* ekf2_derivation: use single source of state definition
The state is defined as an ordered dictionary of group elements and
everything else is generated using that state definition
* ekf2: generated state sample add const reference getter
---------
Co-authored-by: bresch <[brescianimathieu@gmail.com](mailto:brescianimathieu@gmail.com)>
Co-authored-by: Daniel Agar <daniel@agar.ca>
The sparse vector template requires to know which states are non-zero in
the observation jacobian. This complicates the modularity of the code
when the state vector or the derivation is changed.
The computation cost difference is almost negligible for this size.
For FW takeoffs we need to keep the course straight towards the first
waypoint, and not climb in loiter patterns first.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Dataman: Add write function to dataman cache.
RTL and mission have a new common base class mission_base. Both inherit from them and mission, RTL mission, and rtl reverse mission inherit from them and implement their desired functionalities. This simplifies the logic in mission as well as make the logic in rtl mission reverse and mission more readable.
Rtl mission reverse now functional again for VTOL flying back the mission and transitioning to MC at the home position.
Dataman cache has new write functionality to write to dataman while updating write item in its cache if necessary.
Dataman cache is now only updated when the respective module is active. Leads to a higher computation time once on activation, but decreases unnecessary cache updates when inactive.
This is required to support fixed pitch propeller helicopters that have
no collective but only cyclic pitch with two degrees of freedom and
hence only two servos.
The amount of thrust in the body z axis is then controlled using the
motor speed which makes particularly sense on coaxial helicopters
that need to control yaw with changing motor speeds already.
- Preserve tilt variance while doing a yaw reset
- Yaw variance is now correctly set instead of increased
---------
Co-authored-by: Dominique <dominique.leblanc@donecle.com>
Co-authored-by: bresch <[brescianimathieu@gmail.com](mailto:brescianimathieu@gmail.com)>
Even if there is some horizontal motion, a passing check should be
accepted as the terrain can be flat. However, the vehicle must not be
moving horizontally to invalidate the consistency as a change in terrain
can make the kinematic check temporarily fail.
A setpoint of type IDLE can be published by Navigator without a valid position,
and should be handled in the auto function in FW_POSCTRL_MODE_AUTO. If it wouldn't be
handled there the controller would switch to mode FW_POSCTRL_MODE_OTHER, in which case
no attitude setpoint is published and the lower controllers would be stuck with the
last published value (incl. thrust).
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
