Add a more realistic model for the aerodynamic force

[1-Bart-1]

The winch force model is quite unrealistic, not taking into account the kite inertia or tether slack. Some test cases with different amounts of tether slack and inertia should be added, to make sure that the controllers using the inaccurate force model still work.

[ufechner7]

Well, for the tether there is a separate issue (#10), lets not discuss this here.

For the aerodynamic force at the kite and its dependency on the reel-out speed we need indeed a better model. Ideas:

1. Use a recorded flight as input; if we want to have a short time series for the sake of performance, we could also use the recording of just one figure of eight
2. Derive an analytic solution how changes of the reel-out speed effect the kite speed, taking the mass of the kite into account.
3. Combine 1. and 2. such that we use a real (or simulated) flight as input data and correct it with the reel-out speed produced by our controller to get a reel-out speed corrected tether force
4. Add turbulence to the atmospheric model
5. Add gusts to the atmospheric model

Then we could compare the performance of improved winch controllers with the old, PID controller based implementation in a more realistic way.

[1-Bart-1]

I think we could use 1. to test and tune the controller, while we should derive simplified analytical solutions for feed-forward or model-predictive control in the controllers.
4. and 5. would benefit both WinchControllers.jl and KiteModels.jl.