Wondering if anyone out there has some suggestions for how to correctly analyze GPS data for error correlations or trends.
Part of my research involves gaining a better understanding of the position accuracy so that I know how close I can hope to get multiple a/c together into formation.
My idea was to take two MediaTek 3329 recievers, connected to serial ports on an Arduino Mega board, and simultaneously record the output of the receivers on an SDCard using the DIY Drones binary protocol and the same MTK library (slightly modified) used in ArduPlane 2.65. The goal was to get as close to the GPS accuracy I could expect from the APM 2.5 and analyze the error within the posted absolute position error.
From what I can tell from my data, I have no discernible correlation in the errors, but I may not be looking in the right places for trends.
I will attach my test summary if anyone wants to take a look, and give me some feedback.
To evaluate the quality of your GPS data you can take the "Horizontal Dilution of Precision (HDOP)" information besides the fixtype and the number of sats into account. Perhaps it helps?