I've heard that we could use two or more handheld GPS to get a better positioning accuracy using post processing techniques. Anybody can tell me the methodology? Also, can we program handheld GPS such that it records coordinates every second?

  • 1
    Regarding that last question I would assume it is highly depending on the model you have. Add that information or read the manual would be my tip. – Martin Jan 9 '14 at 9:08
  • 1
    I have read about it some time ago, but for what I saw, You would need more that a couple of handheld GPS to get sub-meter accuracy. If I remember, the problem would be to garantee that both gps are using the exact same satellite to calculate their position, and that is quite difficult. – Alexandre Neto Jan 9 '14 at 9:15
  • 1
    It might be possible if one of those is known to have exact position. github.com/tridge/pyUblox (from the guy who started samba and rsync) is an implementation using commodity hardware, planned for realtime landing of a UAV. There was a recent video of it at linux.conf.au 2014, but I don't think the video is available yet. – BradHards Jan 9 '14 at 10:09
  • 1
    Not to push you away from GIS.stack - but how about asking at beerleg.com or some site with professional surveyors. – Barrett Jan 9 '14 at 18:59
  • Update to my previous comment: video is now up at mirror.linux.org.au/linux.conf.au/2014/Wednesday/… – BradHards Jan 13 '14 at 7:15

Mobile Mapper 120 (around 4000$ with Glonass Option)

And a Trimble Geo XT (around 8000$ glonass option)

We use it to do forest inventory, the goal is to send people where we want them to go and not let them choose an area where there is less trees to count (because garmin or other old gps without glonass were not as precise and they had a distance flaws).

95% of the time, points are below a 3.5m distance of the intended target for all brand of gps after post-processing at the office (in the woods). Sometimes 5cm, sometimes 3.4m, it all depend of a variety of factors (weather, terrain, time of day, distance of the base for the post-processing correction, number of satellites, etc...)

But the best of those kind of gps is thier accuracy in the field. You go right on the target.

Best and bad of each GPS:

Trimble geo XT = a tank, it's heavy and big but handled, it never miss to record data for post-processing. It's a little difficult to learn to use it but when you are use to it, you like it. Problem : Cost a lot, the internal software as a annual licence fee to upgrade it (around 600$-800$). Same for the post-processing software Trimble geo office (expensive annual licence fee).

When you record, it record everything (bad or good signal) and you sort it with the post-processing software.

it use waypoints thay you create with the post-processing software, that it's kind of pain in the ass to create.

Why we have just one, because of the price.(and none of the companies that works with us bought it for the same reason) We can have 2 MM120 for the price of one XT.

MM100 (ashtech) and MM120 (spectra precision/trimble) are similar, the only difference is the présentation of the internal software. A littlle smaller very good GPS with SBAS/GLonass features.

The good is you create your target (waypoints) in arc/gis and you put the shapefile in the GPS, for arg/gis user it's the best. Easy and user friendly. And the result is a shapefile/points with data for post-processing.

You can choose your PDOP maximal and the GPS will not record the signal above it.

For us, in the woods, we set up a pdop of 4 and 99% it works well.

If we dont get a signal, we put it at 6 for getting a position (rare)

The bad, is that sometimes (not often) it bugs for unknown reasons. you have to use it seriously, step by step like a robot. And sometimes the data for post-processing is missing...but never the shapefile/point before post-processing.

The internal software is free of licence and free for upgrating as the software for post-processing.

I recommand all those GPS. Very accurate.

We use them for a specific purpose but they can do lot more


First of all, this is not possible with any type of handeld GPS. You need a GPS that is able to store the position of the satellites, not only the average position on the ground.

Then, the idea is to have one or more GPS at a fixed (and if possible precisely known) location which will keep recording all the position of the satellites during the time you travel with your other hand held GPS. Those GPS are called "base stations".

When you are done, you use the information from the fix GPS to post-process your mobile GPS information (see Post processing GPS data with open source software for some softwares).

Note that the base station must be precisely located. This can be done by selecting a point of known coordinates or by taking a lot of measure. The confidence interval on your true position indeed decrease if you take a lot of measurement, but there is a temporal auto-correlation of the errors so that if you stay only a few minutes your average value is biased. Several hours is recommended.

Note that you may find base stations with open data in many parts of the world, which can be used at a few hundred km for metric accuracy.


Recording coordinates every second is possible on most handheld GPS receivers.

For Garmin models eTrex Legend / Vista / Venture onwards, turn on track logging and set recording mode to "time". You can set a sampling interval of 1 second. Note that the time recording mode is not available within earlier models (basically that ones without navigation joystick, e.g. eTrex Summit or eTrex).

For Garmin models eTrex 10/20/30 a track log is saved automatically. Go to Setup > Tracks > Track Log. Select Record, Do Not Show or Record, Show On Map. Select Record method: time and enter a Record interval of 1 s.)

However, note that due to a maximum of 10000 points per track on stated models, you will be able to record less than 3 hours of data. (Maybe it is worth considering a slightly longer sampling interval in favor of a longer overall measurement.) For comparison charts of different handheld GPS receivers, have a look at the OpenStreetMap Wiki, e.g. Garmin eTrex comparison.

Rather than logging GPS position internally with your handheld GPS and post-process later, it could be more effective to log the NMEA stream of the GPS device (e.g. by connecting it to a computer during measurement). Besides the position data, log the NMEA sentences GPGSA and GPGSV (overall and detailed satellite data) to record which satellites are used for positioning of each point. (An overview of NMEA sentences are provided on gpsinformation site for example.)

It is worthwhile to note that several handheld (!) GPS are ready to use a RTCM SC-104 compliant real time correction information to improve accuracy. There are lots of different sources for such a correction, the most professional being a DGPS base station, but there are also some regional correction signals available which can be received by a GSM modem or other radio techniques. (e.g., Garmin even sold -now discontinued- beacon receivers in their marine product range (GBR-21 and 23) but there are lots of other products.)

So your options could be:

  • Set up a stationary base device and a moving one. The base station provides the correction for the moving device (the "rover"). Either use real-time-correction via a radio link, or do post-processing, using the base station position variations and the information which satellites have been used.

  • use a correction signal from an external source, e.g. a beacon receiver.

  • Set up a grid of handheld GPS device that real-time cross-sync via WLAN to correct themselves. (That's what a colleague did, I try to get more information about that.)

As a prerequisite, you should only use devices that support build-in WAAS (Wide Area Augmentation System) correction. Sometimes this feature is not turned on as a default because of power saving reasons.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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