I have found the accuracy of Reach RS base improved dramatically (from meters to centimeters) when connected to e-GNSS network in my country. I am thinking if I could have the rover unit on UAV also connection to e-GNSS network, accuracy of photos geotagging would be also greatly improved. It is not possible with wifi because I often work in rural area with no wifi and the UAV is easily out of range of cellphone hotspot.
Is it possible to use a 4G/LTE device, so the rove on UAV can be connected to e-GNSS network?
Thank you for your response. Yes, I have been doing PPK. But I would like to improve the accuracy of GPS data synchronized to each photo before PPK process. I believe better data quality should lead to better post-processed results. Right now, I still need GCP get satisfactory result. With rover being connected e-GNSS network, I am hoping to do it with less or even no more GCPs.
Would you be so kind to provide any suggestions how to integrate a Reach rover unit with a 4G/LTE device?
I cant see how this would help getting better accuracy of your geotagged images. Flying more accurate doesnt geotag your images better, unless you fly free handed and banging your TX stick continuously.
Your onboard controller/navigation with stock gps should fly your drone within few meters of plotted route stable enough.
So, unless you need your drone to fly with few inches accuracy, for reasons i cant come up with right now. PPK would be way to go and rather improve accuracy in other areas.
RTK as i see it and for the mission explained, this would introduce more things that could go wrong and worst case, degrade your PPK accuracy with more noise. But i`ve been wrong before and if someone knows something i dont, please tell and i would go back to playing with my Lego
But this should probably get you going with the LTE thing
Sorry, I forgot to mentioned that the Reach rover is not plugged into Pixhawk but connected to camera via a cable attached to the hotshoe. Pixhawk uses another M8N GPS. So my goal is not about flying more accurate but about geotagging more accurate.
@sygong, I haven’t had commented on an Emlid post in a lot time, but for this I can’t resist. Flying more “accurately” wastes a lot of energy. The aircraft irrespective if its a plane or multicoptor has to make so many corrections to keep it on course. This is battery power that should rather be used for endurance.
In terms of photogrammetry, RTK is fairly useless. A few meters error or uncertainty in your flight line will not upset your side/overlaps much at a useful AGL. RTK is a nice feature to have for perhaps insepction work where you need to be very stable and consistent with your camera angle/position. But other than that, its a bit of a gimmick. However, the feed from uncorrected Reach back in the autopilot is very useful as a second GPS to aid redundancy. But this does not need to be corrected with a base as RTK.
Thanks to your advices. I didn’t explain my question well. Let me try again. The Reach rover unit is NOT connected to pixhawk on my UAV, but to the CAMERA. So I am NOT trying to fly accurately. When each photo was shot, the camera sent a signal to rover unit via a cable attached the hotshoe, to synchronize with GPS timestamp. Then I do PPK after landing.
I found that the accuracy of Reach RS base improve from meters to centimeters when connected to e-GNSS network (via wifi from cellphone hotspot) in my country. So I am thinking the accuracy of rover (and the geotags of photos) could be improved if the rover unit connected to e-GNSS network. But I don’t know how. The rover will certainly be out of range of wifi hotspot.
Thanks for the thread. I will try if a 3G or 4G dongle will work.
I have learned that, when on the ground, Reach rover can be connected to the base unit via the their radios and to e_GNSS network via wifi at the same time. If the rover will maintain connection with e_GNSS network, I assume the difference is that its GPS data is from e_GNSS rather then directly from satellites via rover’s own antenna (or I am wrong?). I don’t think this is in RTK mode.
Ok. If i where you i would stick one of my RS units to a tripod somewhere in the same area you do fly.
Next i would connect to your e-GNSS network and fix base to a known absolute position and save it. Then start log on base/rover and fly the PPK mission. This i think is the most accurate way.
1.You have the base fixed and static for a long time to get the most accurate absolute position from your ntrip service. This step if given some time should give you more then enough accuracy in your absolute position.
2. when processing rover against base, you then have the best setup with the shortes baseline to get the most accurate relative data possible.
Another way (if possible) is to download historic data from e-GNSS and PPK againts the rover log. But i would rank this lower then the options above regarding accuracy.
Edit: If using multiple refrencestation, you would still do it the PPK way.
1.You have the base fixed and static for a long time to get the most accurate absolute position from your ntrip service. This step if given some time should give you more then enough accuracy in your absolute position.
This is what I do now, and the errors are less than 10 cm after GCP corrections in Photoscan. But I like to try if I can achieve same accuracy without GCPs with the rover unit also connected to e_GNSS.
You could, but it would not be better with rtk. PPK is still the way to go.
Lets say your base is spot on, and the e-gnss as well but your rover is not and gives you 10cm error when process against the base. Next would be to add e-GNSS correction and see if this makes any difference, but you would need to have e-GNSS data logged somewhere, e.g to the base.
Now, e-GNSS might pick up satelites that your base dont but rover does and this way increase accuracy.
But your 10cm error seems to appear from something els the lack of refrence stations…
