Portable RTK Receiver/Base for $799

Discussion in 'Precision Farming & GPS' started by Pheasant Surprise, Jun 18, 2017.

  1. Robt

    Robt Member

    Fair play!
    However you aren't an " average farmer"
    That is some seriou skill! You seem a new member, what's your background? Want a job ;)
  2. Charles Quick

    Charles Quick Member

    Not much background to speak of, we've always been into tinkering on this farm so I had a background interest in electronics, last year I took a year out of agriculture at Nottingham uni to study Computer Science, started getting interested in GPS while procrastinating during January exams... I had a working autosteer in April but it lacked any sort of 'lookahead' function. The author of AgOpenGPS, Brian's Pure Pursuit algorithm is a much more elegant solution than mine ever could be, so I implemented his with a few minor tweaks in the last couple of weeks.

    It's all really just a bit of fun to be honest! My main 'big idea' at the moment is culturing a biocontrol agent for various fungal crop diseases :)
  3. Northern farmer

    Very nice, even runs on 4G :) (referring to some other recent thread). Cheaper to replace your complete system rather than replacing the old 2G/3G modem when needed.

    Also solves the steering valve controller compatibility, this one is compatible with itself. :)

    Seriously, this obviously isn't an option for every farmer but I understand the whole system is going to follow the open source approach. No need for every farmer to do the R&D then, just purchase parts and take care of the installation. 95% of farmers can do that.

    I asked for the Topcon AES-35 kit price to make one of our current screens portable. Seems it makes no sense, at half the price one can get a complete system. Big names on the market should open their eyes.
  4. DrWazzock

    DrWazzock Member

    Can somebody tell me if I would need two of these devices (one as base station, one as roving receiver) along with an iPad and an app to provide centimetre position information accuracy over an area of about a square mile?

    Sorry if I'm being a bit thick but not really up to date with the technology but would like to be.

    We have a vast underdrain network that needs maintenance from time to time. It would be nice to be able to record locations of drains and junctions accurately. It would also be nice to get rid of all of my drill bout markers using same kit. Is this kit a viable solution?
  5. You would need either:

    (a) a minimum two receivers for a standalone setup - one a base and one rover. Although one base can send corrections to an unlimited number of rovers in theory, using radio or NTRIP (mobile network) to deliver the correction messages to the rover(s).

    (b) a single receiver for the rover and an external correction source - another compatible base or NTRIP service.

    I've not played with these particular receivers but the basis of operation of RTK is the same irrespective.

    Fire away...
  6. DrWazzock

    DrWazzock Member

    Thank you.

    Any ideas of expected accuracy? I know a lot depends on many factors.

    I am presuming that if I have one module as the RTK base then it needs more or less a line of sight to the roving module to avoid interference of the correction signal transmitted as a radio signal.

    I presume you need to know the exact grid reference of the base and programme it in somewhere, so that it knows how to calculate the correction factor.

    L2, which this system doesn't have, is what?

    I need about 2cm accuracy, not the 20 to 30 cm accuracy I presently have with my basic system. I don't want to be paying big subscriptions either.

    I am working my through the literature in the emlid website but am grateful for any simple explanation of the above questions.

  7. Charles Quick

    Charles Quick Member

    Not sure about the Emlid, but the U-blox M8P (https://www.u-blox.com/en/product/c94-m8p) can be set to 'survey in' to any accuracy you like. You just set it up, leave it in the same place for a few days and it'll start producing correction data once it is confident enough in its position, down to the accuracy you specify. I have mine set at 10mm.

    If using the inbuilt radio module then close to line of sight is necessary. If you can get a computer and internet next to the base station, a better idea would be to use an NTRIP caster like SNIP (https://www.use-snip.com) and download the correction data using your tablet. I don't think an iPad would work for what you want to do but there are some good Windows tablets out there.

    L2 is the encrypted military signal and more expensive receivers use it to get a fix more quickly by comparing the signal's phase with L1, I think. It does not improve accuracy; you may have to wait an extra minute for your RTK fix without it.

    Hope this helps,
    DrWazzock likes this.
  8. I guess the actual real world accuracy is the 64 million dollar question. The claim is cm class accuracy, but I haven't got any Emlid receivers so I can't tell you honestly!

    All these low cost receivers are, with very few exceptions, single frequency, L1, types. This makes them far cheaper to produce than dual frequency receivers, L1 and L2. The latter calculate RTK corrections based on something called the pseudorange (from L1 signal) as well as the carrier phase observations (from L2 signal). This makes dual frequency, L1/L2 receivers very accurate (but also very expensive). They are the default from the mainstream precision firms the likes of Trimble, Topcon, Leica, NovAtel etc.

