FwPositionControl: don't do takeoff help for vtol

[RomanBapst]

• takeoff help is used for fixed wings, it increases the altitude setpoint
  after a launch. A vtol does not need this as it's already sufficiently
  high up in the air.

Signed-off-by: Roman bapstroman@gmail.com

[RomanBapst]

If anybody has been wondering why in the gazebo standard vtol simulation the vehicle climbed like crazy after a front transition just to descend again later...

[dagar]

This helper function is only used in altitude and position control modes, although I don't see why we aren't using it everywhere.

[dagar]

1. On second thought I don't think we should change do_takeoff_help() for VTOL (or at least not specifically for VTOL). All it's doing is forcing a climb in altitude or position control modes if you're close to the ground after takeoff.

2. This doesn't fix VTOL standard climb after transition. See comparison below.

master (vtol standard CI test)

• http://ci.px4.io:8080/blue/organizations/jenkins/Firmware/detail/master/424/pipeline/826
• https://logs.px4.io/plot_app?log=d3f6999d-c4dd-4cd1-b5f4-4b067e9e8e45

PR (vtol standard CI test)

• http://ci.px4.io:8080/blue/organizations/jenkins/Firmware/detail/pr-takeoff_help/4/pipeline/823
• https://logs.px4.io/plot_app?log=6cbebc08-82a5-43a8-b896-b07bd24ce212
  

[dagar]

It's likely something to do with proper TECS initialization (or lack thereof).

[dagar]

On second thought I don't think we should change do_takeoff_help() for VTOL (or at least not specifically for VTOL). All it's doing is forcing a climb in altitude or position control modes if you're close to the ground after takeoff.

[RomanBapst]

@dagar I think I'm already a step ahead. I was implementing front transition using TECS and discovered this behavior. With current master you don't see this because I assume TECS is only activated after transition.
I currently put TECS into no-airspeed mode during the transition and give it a predefined throttle ramp up.
Like this we can have it control altitude during the front transition and it's states (except the throttle states) will be initialized after transition. In SITL I get smooth pitch and throttle setpoints and we can avoid having to hack altitude controllers for the transition into the vtol module. I'm not entirely sure how it will work in practice but it's worth a shot. I'm currently testing in SITL and will push a PR most likely today.

[bresch]

@RomanBapst Why not just setting FW_CLIMBOUT_DIFF to zero for a VTOL if it is just used for FW takeoff and landing?

[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.

[stale]

Closing as stale.

[RomanBapst]

@dagar I think this should be fixed for the release because as is it's broken and it causes the plane to increase altitude after transition.

Since the FW position controller does not run during hover mode the ground reference altitude is not set correctly, it's actually set as soon as the vehicle goes into transition mode.
That means that currently for VTOL we always get an increase in altitude after transition no matter how far you are from the ground. I've confirmed that this does not happen if you set the climb-out diff parameter to 0. But that is not the solution as this parameter is used in other parts of the code, not just for the takeoff help.
Additionally, even if we fix the ground reference we still only allow takeoff help for a duration of 10 seconds after the vehicle goes in air. Practically this means that the takeoff help is anyway useless for VTOL as usually the transition to FW takes place after that.

Let me know what you think.

[dagar]

Alright, can we start by getting it rebased?

[RomanBapst]

@dagar No, if we agree how to do it I will quickly do a PR. I just reopened this as a placeholder.

[RomanBapst]

@dagar We will test this in the afternoon.

[bresch]

Makes sense to me.

[RomanBapst]

@PX4/testflights Please test on the deltaquad. The vehicle should not climb noticeably after the front transition (assuming you don't command it to climb).

[dagar]

I've given this a quick test in SITL to verify the behavior. @PX4/testflights please still test this as a VTOL sanity check for the release.