Why FOPD spread at last?

[cnpcshangbo]

I tried to increase the solver sampling period from 1/100 to 1/1000, but the curve didn't get better.
It is partially because the model was identified for 1/100.
I am not going to continuous this branch now: https://github.com/cnpcshangbo/simscapeFOPD/tree/Testing-solver-1/1000
I am going to use the currently best result to write up my paper.
numerical simulation result 169545d

[cnpcshangbo]

From the video I can see that the FOPD spread at last in roll, pitch and yaw direction. I could lock those dimensions to tune the altitude itself.

https://drive.google.com/file/d/1dZxUR8A1qw6urSF6Zie1-eXuZx_CHXYn/view