7
\$\begingroup\$

I am currently looking for an antenna to determine GPS location. The question is a general one, but I have listed the equipment that I am using to help ensure that answers are more likely to be applicable to my situation.

The antenna will be connected to a Teltonika 955 router.
I have an antenna from the same brand, Teltonika PR1KCL28.

Question: In theory can I achieve a significantly more accurate GPS position with some antennas than others on a given GPS router.

Note, I'm not asking for recommendations for specific products.

\$\endgroup\$
4
  • 1
    \$\begingroup\$ My understanding is that there are more accurate antennas based on rejecting reflections--direct path to the satellite or no solution. It's not my field, though, this is just something I've read about it. \$\endgroup\$ Commented Aug 18, 2023 at 0:06
  • 1
    \$\begingroup\$ The more satellites you can receive with your antenna, the better the solution you're likely to compute. So clearly a good antenna, positioned well, can improve the positional accuracy. Once you have good signal from all the visible satellites, you're unlikely to get significant further improvement. \$\endgroup\$ Commented Aug 18, 2023 at 8:36
  • \$\begingroup\$ Simple averaging will help with some sources of error (random ones) but not others (systematic). There are likely to be better software filters to use to discard outliers; some of these will depend on your expected rate of change of coordinates or direction \$\endgroup\$
    – Chris H
    Commented Aug 18, 2023 at 12:34
  • 1
    \$\begingroup\$ I believe you can (with a good amount of tinkering) use an NTRIP client on your RUT955 to do RTK. That will give you so much more accuracy than changing the antenna. \$\endgroup\$
    – AndreKR
    Commented Aug 18, 2023 at 14:52

5 Answers 5

13
\$\begingroup\$

The location that a GPS receiver computes is the location of the antenna's "phase center", which is the point relative to the physical structure of the antenna at which rays from all directions have the same delay.

A poorly-constructed antenna might have an ill-defined phase center, in which case, the computed position will vary with the relative position of the satellites, increasing the overall error beyond what the receiver is capable of. But once you have an adequate antenna, the overall error budget is dominated by the receiver itself, so further improvements to the antenna will have negligible effect.

\$\endgroup\$
6
  • \$\begingroup\$ Is there a way for me to distinguish a poorly-constructed antenna from a good one? Can I derive this from the specifications for instance ? \$\endgroup\$
    – cobdmg
    Commented Aug 18, 2023 at 13:04
  • 2
    \$\begingroup\$ Even with Wide Area Augmentation System (WAAS), GPS accuracy is around 3 meters. Do you think that changing the phase center of the GPS antenna by an inch or two is going to affect that accuracy? \$\endgroup\$
    – SteveSh
    Commented Aug 18, 2023 at 14:14
  • \$\begingroup\$ Hm, wouldn't the phase center always lie somewhere inside the "hull" of the shape? So I suppose it would be a difference of a few cm at most \$\endgroup\$
    – pipe
    Commented Aug 18, 2023 at 14:46
  • \$\begingroup\$ @cobdmg: It's been a while since I last looked for antennas, but a general guideline would be that the antennas whose datasheets mention "phase center" at all are probably worth considering; the ones that don't should probably be avoided. \$\endgroup\$
    – Dave Tweed
    Commented Aug 18, 2023 at 15:43
  • \$\begingroup\$ @SteveSh: There are many kinds of GPS receivers. Precision (cm level) surveying is done using RTK (real-time kinematic) receivers that use techniques far beyond what commercial modules can do, such as directly measuring the carrier (not code) phase angles. \$\endgroup\$
    – Dave Tweed
    Commented Aug 18, 2023 at 15:45
6
\$\begingroup\$

A guess: No remarkable accuracy lifts are available by getting a higher spec antenna. You'll stay in the same accuracy league as phones or hiking navigators. That means 3 meter accuracy at best. The variations in the atmosphere can easily add 10 meters more error. Car navigators try to fix it by guessing that the user drives consistently on a known road, but your device probably doesn't make such assumptions.

Stepping to surveyor's or guided roadwork bulldozer accuracy league needs more advanced equipment which are based on finer analysis of satellite signals and may utilize carrier phase angle and even not-commonly available high bitrate codes in the signal. They also are connected to a reference station which has known coordinates. The known local reference makes possible to take into the account the variations in the all-time restless atmosphere. With direct carrier phase detection and a reference the accuracy can be as good as 1...2 centimeters.

