Model Improvement/Debugging and Force Model for Delta Quad Plane

[manumerous]

Model Adaptions

This PR includes many smaller adaptions to improve and debug the model. The most important ones are:

• Fixing of the symmetric sigmoid function which was wrong due to not setting the right scaling parameter to adjust the width of the band needed to change from 0 to 1. This is documented in #62
• Disabling the regression intercept. Since the rows of the regression matrices correspond to 3 different linear and 3 different rotational axis it makes no sense to try to fit a single intercept through them all.

Delta Quad Plane Model

A quick model for a delta quad plane model was created to test with some real data from logs.px4.io. Since i can not know the the actuator position and the logs found online do not contain angular acceleration as a topic it is just set up to estimate forces for now.

For this reason it was convenient and fast to handle it in a different file. But with some though that could be improved in the future. A description of those improvements can be found in the issues #65 and #66.

However, the separate aerodynamics model for deltas and the config are needed regardless of whether we unify all quad_plane_models.

[manumerous]

* With the two different models only differing actuator configurations the amount of boiler plate code that needs to be added for newer models are too large. Lets discuss how we can improve on this later. (not a blocker for this PR)

@Jaeyoung-Lim

Just to be sure that we do not confuse anything here:

As mentioned above the models differ in the aerodynamics module and the inputs of the aerodynamics module. The computation of the rotor features is identical (apart from of course the values in the configuration files). We can discuss this directly in the issue #65 that i linked.