-
Satellite navigation Q&A
by
Galactic Penguin SST
on 03 Jan, 2013 15:46
-
I'll start with two questions:
1. Why would all the GNSS systems in the world all use multiple frequencies for transmitting the navigation signals, given that common usage is usually limited to one signal channel (e.g. most GPS systems we use make use of the L1 band)? What features does the new frequencies have (e.g. the L5 signal of GPS)?
2. How does the GPS (or other systems) receivers correct the error caused by atmospheric effects and changes?
-
#1
by
knotnic
on 03 Jan, 2013 19:33
-
-
#2
by
john smith 19
on 12 Aug, 2013 07:07
-
I'll start with two questions:
1. Why would all the GNSS systems in the world all use multiple frequencies for transmitting the navigation signals, given that common usage is usually limited to one signal channel (e.g. most GPS systems we use make use of the L1 band)? What features does the new frequencies have (e.g. the L5 signal of GPS)?
These systems are designed to
not need a directional antenna so the variable they sense is the time between when the satellite sent the signal and the time they received it and the sending stations identity (because if you can see station 4,5,6 but not 2 or 3 that tells you quite a lot about where you are). But the time the signal takes to travel through the ionosphere varies and is frequency dependent. If you have 2 frequencies [edit and can read the signal data as the military one is encrypted, I presume with the key buried in some of the unused/unidentified bits in the civilian code packets] you can calculate the delay and subtract it out.
2. How does the GPS (or other systems) receivers correct the error caused by atmospheric effects and changes?
For civilian GPS the GPS interface documents (there's no actual NTIS published standard as its a DoD programme) specify a model of the ionospheric delay which (I think) has some variables for time of year and day. That's the default value you subtract from the time.
-
#3
by
baldusi
on 12 Aug, 2013 22:26
-
That's one of the reasons to introduce the L3 and L5 signals. You can get a differential from the signals and filter out most of the atmospheric errors.
-
#4
by
Nomadd
on 13 Aug, 2013 14:32
-
Refraction is also different for different frequencies and one of the biggest limitations for accuracy, so having various frequency bands can let you adjust for that factor by averaging or using more complicated algorithms.
-
#5
by
macpacheco
on 08 Jan, 2014 19:09
-
Ionospheric errors are the largest factor in GNSS (GPS and alike) systems (except for receivers operating in space).
The basic ionosphere model in GPS typically results in 5-20 meter errors (even more in extreme scenarios), from iono errors alone.
Most GPS receivers today use only L1 because:
- newer civilian signals (L2C and L5) are in testing and have a very small number of satellites capable of broadcasting those signals
- multi frequency receivers are expensive
- currently the only way to correct for iono errors is the semi-codeless method, the two operational military signals L1 P(Y) and L2 P(Y) broadcast the same data, so by comparing the delay between the two signals, iono corrections can be calculated (and then applied to the L1 C/A civilian signal), however those methods used to be protected by a lot of patents, so you have the cost of the patents, beyond the added L2 receive hardware and extra processing required, a single frequency GPS chip costs less than US$ 1, a dual frequency GPS chip costs US$ 50-100, so you can see why your smartphone doesn't have semi-codeless
- all this hard work is actually done by all GPS augmentation systems (WAAS, EGNOS, DGPS), so being able to receive those signals improves accuracy by 80% of doing the semi-codeless method on the receiver, currently all augmentation systems use L1 C/A + semi codeless to calculate GPS errors
The actual iono errors are location dependent, so you need to be within the coverage area for that system for the iono grid received to be useful
With WAAS errors get down to typically less than 2 meters (already considering all other error factors).
A dual frequency using L1+L5 with the other augmentation data provided by WAAS should get errors typically to less than 50cm. L1+L2C is a little less precise (the larger the frequency difference the better).
For the in depth technical details, google gps total electron content.
The other errors of GPS are ephemeris+clock errors (augmentations correct those down to a few cm), multipath (some GPS receivers can reject multipath components, most can't) and receiver residual errors (mostly due to the complex trig math receivers need to do).
-
#6
by
Danderman
on 31 Jan, 2014 17:41
-
OK, so GPS data is transmitted on a certain set of frequency bands, but the question is:
Can GPS data be transmitted on other frequencies, assuming that the receiver is set up to receive that signal, and the GPS data can be "decoded" or the frequency could be altered on the ground and converted to standard GPS.
An example of this would be to have a GEO comsat re-transmit GPS data via Ku band spot beam to a user on the ground who, for whatever reason, cannot receive standard GPS transmissions.
