I just thought I’d throw it out to the community, and see if anyone has any thoughts on improving the vertical accuracy of RTK devices.
I’m currently using a Facet L band (without L band) for a project. While it is pretty good when its got an RTK fix (10-15mm), it can drift around a bit. I.E if its got 14mm accuracy, its seems like it can be ± 14mm from zero?
For this project, I need to know my elevation as absolutely best I can.
I’d like to know if there is anyway of getting it better?
Should i be using a different nmea message, rather than GGA?
I am currently just averaging the received data out in my app, which seems to help. Just looking for other ideas.
Vertical measurement is very sensitive to tropospheric delay i.e. humidity (quantity of water in atmosphere above the antenna). The first recommendation is to use a reference station with an elevation similar to the one to the rover. I read somewhere no more than +/-300 m. And, if you want to have very precise measurements, for instance for hydrographic studies, I would suggest using a temporary reference station in the middle of the studied area. If possible, select day with stable high-pressure weather instead of stormy one. For the rover, be careful with tree cover. If the receiver fight to get the fix, the fix may be wrong. Better move away to an open area and finish with a classic levelling.
You can establish very accurate coordinates on a position by enabling RAW measurements on your Facet ( & turn off L-Band) and collecting 6 hours of data for Precise Point Positioning (PPP) that’s used by a couple of different online services. This is how a Reference Station is established.
When we commission a typical CORS, we usually log 24-hour datasets for 2 weeks, all treated as individual survey missions. You will get better results if you wait a few days before performing the online processing for a particular mission.
Two weeks is probably overkill for your Project. Six hours of logging data for PPP will get you very accurate results.
The trick is to submit RAW data that’s a few days old for best results. By then, the online service knows the accurate locations of the birds with respect to time (in the past), instead of using their predicted orbits in the position calculations.