Indoor Flight Crash(in Stablized mode) based on EKF2

[wkyoun]

The purpose of this indoor flight was to test whether UAV can fly in stabilized mode without relying on nothing(No GPS, No optical Flow, No Lidar Lite measurement update was used).

Since we are not relying on GPS, Optical Flow, Lidar Lite, EKF2 measurement update was executed only for Down position from barometer, and that's why other innovation value is always zero.

The flight mode of UAV was stabilized mode during the flight.
The takeoff of UAV was okay, and hovering of UAV was okay at the first time,

but after a few seconds, I don't know why UAV fly in weired direction, and crashed.
(So, I changed into kill switch mode)

https://review.px4.io/plot_app?log=21bc6c2a-7932-408a-bfd9-822ce1de5473

Would you please look at the log to see what might be the reason?
@priseborough @mhkabir @dagar

We used the external magnetometer from GPS, and when crash happen 200 seconds(3 minute 20 sec),
and connection of magnetometer from GPS was disconnected, and that's why magnetometer norm and innovation value become crazy value at the time(200 seconds(3 minute 20 sec))

2018-3-30-01-57-40_indoor_crash.ulg.pdf

From above pdf, the performance of EKF2 was not bad, and sensor calibration and title alignment and yaw alignment was completed appropriately.

I guess that the crash might be caused by lack of my flight control skill, but I am not sure.
I certainly felt that I cannot control the UAV