Inherent in this question is the problem that such a set-up would emulate a single GPS transmitter, and I have no idea if GPS is viable with just a single transmitter.
-
#7
by
spacecane
on 31 Jan, 2014 17:56
-
Inherent in this question is the problem that such a set-up would emulate a single GPS transmitter, and I have no idea if GPS is viable with just a single transmitter.
No, GPS can not work with a single transmitter. In simple terms, the GPS data gives a map of where all the satellites in the constellation are right now and then the receiver determines which satellites it can see. That gives a general idea of the location of the receiver. Then, it times the difference in arrival time from each satellite to "triangulate" where it is.
Cell phones use AGPS (Assisted GPS). In that case the cell tower will transmit data of roughly what the location is and what satellites should be in view. This speeds up the process of calculating the location.
-
#8
by
ddunham
on 31 Jan, 2014 17:58
-
Can GPS data be transmitted on other frequencies, assuming that the receiver is set up to receive that signal, and the GPS data can be "decoded" or the frequency could be altered on the ground and converted to standard GPS.
GPS doesn't depend on the specific frequencies chosen. If the transmitters and receivers all used a different frequency, then the system would work.
An example of this would be to have a GEO comsat re-transmit GPS data via Ku band spot beam to a user on the ground who, for whatever reason, cannot receive standard GPS transmissions.
Inherent in this question is the problem that such a set-up would emulate a single GPS transmitter, and I have no idea if GPS is viable with just a single transmitter.
That won't work. The receiver needs to have the actual timing that the signal takes from the satellites (at least three) to arrive at a solution. Dropping the number of transmitters to 1 or making major changes to the transmission length will fail. In the "retransmit" case, the receiver on the ground would decode the signal and would arrive at a position solution that coincided with the retransmitting satellite, not the receiver on the ground.
-
#9
by
Danderman
on 31 Jan, 2014 17:59
-
Your statement implies that a minimum of 3 satellites are required to get position.
-
#10
by
baldusi
on 31 Jan, 2014 18:07
-
Your statement implies that a minimum of 3 satellites are required to get position.
Ny. You can guess you position if you know your time with enough accuracy. But you need four to get the total available accuracy (the variables are x,y,z and time).
Also, getting your signals on a plane (equator) would mean that you'd have to know if you are in the Northern or Southern Hemisphere before hand.
But both the Chinese and Indians do use GEO and GSO GNSS and it works well since they don't have high latitudes.
-
#11
by
kevin-rf
on 31 Jan, 2014 18:15
-
It all depends on what you are trying to achieve, The GPS signal is tied to an atomic clock. It is a time signal. Re transmitting it from a GSO satellite will delay the signal reaching the receiver and must be taken into account. In reality the GSO satellite would need it's own atomic clock.
If you have your own atomic clock, you can see how far off you are from a single signal, this will tell you how far you are from the transmitter. That could be a very useful value to have.
If you are looking to establish your location in X,Y,Z you need at least three, actually I think four. The fourth is used as a time reference, since none of us walk around with an atomic clock in our pocket.
1 and a time reference gives you a distance to the transmitter.
2 and a time reference gives you two spheres and an arc that you are on, if the arc intercepts the earth, it gives you two possible locations.
3 and a time reference gives you three spheres which only have one point where all three spheres touch. Your location.
4 gives you your time reference without keeping that atomic clock in your back pocket.
I thought the Chinese where doing GSO based satellite navigation, one would assume the clock is on the ground and the time signal takes into account the trip to GSO.
Btw. If you have a portable atomic clock, GPS is not the only signal you can use. You could tune into one or more pulsar to determine your location anywhere in the solar system... Just saying.
Edit: What Baldusi said
-
#12
by
Danderman
on 31 Jan, 2014 18:21
-
But both the Chinese and Indians do use GEO and GSO GNSS and it works well since they don't have high latitudes.
Do the Chinese use a single transmitter from GEO, or multiple satellites?
-
#13
by
Danderman
on 31 Jan, 2014 18:22
-
Btw. If you have a portable atomic clock, GPS is not the only signal you can use. You could tune into one or more pulsar to determine your location anywhere in the solar system... Just saying.
Which raises the question as to why Mars probes don't carry atomic clocks to determine location from pulsars.
-
#14
by
kevin-rf
on 31 Jan, 2014 18:33
-
-
#15
by
pericynthion
on 31 Jan, 2014 19:21
-
Btw. If you have a portable atomic clock, GPS is not the only signal you can use. You could tune into one or more pulsar to determine your location anywhere in the solar system... Just saying.
Which raises the question as to why Mars probes don't carry atomic clocks to determine location from pulsars.
