hand turns within 50m - well within GPS error within limited sky view and hence a large HDOP error. That wrong turn will eventually be detected but the GPS algorithm to get you back to the correct turn off may involve a long and tiring circular route back to the same turn off point rather than telling you to Stop and backup the same road you've just traveled down.

For example the variable nature of the speed of the vehicle may increase errors if the velocity filtering is poorly implemented. GPS systems also cannot detect direction or acceleration very well without filtering a number of position data sets so the update rate of the GPS is important. The GPS circuit might be updating every 1 sec or every 10 or 30 sec for example to save battery power.

A lot depends on the accuracy of the electronic map. The map may have been produced from a very old survey on a different map datum for example. It may have been produced from satellite or aerial imagery or ground survey. There is a lot of computation that is also required for datum conversion to cater for different map datum transformations i.e. WGS84 to OSGB36 for UK Ordnance Survey maps (which are very intuitive easy to use and accurate grid type maps). This conversions all have inherent errors transforming elliptical polar coordinates such as WGS84 to grid mapping types where they tend to loose some accuracy at the edges of the grid.

http://www.movable-type.co.uk/scripts/latlong-convert-coords.html

Solar storms may increase TEC atmospheric errors (most GPS don't have access to both L2 and L1 bands, then there is multipath errors where more costly GPS antennas are required to reduce this error. Turning on WAAS or EGNOS will also affect circular and spherical positional error probabilities.

Out of band interference from other microwave signals such as Lightsquared will most likey produce random additional errors.

Then there is deliberate Jamming and Spoofing.

All these error can sometimes conspire to send folks over a cliff.


Edited by Am_Fear_Liath_Mor (05/16/11 11:24 PM)