GPS (Global Positioning System) was invented as a precision guidance system for ICBMs (Inter-Continental Ballistic Missiles). Before GPS, missiles were steered by inertial navigation systems (spinning gyroscopes). That would get them to the right city, but not necessarily anywhere near the right block within that city. With GPS, you could get to within about 10 meters of the target that was marked out on the map. That meant that you could take out an army barracks or an airforce hangar building instead of the whole city. This in turn meant you might fight a war without killing several million civilians. That was a great improvement.
Of course, we wanted to give this technical advantage to our own missiles without giving our enemy's missiles a free ride on the system. So the technology was classified, and the receiver modules that were sold to the public for civilian use were crippled in 2 different ways:
1) They were blocked from providing any position data at all at altitudes above 10,000 feet (3,000 meter)
2) The readout precision was "smudged" so that it had a random variation of about 50 meters. To get a "good" position, you had to average over a 24-hour period.
The better signal was encrypted so that only military receivers could get it in "normal" times, and the civilian signal could be completely disabled (globally or in a specific international region) when the military commanded it to be.
This all changed in the early 1990'es, when the civilian signal was allowed to get the full precision during the first Gulf war, so that the US Army could give a cheap civilian receiver to every vehicle in the US Army. Where a US Army dropped into a foreign desert used to be severely handicapped by the unfamiliar terrain, they now had better map data than the Iraqi army and could easily outmaneuver Saddam's army. A couple of years later, Bill Clinton quietly made the improvement permanent, and we have all seen the benefits of that.
Having learned about the GPS system before it was ever launched, the Soviet military quickly decided that they needed a system like that, too. Their clone was called GLONASS but like everything else produced by that system it did not work nearly as well as their literature indicated that it did. In fact, they had trouble keeping the satellites operating for more than a year at a time, so the system never achieved worldwide full service. They are now finally scheduled to reach full service this year.
Since then, both China (Compass) and Europe (Galileo) have designed similar systems and started to launch satellites. It will be a few years still before all 4 systems are each capable of independent world wide service. But the commercial builders of GPS receivers are building devices that will make use of any additional satellites they can see. At least one company, JAVAD (headquartered in Moscow, but chaired by an Indian), has promised that their newest series of high-precision receivers will be able to compute a fix from four arbitrary satellites, even one from each of the four constellations. That is VERY impressive.
Most of us have used a GPS receiver, but few have any idea how they work. Most people seem to think that the receiver transmit a signal that is picked up my the satellites. That is exactly backwards. While there are minor differences between the systems, they are essentially similar, and the following description is more or less true for all of them, although the one I know best is the American GPS system.
The full GPS constellation is 24 satellites (plus a few spares) in medium high orbits at different inclinations betweenn equatorial orbits and polar orbits, designed so that most points on earth will have between 8 and 12 satellites above the horizon at all times. For a given service (such as the civilan "public" signal) all satellites in the constellation transmit on THE SAME frequency, using a direct-sequence spread spectrum signal with a long spreading code - a different code for each satellite. Similar to CDMA mobile telephone encoding, this allows a receiver that knows the proper code for a specific sender to extract that one signal out of the resulting mess. And a receiver can have a single front end (tuned to the shared radio frequency) for all the satellites and apply a separate back end loaded with a separate code for each channel that it needs to decode. Each of the satellites then transmits a complex data stream at a low data bit rate which contains a description of the orbit for each of the satellites as well as a very precise time signal - in the nanosecond range of precision. All of these data streams are precisely synchronized to an absolute time reference. Each satellite actually contains an atomic clock (and two spares!)
When decoding these data streams, the hand held receivers can triangulate to find their position. Since they know the precise position of each satellite (from knowing the orbit of each and the precise current time) and they can measure the relative delay between the multiple satellites that they can receive, they can compute their own position in 3 dimensions from 4 satellite data streams. With more satellites, they can improve on the solution by solving for multiple different combinations of 4 and then averaging.
Even with the best equipment and without the deliberate introduction of errors, this tends to only get to about 10 feet (3 meters) of precision for a variety of reasons. One of these is atmospheric disturbances, especially in the ionosphere, which can affect areas up to a few hundred kilometers at a time. But the error induced by this effect (and several others) tends to be similar for all receivers within a particular geographic area. This means that you can install a stationary receiver, learn its exact location by averaging over several weeks of time, and then compute its error vector by comparing its reading with the known location in real-time. You can then broadcast this correction factor to other receivers on the neighborhood, for example using a local ethernet network or a cellphone network. (Some geostationary satellites broadcast such a corrective data stream in a way that lets some hand-held receivers apply the correction in real time. This is called the WAAS (Wide Area Augmentation Service). While it is not as precise as a local differential reading, it may still get you from 20 feet to 5 feet. A local differential correction obtained from a station a mile or so away may get you in the centimeter range.
Are you impressed yet? I sure am, and I work with wireless engineering every day.
Another article soon will talk about novel applications for GPS and other GNSS (Global Navigation Satellite Systems).