So far, Mars probes have all been returning significant amounts of science data and therefore have had the support of the Deep Space Network or Russian equivalent. If you have that, it's a no-brainer - coherent ranging from Earth allows highly accurate navigation with minimal onboard equipment. It essentially comes free with the radio.
Pulsar nav on the other hand is attractive for autonomous spacecraft because it doesn't require ground support, but it does need not only an accurate onboard clock but also a dedicated receiver and (potentially quite bulky) antenna suitable for the radio frequency of the pulsar. Complicating this is the fact that the frequency band concerned varies widely from pulsar to pulsar. Pulsar navigation has been proposed using both radio-frequency and x-ray pulsars;
this paper gives a decent introduction.
-
#16
by
Danderman
on 31 Jan, 2014 19:30
-
Just a wild guess, but weight.... Like I said, for pulsars you need an atomic clock, and receiver that can lock onto pulsars. Think big dish.
I believe China has multiple GSO navigation satellites. Google is your friend here, http://en.wikipedia.org/wiki/Beidou_Navigation_Satellite_System
The Chinese system is a hybrid, with satellites in all sorts of orbits.
Anyway, the answer to my question is that 4 satellites seems to be the number of satellites over the horizon required to get a GPS location, assuming you don't have a satellite dish that can pick up pulsars.
-
#17
by
baldusi
on 31 Jan, 2014 19:35
-
Btw, you can get quite accurate positioning by star tracking and a good timing source. I believe that's what Mars probes use. But it's not real time.
And yes, the Chinese and Indian systems are dedicated GSO and IGSO birds. Plus MEO for Beidu.
-
#18
by
baldusi
on 31 Jan, 2014 19:43
-
May be I didn't made myself clear about the single ane problem. If you only put your signaling satellites on the equatorial plane, you wouldn't know if you are in the north or south hemisphere. That's why you need the IGSO birds.
Regarding altitude, is not such a big problem. The time difference gives the distance, and 36.000km is not that different from 22.000km. In fact, the troposphere perturbations are a bigger source of error.
The fact is that if you want to cover 95% of the population, you need to go upto 50degrees of inclination.
And if so the optimum fleet is like Glonass, GPS and Galileo.
-
#19
by
pericynthion
on 31 Jan, 2014 20:10
-
Btw, you can get quite accurate positioning by star tracking and a good timing source. I believe that's what Mars probes use. But it's not real time.
Do you mean asteroid tracking? Deep Space One demonstrated that. You can't navigate by stars alone, there's not nearly enough parallax.
-
#20
by
Danderman
on 01 Feb, 2014 17:12
-
OK, let me ask my question a different way:
Let's say you wanted to generate GPS type data in lunar orbit. Could you embed GPS data in future lunar satellite transmission systems so that a rover on the ground would not have to carry both a GPS receiver and a standard comm link, one transmission path could handle both datasets?
-
#21
by
kevin-rf
on 01 Feb, 2014 18:27
-
I think I saw the quote somewhere on NSF once where they mentioned that a GPS satellite was about as small as it was possible to shrink the package. If they could shrink it smaller, they would have.
To do "GPS" around the moon or another planet you need an accurate time source.
Around the moon you run into another problem, the moon's lumpy gravitational field. To be able to accurately use GPS you need to know precisely where the time signal originated from. That requires precise information on the satellites orbit. The moon's lumpy gravity makes that more difficult.
There are a few other options, and not all require a precise clock.
1. Instead of being passive, have an active system. Meaning you send a signal to the satellites and listen for a response from the satellites. You can determine precise distances to each of the satellites whose location is known. You do need to take into account the latency of the satellite receiving your signal and sending a message back to you, but that should be a constant. This then results in you only needing three satellites for "real time" positioning. But you need a transmitter and receiver.
*If you don't need real time, since a satellite will change position with time, multiple pings and knowledge of a precise orbit will allow you to determine your location to a high degree of accuracy.
2. If you are talking rovers on the moon, or Mars, or Comet ISON for that matter, ground based navigation beacons could work. You really only need two or three. It would provide coverage over a limited area, but it's not like anyone will be doing a Cannon Ball run across Mars in the near future. A "human" rover would not be roving very far from it's home base. You don't really need global coverage. I can even think of how to embed a precise time signal into the system without an atomic clock...
3. You could look up Transit. It was the first satellite navigation system. It worked on the principle of precisely knowing the satellites orbit and measuring the Doppler shift of it's radio beacons. It was not realtime, allowing the end user to get a location fix only after the satellite completed a ground pass. I would assume twice a day. But it was good enough for Polaris.
