I am currently working with tractors using a RTK enhanced automated steering system. One of the biggest problems is that even RTK can have a inaccuracy of 2cm. This can be too much for some possible field applications like sowing. Therefore I am trying to determine the inaccuracy of such a steering system in dependence to the distance of the RTK-station. My problem right now is, that I don't have a way to measure the inaccuracy of the tractor driving the track.

In my current setup I want the tractor to drive 100m straight between two given points. Thereby I want to measure the inaccuracy. This I want to repeat with several distances to the RTK-station (1km, 2km, 3km, 4km, 5km, 10km, 15km, 20km, 30km, 40km). I hope that I can get a function which tells the inaccuracy in dependence to the distance of the RTK-station.


  • Is there a way to process the points given by the GPS device to determine the inaccuracy?
  • Do you know any visual validation method? (with stereo cameras or something)
  • Do you have any other idea of how to measure the inaccuracy ?

Assuming you have line-of-sight to the whole of the route you want to drive from a single observation point, an optical TotalStation mounted at that observation point would allow you to track a prism mounted on your tractor with millimeter accuracy. You'd need two surveyed locations to set up the measurements (usually the observation point and another location which is visible from the observation point).

Once you have the actual track of the tractor from the TotalStation, you can compare this with the location given by the GPS. Since you want to compare prism location with GPS location (probably on a second-by-second basis), you'd probably mount the GPS antenna directly on top of the prism (prisms often have a 5/8" thread top and bottom, so they screw onto standard antenna mounts).

Disclaimer: I'm not a surveyor so I can't give details, but I've seen this done.

| improve this answer | |
  • For this setup to function properly, you would need a robotic total station capable of tracking the moving prism. The prism would preferably be an omni-directional prism or a 360° prism and the total station can be practically setup anywhere that has a line of sight to the track the tractor is following since I suppose you already have the coordinates of the start and end point of the line. This is much more practical when doing several lines in a row as no change of station is needed. – Techie_Gus Feb 26 '17 at 18:36
  • Thanks for the detail. Doesn't look like the original poster cares. . . – Trams Mar 17 '17 at 21:26

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