Flight - NAV_POSHOLD causes copter to jitter

[digitalentity]

Poshold introduces alot of jitter to the quad, automatically applied rate control is not smooth. Some sort of LPF is probably needed.

[digitalentity]

Porting this over might help:
https://github.com/Crashpilot1000/HarakiriWebstore1/blob/master/src/mw.c#L763
Should get and examine some blackbox logs to investigate the cause. Possibly IMU code is overresponsive to vibration. Adding an IIR LPF here https://github.com/digitalentity/cleanflight/blob/navigation-rewrite/src/main/flight/imu.c#L203 might help.

[digitalentity]

fa0e6ccda374762663ad057c131e7ec700496a5a should improve the situation. I'll remember of YAGNI and consider adding a LPF to IMU velocities after I get some blackbox logs showing that the jitter is actually there.

[digitalentity]

Applying expo curve to attitude adjustment improved the situation, but jitter is still there. Probably some PosR PID tuning is required.

[digitalentity]

Video of jitter during POSHOLD filmed by onboard camera - http://youtu.be/9tJJ4sl6iJw
The wind was rather strong - about 12-15mph.

[digitalentity]

From blackbox logs it seems that actual jitter is caused by heading hold. Will try disabling it and test again tomorrow when my batteries will be all charged.
EDIT: Preliminary fix for this is ready to be implemented in the code, but I need to confirm that heading hold is the cause of jutter.
EDIT2: The problem was actually my re-implementation of heading PID right from the start. Should be fixed now. Yaw expo and heading error LPF are still being considered if the fix won't work.

[digitalentity]

It was yaw indeed. Fixes in 4d5bc1b07745eed38c27d2214a41c571182e36f5 and a4e9f90da3c1b6c7e944d8324cfd1d4e802b83c3