The real issue is GPS like capabilities are costly and several options exist for localized navigation. The GPS advantage is it is 24/7, mindless, and accurate. That capability does not come cheap.
-
#22
by
Danderman
on 01 Feb, 2014 18:49
-
Around the moon you run into another problem, the moon's lumpy gravitational field. To be able to accurately use GPS you need to know precisely where the time signal originated from. That requires precise information on the satellites orbit. The moon's lumpy gravity makes that more difficult.
Are very high lunar orbits that impacted by the lumpiness? For a spacecraft in a 10,000 km orbit, are perturbations that great?
BTW, is there "Extra credit" for no atmosphere on the Moon?
-
#23
by
Danderman
on 01 Feb, 2014 18:55
-
3. You could look up Transit. It was the first satellite navigation system. It worked on the principle of precisely knowing the satellites orbit and measuring the Doppler shift of it's radio beacons. It was not realtime, allowing the end user to get a location fix only after the satellite completed a ground pass. I would assume twice a day. But it was good enough for Polaris.
Since exploration systems really don't need to know where they are 24/7, I presume that a Transit-type system might be useful for lunar exploration. Perhaps a single satellite that used laser communications with the Earth so that its position were known, and then which transmitted that positional data with timing to the user on the lunar surface could provide daily positional updates to rovers.
Assuming that Elektra communications did not preclude this, a "Mars navigation" satellite could do the same at Mars and transmit positional data to Mars rovers using the Elektra radio system. The receiving spacecraft could extract positional data from the Elektra data stream, assuming that Elektra does not use some sort of fixed digital protocol. In that way, rovers at Mars would not have to include two receivers, one for commands and the other for navigation.
I guess I should mention that the requirement for positional information could be for any surface where there is a rover or lander or prospector, etc.
-
#24
by
pericynthion
on 01 Feb, 2014 19:32
-
Are very high lunar orbits that impacted by the lumpiness? For a spacecraft in a 10,000 km orbit, are perturbations that great?
The mascon perturbations are much smaller in higher altitudes, but then you have more perturbations from the Earth. Nonetheless, I think by this point both of those are sufficiently well characterized that it wouldn't be a significant problem for navigation.
In response to your previous question: yes, IMO it makes a great deal of sense to combine communication relay capability with navigation for constellations orbiting around the moon and other planets.
-
#25
by
Danderman
on 01 Feb, 2014 20:47
-
Are very high lunar orbits that impacted by the lumpiness? For a spacecraft in a 10,000 km orbit, are perturbations that great?
The mascon perturbations are much smaller in higher altitudes, but then you have more perturbations from the Earth. Nonetheless, I think by this point both of those are sufficiently well characterized that it wouldn't be a significant problem for navigation.
In response to your previous question: yes, IMO it makes a great deal of sense to combine communication relay capability with navigation for constellations orbiting around the moon and other planets.
Thank you.
Now I have to find a Mars Q&A to figure out if navigational data can be incorporated into the Elektra data stream.
-
#26
by
Jim
on 01 Feb, 2014 21:36
-
Now I have to find a Mars Q&A to figure out if navigational data can be incorporated into the Elektra data stream.
Not autonomously. Navigation data is produced on earth and uplinked. That is the way TDRSS works.
-
#27
by
Danderman
on 01 Feb, 2014 21:54
-
Now I have to find a Mars Q&A to figure out if navigational data can be incorporated into the Elektra data stream.
Not autonomously. Navigation data is produced on earth and uplinked. That is the way TDRSS works.
TDRSS receives navigation data for its own use, or does TDRSS re-broadcast it?
-
#28
by
baldusi
on 01 Feb, 2014 22:17
-
Please remember that the GPS constellation requires constant orbit determination by the controllers, to feed back that info into each bird. I understand that the onboard atomic clock are constantly checked for consiatency and accuracy. Thus, you need a lot of ground support to keep a GNSS system as accurate as current implementation. And the atmospheric diaturbance can be eliminated by using two, different frequencies.
So, you'd lose a lot of your accuracy. And for slow moving targets, like a rover, it's probably easier to just use the startracker. You know, like sailors have been doing for centuries.
-
#29
by
kevin-rf
on 02 Feb, 2014 04:25
-
I believe the word is sextant
-
#30
by
Jim
on 02 Feb, 2014 21:18
-
TDRSS receives navigation data for its own use, or does TDRSS re-broadcast it?
TDRSs are just orbital tracking stations. They provide ranging data to the ground which processes it and uplinks a state vector to the spacecraft that needs a nav update.