As Dave Tweed notes, the centimeter-class accuracy, of course, needs a high spec antenna which has a well defined phase center for reliable phase angle comparisons for different satellite signals. I guess you do not have such equipment, so a high spec antenna will be useless.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Not an answer to the question. The errors that dGPS removes are unrelated to antenna-induced errors. \$\endgroup\$
    – Dave Tweed
    Commented Aug 17, 2023 at 17:27
  • 1
    \$\begingroup\$ I did a rollback on your edit. The upvotes indicate that your answer does indeed contain some valuable information. If you really want to remove this answer, you can use the delete feature. But changing the content to "An useless answer. Deleted the content." isn't the way to go. \$\endgroup\$
    – Velvet
    Commented Aug 18, 2023 at 9:19
  • \$\begingroup\$ Ft. Equity - I wasn't trying to annoy you with my edit. As I said in the edit comment - the answer was useful. Not very well targeted at the original question BUT still useful to others. \$\endgroup\$
    – Russell McMahon
    Commented Aug 18, 2023 at 15:36
3
\$\begingroup\$

The only situation where you’ll get a significant improvement is in a situation where the antenna gets a marginal signal some or most of the time. It’s important to understand that since the satellites are in low orbit, they will appear fairly close to the horizon most of the time, you might only get one at a time that’s above 45 degrees elevation. Therefore the location of the antenna is generally more important than the quality of the antenna itself. For really good coverage the antenna needs line-of-sight to the horizon in every direction.

\$\endgroup\$
2
\$\begingroup\$

There are antennas out there that promise more, for certain applications. The antenna you've linked looks appropriate to use with your receiver.

An example of a better antennas, used in highest precision applications (such as stationary reference stations) are 'choke ring' antennas. A precise, CNC-machined shape attenuates signals arriving from the sides or bottom, without interfering with signals arriving from the top. This reduces the amount of interference from signals that have reflected off the ground.

The receiver you linked is probably not precise enough to benefit from these costly antennas.

Choke ring antenna

(Image from wikipedia)

An example of a worse antenna can be found in sports-oriented GPS receivers and phones. They tend to use an 'omnidirectional' antenna, so reception is equally good whatever angle you're holding the device at. This of course reduces their ability to reject multipath noise.

The antenna you've got is a patch antenna, which is directional but not as expensive as a full choke ring antenna. Your antenna also has an 'LNA' or 'Low Noise Amplifier' which amplifies the signal as it goes down the 3m cable to your receiver.

This setup is normal for systems like yours - roof mounted, but not costly multi-frequency survey equipment.

If you are having GPS reception problems there are a number of things you can check:

  • I see your antenna has 5 cables, 4 of which are for 'LTE' - are you completely sure you've got the GPS antenna connected to the GPS receiver?
  • Is your antenna mounted high up, ideally at roof level, with a good view of the sky? This isn't always vital (cell phones can navigate just fine from ground level inside a metal car) but if you're having reception problems it makes sense to check.
  • You can get software for GPS receivers that will create a 'sky plot' over the course of a day or so, showing the signal strength of different satellites at different positions in the sky - this will tell you if nearby buildings etc are blocking signals from certain places in the sky gps sky plot (Image from Future Automotive GNSS Positioning in Urban Scenarios)
  • If you've got worries that your antenna might be broken, patch antennas don't cost much. Spend $30, see if the problem disappears.
\$\endgroup\$
0
\$\begingroup\$

Yes, somewhat, to a point. Then no.

GNSS (more generic term for "GPS" as GPS is american) signals are broadcasting the time of transmission, where they are, and where the other satellites are, that's it. Receivers pick several satellites that they "hear", and for each, look at the receive timestamps, update their drifting clock, and do the triangulation when at least 3 satellites are in view.

Each satellite you track beyond 3 satellites increases accuracy, since the receiver is able to average out each measurement. But the receiver can only process signals that it can hear, and it can only process as many satellites at the same time as it has channels.

Based on this, you can:

  • Upgrade your receiver for a better clock or for a higher number of channels (i.e. number of signals it listens to at the same time)
  • Increase the number of satellites in view by increasing the signal to noise ratio of the receiving chain (antenna, amplifier, matching...), if the receiver's channels are not already working 100%.

Item 2 is within the scope of the question. One easy way would be to increase the gain of an antenna, but unless you're willing to get a physically bigger one, more gain with same size means more directional, which could mean less satellites in its beam.

In any case the return diminishes as some errors can't be averaged out in this way, and even with infinite channels, there are only up to ~15 GNSS satellites visible. To get more accuracy, you'll need either nearby reference stations (e.g. RTK, DGPS), or combine the readings of a GPS with those of a more accurate (but drifting) Inertial Measurements Unit. This is called sensor fusion, and usually done using a Kalman filter. That's how smartphones are able to track your position through tunnels when no satellite is in view.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.