I am a bit confused by the way you indicate time. The flight lasted 14Min 32.8Sec. For reference are you using Ticks or something else?
I looked at the Graphs and I have seen this same pattern on my own Mavic when it is indoors. From the information I have looked at I came to the conclusion that if the IMU detects almost no movement it seems to keep the flight state in GPS mode even if the GPS signal has dropped out.
So far I have only seen this happen when the GPS drops out for very short periods and the craft has minimal movement. It might be DJI has programmed in a hysteresis effect in case the GPS lock is going in and out very fast. I think this flight demonstrates that happening.
I sampled the last bit of the Flight.
GPS Data
View attachment 13103
Movement during that period
View attachment 13104
Rob
I am also confused about his timestamp... The meaning of "32" at begining???
Your graphics are very good... I believe that between 828 and 837, when the red GPS line drops (by the end of my flight) that was the moment I crossed my window and flew indoors. At this moment RC screen changed to OPTI and Mavic became much more stable. I flew 30sec (+/-) before landing and shutdown (manually).
I believe your hysteresis theory is very right!