Refactor: effective matrix calculation from module to lib

[Perrrewi]

Solved Problem

Refactoring of control allocation started here.
This PR enable the tests and will unlock the computeEffectivenessMatrix in SIH.

Note that the refactoring should not affect the performance, only make it accessible as a library.

Test coverage

• Unit/integration test: Yes
• SIH test with QGC and AMC 34.12

[Jaeyoung-Lim]

This is undoing the changes that were done in the previous PR

[Jaeyoung-Lim]

These were intentionally left as part of the control_allocator module, such that the control allocation library can stay generic.

In theory, there is no reason for the actuator effectiveness to be tied to vehicle types. The reason that this exists, is tied to how the control_allocator module is implemented. This is why I wanted to leave this out of the control_allocation library.

[Perrrewi]

I noticed the description was wrong, it should have been computeEffectivenessMatrix that needs to be accessible for SIH.

[Jaeyoung-Lim]

I think this doesn't change my comment on the PR. Specific implementations of effectiveness matrix should not be part of the library. Only the base class should be part of the common lib. If SIH depends on specific versions of the actuatoreffectiveness implementation, it should use the VehicleActuatorEffectiveness library, which is built with control_allocator

[Perrrewi]

Thanks I made some changes.

Only moved the most generic code to the library and let the vehicle specific stay in the VehicleActuatorEffectiveness module library.

[Jaeyoung-Lim]

I still don't understand why these changes are needed for enabling SIH. For me it seems as you are just moving the location of the implementation without any added value. The actuatoreffectiveness classes you have moved are still specific for the control_allocator, and I would prefer if they don't propagate to the library.

[Perrrewi]

I noticed that there are other libraries inside modules, but in general, shouldn't we place shared content in a non-module library?

[Jaeyoung-Lim]

The reason I did not copy over the entire library was because the actuator effectiveness was not just actuator effectiveness semantically, but more of an implementation for the control allocator. Therefore I have only copied over the base class so that the control_allocation library is reusable without depending on how control_allocator works.

The main intention for this is that we want to be able to write our own control allocators e.g. metric control allocation.

If youtr intention is to fix the pseudo inverse test, we need to rewrite the test to not use to actuatoreffectivenessrotors.

If your intention is to use airframe specific control allocators in SIH, you should just use the library that contains that implementation, not move the module specific implementation to a generic library.

[Perrrewi]

Alright, I will close this PR and proposed a fix for fixing the unit test for now. Thanks for the review!