Attitude glitches in velocity control

[bresch]

Describe the bug
After flying with tightly tuned multicopters in position control mode (basically a racer with a GPS), I realized some attitude twitches that I don't see in stabilized mode. I'm now able to see that those twitches are created by the velocity controller.
In the log below, you can see that the velocity setpoint and estimate is perfectly smooth and that the thrust/pitch/pitchrate setpoint have glitches. Pitch and pitchrate setpoint glitches are a consequence of the thrust glitch and the thrust setpoint should have the shape of the error between the velocity estimate and setpoint (for my tests, I removed the integral and derivative gains to simplify the problem).

Reproduction in SITL with fullrate logging of the relevant topics:

To Reproduce

• jMAVSim test in position control mode (default smoothing or velocity smoothing).
• log vehicle_attitude_setpoint, vehicle_local_position and vehicle_local_positon_setpoint at full speed
• set MPC_VEL_MANUAL to a small value to be able to give constant full stick and stay far from any saturation (3 m\s for example)
• takeoff and fly full stick straight to the North

Expected behavior
A smoooooooth flight

Log Files and Screenshots
https://logs.px4.io/plot_app?log=a63b076d-b174-40d0-80f3-a57bc4f6a87b

@bkueng @MaEtUgR @Stifael Can you also reproduce the issue?

[bresch]

It looks that the glitches in SITL are just due to jitter in the execution of the control loop. When the setpoint is generated later than the estimate, it has time to move and this results in a glitch.

Now I'm not sure if this is what I see on the real vehicle and will fly with high logging rate to confirm that.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[bresch]

I've not seen that problem anymore.