Code-phase vs. carrier-phase GPS

[+EraSeek]

So educate me; how close are we to getting a carrier-phase consumer GPS?

 

"Code-Phase vs Carrier-Phase

 

In essence this method is counting the exact number of carrier cycles between the satellite and the receiver.

 

The problem is that the carrier frequency is hard to count because it's so uniform. Every cycle looks like every other. The pseudo random code on the other hand is intentionally complex to make it easier to know which cycle you're looking at.

 

So the trick with "carrier-phase GPS" is to use code-phase techniques to get close. If the code measurement can be made accurate to say, a meter, then we only have a few wavelengths of carrier to consider as we try to determine which cycle really marks the edge of our timing pulse.

 

Resolving this "carrier phase ambiguity" for just a few cycles is a much more tractable problem and as the computers inside the receivers get smarter and smarter it's becoming possible to make this kind of measurement without all the ritual that surveyors go through." -Trimble

[+Indotguy]

Huh?

[+EraSeek]

Well it means we are on the verge of having centimeter accuracy in consumer grade GPS's.

Look at the advancement we have had since 2000. SA set to zero. Multi-channel recievers. The advent of WAAS to correct Ionospheric errors, timing errors, sat position error, which is availible to most recievers anywhere in North America, and now the European and asian versions. The new sats with a more powerful signal, and the advent of LC2 (duel band for civilians) to also eliminate ionospheric errors. As I understand it a new system where sats talk to each other and know where the other is. The new Sirf chips with higher sensitivity and the ability to grab signals where you couldn't before. And now the possiblitiy of code/carrier-phase GPS's giving you what surveyors have without multiple recievers and post-processing and expense (hopefully).

[John E Cache]

That quote makes no sense to me. I must be missing something. How would you count cycles between the sat and receiver?

[gallet]

it's all explained here

[John E Cache]

it's all explained here

That is a pointer to the same quote. It still makes no sense. Show me a picture where they are counting cycles between the sat and receiver. It is absurd when I try and picture it.

[gallet]

John E, I have led you to the water, you'll need to drink it yourself.

[+oisact]

I'm no expert, but my interpretation is this:

The satellite encodes data on (modulates) a radio wave. The radio wave is transmitted continuously, and the GPSr locks onto it and observes it for changes, which represents the encoded data. So if a high data bit needs to be sent, a property of the wave is modified accordingly. Because the data rate is much slower than the frequency of the radio wave the GPSr will receive many cycles representing a specific bit of data, so it has a (relatively) long time to detect the change. The faster it can register the change, the more in-synch it will be with the satellite, and thus the more accurate it will be. If the GPSr could monitor individual cycles, and detect the exact waveform cycle that the data bit was first applied, then the precision could lead to centimeter accuracy.

 

This would probably require some real-time post-processing, like recording thousands of cycles, and then analyzing them to see the exact cycle the modulation was changed. A thousand cycles only takes 1.5 microseconds (a millionth of a second), so it would of course seem real-time to the end user.

 

Dan

[John E Cache]

I'm no expert, but my interpretation is this:

The satellite encodes data on (modulates) a radio wave. The radio wave is transmitted continuously, and the GPSr locks onto it and observes it for changes, which represents the encoded data. So if a high data bit needs to be sent, a property of the wave is modified accordingly. Because the data rate is much slower than the frequency of the radio wave the GPSr will receive many cycles representing a specific bit of data, so it has a (relatively) long time to detect the change. The faster it can register the change, the more in-synch it will be with the satellite, and thus the more accurate it will be. If the GPSr could monitor individual cycles, and detect the exact waveform cycle that the data bit was first applied, then the precision could lead to centimeter accuracy.

 

This would probably require some real-time post-processing, like recording thousands of cycles, and then analyzing them to see the exact cycle the modulation was changed. A thousand cycles only takes 1.5 microseconds (a millionth of a second), so it would of course seem real-time to the end user.

 

Dan

I get that part. What has that got to do with the number of cycles between the sat and receiver. I see the ticks are smaller and that picking the right one would give better time measurement. I only see a few cycles in their picture. What am I missing?

[+oisact]

The satellite generates and transmits a random sequence, and the GPSr generates the same sequence at the same time. The two devices are synchronized to time (the sat via an atomic clock, the GPSr in turn synchronized to the satellites).

 

The data sequence the the GPSr receives lags behind the sequence it generates internally, due to the fact that even at the speed of light there is still a measurable delay, which represents how far the GPSr is from the satellite. Currently consumer GPSr receivers measure that delay going by the data stream, which is sent at around 1 MHz. That gives an accuracy of around 1 microsecond. They use some additional tricks to increase that accuracy. The carrier signal is 1.5 GHz. If the GPSr can count the time delay difference in cycles of the carrier it will allow nanosecond accuracy, leading to centimeter accuracy in position.

 

Dan

[+GPSlug]

Say you're measuring the distance to the satellite with a ruler. So you count up all the times you have a full ruler (say, 82 million times) and add the fractional part of the ruler (say, 1/4) to get the full distance (82 million feet and 3 inches).

 

For the code phase, the ruler is 1 millisecond long (299792 meters), so it isn't too hard to figure out how many full rulers you need to add to the fractional part and get the pseudorange.

 

For the carrier phase, the ruler is 19 cm long (wavelength of L1). The measurement accuracy of it (the fuzziness of the fractional lines on the ruler) is a couple of orders of magnitude better than for the code phase, but you have no idea how many full rulers there are between you and the satellite. So you ballpark it with the pseudorange to get the "carrier phase ambiguity" (how many rulers) plus or minus several wavelengths. Then you need a bunch of calculations to figure out which number makes the best fit considering all the satellites.