    The main downsides that I'm aware of on an L1 only receiver, are:
    a) the receiver cannot correct for the effects of atmospheric (tropospheric) delays, which on certain days and weather conditions can cause dilution of precision as the signals can take longer or shorter times to travel from space through the atmosphere. This is more noticeable when the base and rover are quite far apart.

    b) The time to get convergence (an RTK fix) Is slower than a dual frequency setup. On a dual frequency unit you should be able to get convergence in under 5 minutes, often 2 or 3 minutes. Again I've read differing views on how long this takes on L1 only receivers, but I've read it could take 30 minutes plus.

    So basically this boils down to how absolutely accurate, how absolutely repeatable (measured in days, weeks, months and years) the "position" will be with L1 only RTK, as against a traditional dual frequency setup, which we have taken for granted will be completely repeatable (as long as the base doesnt shift, more on that in a moment).

    On the base positioning/line if site question. The most important thing is the base position needs to be open sky as possible, with as little shadowing from trees etc up to about 15 degrees from the horizon viewed from the antenna position. It should also be fixed and immovable as you will introduce an offset error should the base position not be exactly the same, or if you're tearing it down and setting it up repeatedly as a portable base, set in up in precisely the same spot. The fixed bolt in an old post is the usual trick.

    The precise absolute reference position of the base is good to have but not a show stopper (unless absolute position is necessary). It's far more important, from a simple repeatability standpoint to set the base up in the exact same fixed position, even if it's absolutely reference co-ordinates are slightly off, IYSWIM.

    Getting a precise reference position with one receiver can be done (but not in real time) using something called post-processing. You run the base for several hours, all the while recoding/logging the observation data. The longer the period of logging data the more accurate the final result, a couple of hours is minimum for good results. The generated log file is sent to one of the free post processing websites (for example Trimble CenterPoint RTX Post Processing) and they calculate the accurate fixed position of the base.

    Strict line of site, from base to rover, isn't necessary for RTK to work. The question for radio is good, because it generally depends on radio output power on the base. Unlicensed is a max of 1 watt, good for a several miles depending on the terrain between base and rover. The Emlid units use a different radio than standard 400Mhz band UHF radios typically used here for RTK. Worth looking into. Otherwise you could set up your own NTRIP box (software can be got for free) and get your own corrections back on the rover using a mobile connection.
    Kiwi Pete and DrWazzock like this.
  9. Gpsvince

    Gpsvince Member

    Very impressive!
  10. Well the AgOpenGPS project got a lot more exciting. Driving by wire is freaky but pretty fun, can steer the tractor with my smartphone over bluetooth! Next steps are to focus more on autonomous operation.

    Contour is certainly the most challenging, but works pretty good. Overlap set to 10 cm. Going at 8 to 15 kmh. No crazy wheel whipping back and forth like most systems - more like hands free in a Benz. Milwaukee 18 V drill as steer motor. About 150$ total for auto steer hardware excluding gps.

  11. Nice work! When you’ve figured out how to hack/code PVED-CL you’ll be ready to conquer the world.:D
  12. Robt

    Robt Member

    Pved-cl is so last year..........
  13. I’m just playing catch-up. Up to 2014 now ;)
    Robt likes this.
  14. Got the headland turn figured. Works quite nicely. Also using an IMU + GPS for stable heading. Super straight.

  15. The next step, doing the whole field autonomously with "smart" turns based on whether or not the pass beside is applied

  16. Northern farmer

    I'm voting Briantee to stop here. It is OK to build applications that make farmers' life easier but now he is going to make farmers redundant. :)
    Pheasant Surprise and Briantee like this.
  17. Off topic, but I think the implications of the automation (r)evolution are pretty staggering...lots of occupations are in the cross hairs. What are we going to do with ourselves :p
  18. Write software? ;)

    I'm sorry but automation of the majority of the farm jobs just isn't going to happen without a monkey (a real farmer) in the seat. Maybe i'm just not techy enough, but this is 1% practical and 99% media just looking for a story to print and hype.
  19. Another video on progress. Have headland sequencing of events, auto headland turning, headland generation based on boundary and touch, and much more. Project is getting really fun now!

    Pheasant Surprise likes this.
  20. Looking really good Brian!

    What’s with the black helicopters reference? Are you expecting guests? ;):alien: [We get the army boys from Wattisham flying in quite low as they return from presumably exercises]

Share This Page