GPS
Will a GPS receiver tell me exactly where I am?
There's a long and a short answer to this question.
The short answer is "no."
The long answer is "no, not exactly, but it can give you a darn accurate guess!"
Most people are surprised to find out that a GPS cannot tell you exactly where you are. They are surprised because they have normally fallen for the marketing hype and buzz phrases such as "pin point accuracy" and "real time satellite tracking". The surface of the Earth is pretty bit - approximately 510 million square kilometers - so expecting "pin point accuracy" from a handheld device costing under 100 bucks receiving signals from satellites 20,000 kilometers above the Earth orbiting it once every 12 hours, each emitting a signal no stringer than a 100 watt lightbulb is asking an awful lot.
So what can you expect from GPS?
The key to GPS being able to give you a good idea of where you are on the Earth's surface relies on each of the 24 active satellites knowing what time it is. But this isn't the "quarter past three" kind of time that we are used to. The GPS network relies on each satellite being able to know the time accurately down to within 3 nanoseconds - that's 0.000000003, or three billionths, of a second. This kind of precision can only be achieved by having ground based atomic clocks constantly updating the clocks on the satellites. Having this sort of precision timing is important because the receiver must determine exactly how long it takes for signals to travel from each GPS satellite. The receiver then uses this information to calculate its position.
Given a clear view of the sky with good satellite lock a GPSr can normally give you horizontal positional accuracy of something in the range on 10 - 20 meters. So, take your GPSr outdoors, give it a clear view of the sky, let it acquire as many satellites as it can and it will give you a set of co-ordinates. Wait a while and you'll notice that the co-ordinates change slightly over time as the GPSr continually recalculates it's position based on the steams of data it is receiving from the SV (space vehicles - another word for the GPS satellites). These fluctuation will, if you plot them on a map, should all fall within a circle of about 20 meters radius. To do this you'd need a really accurate map, a circle of this size on a hiking map or road atlas would be a no bigger than the end of a felt-tip pen or smaller.
On some GPSr you'll come across a screen that shows you an accuracy figure for the co-ordinates give. Two things to bear in mind about this figure are:
- It is an estimated accuracy - if the GPSr knew exactly what this error was it could compensate for it and give you a better reading!
- Manufacturers can be a bit optimistic with this figure!
So why is the system not pin point accurate?
There are a lot of reasons. Firstly, the Earth's atmosphere distorts the signal a bit, especially the ionosphere and troposphere. This alone contributes about 5 meters of uncertainty to the reading you get. The 3 nanoseconds of clock inaccuracy also results in an error that works out to about 2 meters. Then the orbits of the satellites aren't perfect, adding a further 2 meters. Then there is what's known as multipath corruption. Multipath error arises from one of more of the signals from the SVs being reflected from the local surroundings before being received by the GPSr. This can add a further 1 meter to the uncertainty total. Finally you have receiver error (from rounding in calculations and so on) which adds a further 0.5 meters of uncertainty.
Add all this up and you get an uncertainty of about 10 meters under ideal conditions. Poor distribution of satellites in the sky, part of the sky obscured or weak batteries in the GPSr can add to this figure.
Remember too that these estimates of uncertainty are for GPSrs that have been switched on and allowed to get a good lock on the satellites in view. The co-ordinates that you get from a receiver that has just acquired the minimum of three SVs needed to give you an initial set of co-ordinates can be out by much more than this - anything up to 150 meters! So, whenever you switch on you GPSr or come out from under signal cover (trees, tunnels, valleys and so on), give it a chance to reacquire and get a good signal lock before trusting the signal! The same is true of in-car GPS navigation systems - trees, tunnels and built-up areas can affect the signals you receive too so take care!
If in any doubt as to the quality of the data, find a spot with a clear view of the sky (or the best you can find) and stop. Give the GPSr a few minutes to get good signals before moving off again. Funnily enough, it's similar to the advice that anyone navigating with a map and compass might follow!
Navigating to a spot - what does this error mean?
Let's say you are trying to navigate to a particular waypoint using a GPSr (a geocache say!). You've plugged the co-ordinates in and set it as the active waypoint and off you go. If you are some distance away you might do the first part of the journey by car. Quickly the kilometers tick away as you make your way closer and closer. There probably will come a point at which you'll have to leave the car and set off on foot. OK, the GPSr is still counting down the meters or kilometers, only it's much slower. You follow the direction arrow as best as you can taking into account that paths aren't straight but all in all the numbers keep going down.
100 meters to go! Wow! This is getting exciting now.
50 meters ... getting closer now ... 40 meters ... 30 meters ... 20 meters ...
At this point you might be expecting a countdown down to 0 meters where you can just reach under a log or behind a bush and retrieve your geocache. However, in the GPS world things don't work out that way. Once you get to within 20 - 30 meters of your destination some strange things will begin to happen to the numbers. They might suddenly start to go up again. Or they might stick at around a ballpark figure no matter which way you travel. Whilst you are standing still at one spot the GPSr might one moment be telling you that you are 17 meters away from your destination and the next that you are less than 5 meters away. The navigation pointer on your GPSr might also begin to behave in a strange Bermuda Triangle sort of way, giving you weird, conflicting readings.
Why is it behaving erratically all of a sudden when it seemed fine all the time before? Well, you've probably guessed it - it's the errors coming into play. Over long distances, errors of a few meters don't matter much but as you get closer to your destination they matter more and more and if you are dealing with errors in the region of 20 meters, the final 20 meters of so are going to the most problematic! Generally the final 20 meters or so is more intuition than following what the GPSr is saying!
In a next article we show you a few tricks you can use (some simple, some more high-tech) to help you make the final 20 meters or so a bit easier as part of our geocaching articles! more >>
Links
Links open in new window:
Kathie Kingsley-Hughes
Last updated: May 4th 2004
Print This Page |
Email me when this page changes
| Search This Site
- Book: Hacking GPS
- GPS Blog
- How NOT to Lose a GPS
- GPS Devices
- Garmin Geko 201
- Accuracy
- SiRF Technology
- VisualGPSXP
- Screenshot Your GPS
- Garmin GPS Firmware Updates
- NMEA Time
- Geocaching
- Pocket PC/PDA
Out Now!
Take control of your GPS with the new title Hacking GPS by Kathie Kingsley-Hughes >> Go!
|
|