 

It's a lot more complicated than that, though, because the amount of various errors on the range mean you need to calculate everything relative to another receiver in the area to cancel out a bunch of the errors. And a bunch of statistical stuff is needed to make sure you do it right.

 

And when you get a blip in the signal because you've gone under a tree, you lose count ("cycle slip") and have to start over again.

 

"RTK" (for "Real-Time Kinematic") is the term you want to search for to learn more.

 

 

But back to the OP, there are trade-offs between high-sensitivity / high-availability and high-accuracy. The consumer market will favor availability for a long time to come.

[+EraSeek]

Here are the three pages I was reading from: http://www.trimble.com/gps/dgps-advanced2.shtml

Thanks for the input. I always learn from you guys.

[John E Cache]

Random thought: Doppler shift changes the carrier frequency at the receiver. I wonder how that will affect carrier phase measurements.

[+oisact]

Random thought: Doppler shift changes the carrier frequency at the receiver. I wonder how that will affect carrier phase measurements.

 

It shouldn't. Regardless of if the received frequency, the transmission frequency is known, and the EMR still travels at the same speed. So you just count pluses and multiply by by the transmission frequency instead of the received frequency. Also, the receiver should know what the doppler shift will be at any given time (based on the GPSr's last known location), just like it knows where the satellites are in the sky. Receivers (especially the very sensitive ones) most likely have to account for doppler shift already, especially since they are operating at gigahertz frequences.

 

Dan

[John E Cache]

Random thought: Doppler shift changes the carrier frequency at the receiver. I wonder how that will affect carrier phase measurements.

 

It shouldn't. Regardless of if the received frequency, the transmission frequency is known, and the EMR still travels at the same speed. So you just count pluses and multiply by by the transmission frequency instead of the received frequency. Also, the receiver should know what the doppler shift will be at any given time (based on the GPSr's last known location), just like it knows where the satellites are in the sky. Receivers (especially the very sensitive ones) most likely have to account for doppler shift already, especially since they are operating at gigahertz frequences.

 

Dan

Last known location? I just want one known location. I am still having a problem with visualizing the counting of the pulses between the sat and receiver. You need a really long ruler? To count pulses you have to know when to start counting and when to stop don't you? I think I am missing something still.

[+Klatch]

When the hider and seeker of geocaches have centimeter accuracy it will sure take a lot of the challenge out of the find - or require a lot more devious hiding techniques.

[+EraSeek]

Random thought: Doppler shift changes the carrier frequency at the receiver. I wonder how that will affect carrier phase measurements.

 

It shouldn't. Regardless of if the received frequency, the transmission frequency is known, and the EMR still travels at the same speed. So you just count pluses and multiply by by the transmission frequency instead of the received frequency. Also, the receiver should know what the doppler shift will be at any given time (based on the GPSr's last known location), just like it knows where the satellites are in the sky. Receivers (especially the very sensitive ones) most likely have to account for doppler shift already, especially since they are operating at gigahertz frequences.

 

Dan

Last known location? I just want one known location. I am still having a problem with visualizing the counting of the pulses between the sat and receiver. You need a really long ruler? To count pulses you have to know when to start counting and when to stop don't you? I think I am missing something still.

 

I don't understand this end of things real well but the link I provided has animations on all three code/carrier pages showing this process. Perhaps it will help.

[+EraSeek]

When the hider and seeker of geocaches have centimeter accuracy it will sure take a lot of the challenge out of the find - or require a lot more devious hiding techniques.

 

I have been around a long time. I remember when SA was still in effect. Then you were dealing with a football field size search area. Geocaching started when that was dropped. Then you had a basketball to tennis court sized search area. WAAS came along and everyone said it was going to kill the fun because now you has a tether-ball sized search area when you had WAAS. You are right, the search area lessened (in good reception areas) and the hides got more devious, but still the game go on. In fact such a case may be argued that it has also increased creativity in many cases (but certainly not always). My hope is that is what we all strive for--increased creativity. Regardless, centimeter accuracy will not change the game very much from where we are now with WAAS.

[+Klatch]

When the hider and seeker of geocaches have centimeter accuracy it will sure take a lot of the challenge out of the find - or require a lot more devious hiding techniques.

 

I have been around a long time. I remember when SA was still in effect. Then you were dealing with a football field size search area. Geocaching started when that was dropped. Then you had a basketball to tennis court sized search area. WAAS came along and everyone said it was going to kill the fun because now you has a tether-ball sized search area when you had WAAS. You are right, the search area lessened (in good reception areas) and the hides got more devious, but still the game go on. In fact such a case may be argued that it has also increased creativity in many cases (but certainly not always). My hope is that is what we all strive for--increased creativity. Regardless, centimeter accuracy will not change the game very much from where we are now with WAAS.

I think it will have more impact than you think. Today you have the inaccuracy of the hider's GPS and the finder's GPS. It is not unusual (actually typical) to find the cache 20 or 30 feet from where your WAAS enabled GPS says it should be. With centimeter accuracy there is no inaccuracy, to speak of, other than any incompetence on the part of either user. I hope YOU are correct.

[+trainlove]

Unless your, and all our GPS receivers, had an atomic clock (whether cesium or rhudibium) synchronized with the GPS system (forever without interruption) then no such thing can occur.

 

That's not how the system is designed. And that's why GPSr's only cost $40-$400 instead of $10,000.