Blackjack: Ultra-Accurate GPS Measurement

Conrad_Bombora writes: "NASA's Blackjack Global Positioning System (GPS) receiver, flying on the Argentine satellite SAC-C provides a new way to study Earth's gravity field and atmosphere. The Blackjack looks at how the radio signals from the constellation of GPS satellites are distorted or delayed along their way. While a typical GPS receiver can determine its position to about 22 yards, the BlackJack can pinpoint the position of its host satellite continuously an accuracy of about one inch, and can be used for a variety of Earth studies." The paragraph I find most interesting says "the BlackJacks are also equipped with small down-looking antennas to attempt to receive GPS signals that reflect off the oceans."

Comments

Precision GPS

[crisco]

Sub centimeter accurate GPS has been available for at least 10 years to the civilian population. Its also known as Survey Grade GPS. Companies such as Trimble Navigation, Topcon and others provide reciever systems that consistently provide land surveyors with GPS systems that routinely provide results repeatable and verifiable to about 0.5cm.

The receivers use a combination of C/A code (the coarse code that inexpensive receivers use to get you within about 10M) and processing of the GPS carrier wave itself to perform measurements to that accuracy.

Five to Ten years ago you would set up two of these recievers recording GPS measurements and let each run for 15 miniutes or more, then process the data sets against each other to determine the relative positions of the two antennas. Then advances in computing the position for each epoch of satellite data recieved allowed one reciever to be mobile during data collection, only stopping to increase accuracy for each unknown point. Combine this with a radio transmitting the stationary receiver satellite data and a mobile processor powerful enough to do all the fun matrix math involved and you have a Real Time (within a second or so) Kinematic (moving) Survey Grade GPS system. Costs you about $40,000 or so.

One reason these aren't useful to Saddam is the fact that high dynamic situations (like an ICBM) break the entire system, from the C/A solution to the carrier wave processing.

If you want to know more, here is one article that goes beyond the basics of GPS positioning.

Oh, and to respond to some of the people on /., the military USED to introduce error (called Selective Availability) into the C/A code, reducing the accuracy of the measurements from a single GPS reciever from about 10m to 75m or more. That introduction of error has been turned off, though it can be turned back on in case of a national emergency. The military also has an encrypted transmission from GPS satellites called P code, it achieves a higher level of accuracy (with military recievers that can decrypt it) than the C/A code does without SA. Using two recievers (or a reciever and a differential correction signal such as that from the Coast Guard) narrows the error down to around 1m, the differrence is made up by errors introduced by the ionosphere, other atmospheric variables and the internal accuracy of the reciever clock itself. FWIW, GPS recievers are being used to measure atmospheric water vapor content, to aid in weather models and prediction.

Civilian GPS precision is better than rated.

[hey!]

The official ratings given for civilian GPS accuracy are very conservative.

My own experiments with plotting several thousand fixes on the same location showed that 99.9% of the time they fell within a circle of 6m. Other people trying this experiment with longer periods reported that the points fall within 2.5m of the mean 50% of the time, and 7m 95% of the time. With the USCG differential beacon they get within 1.6m 50% of the time and 4.2m 95% of the time.

I've worked with units from SatLoc and Racal that have sub 2m and sub meter accuracy from their own private satellite differential broadcast. The subscription fees are high though. You can also set up your own differential broadcast station for not too much money and get highly repeatable fixes if you aren't within the range of the USCG broadcasts.

Supposedly you can take these survey grade units and get to within 1cm by averaging out the fixes over 24 hours. It does take time though. The reason my own experiments seemed better than the people who took more fixes is that if you average over a short time, say ten or fifteen minutes, the position probably won't change much; every so often the fix slews to a new area a couple meters away and stays put around there for some time. This may have something to do with satellites coming up over the horizon. In any case, a long term averaging is needed to improve precision to what is being reported in the article.

By the way my military acquatences say that it's hard even for them to get their hands on the real good GPS stuff -- they end up buying commercial equipment. We've had some military guys look at the Racal unit despite having a subscription fee of several thousand dollars.

The missing link...

[Verteiron]

First one I found, on NASA's site.