Congrats! I just ran a data set through from an Autel Evo2 Ent. drone and the result was FLOAT for 100% of camera locations. Using all the same files and processing with RedToolbox results in 98.8% cameras FIXED and 1.2% FLOAT. So, something is certainly different in the way ES interprets the files.
I just ran another set and Emlid studio returned 94% FIXED solutions. So maybe there is something specific in the first set that is causing it to not FIX.
RedToolbox processing resulted in 100% FIXED on this project.
I’ve managed to obtain a 96.5% of Fix by changing the Filter type to Combined. SNR on a drone is slightly jumping, and the base’s log interval is 30 seconds. That’s why processing in both directions gives better results, I think.
Thanks Svetlana. I’ll get back to trying this out again with your suggestions. I have spoken to Hans (RedToolbox dev) about that project (which is just a test) and he said the formula for extrapolating the exact time from the Base log when the interval is 30 seconds is some of the magic sauce. That log was in “archive” mode when I pulled it from the CORS site weeks after the date. At the time, I had not thought about using PPK for that set. That’s why it was an archived log. If the CORS logs are pulled sooner, they are 1hz. and much less interpolation required (better) .