Hi All,
I've been trying to understand how to get a quadrotor helicopter under control against wind, keeping it in the planned path and speed, despite the wind.
I have got a set of equations (six of them to be precise -x,y,z,theta, phi and psi), all of them showing accelerations in the x,y,z and angular accelerations for pitch, roll and yaw.
I built a simulink model and tried to run simulations to predict how the quad should behave but it seems to be pitching violently.
I need to get the model working correctly and then find the inverse, which I intend to use to train the neural controller.
Please can someone give me any help on
- a set of quadrotor equations that will give me the best results
-ideas on how to build a good model, one that I can easily find the inverse.
-any other advice on how to achieve neural control of the quad.
Thanks.
Tim.
Replies
You shouldn't be inverting the matrix:
http://www.johndcook.com/blog/2010/01/19/dont-invert-that-matrix/
If you post your current simulink model, maybe we can help (just a snapshot). Try giving less authority to your learning algorithm.