SEDSAT-2 ADCS Control algorithms
From SEDSWiki
Contents |
Simple Algorithms
proportinal control, etc...
Advanced Algorithms
Adaptive Control
General & SWOT Analysis
The method chosen is the dynamically stable “Lyapunov’s Direct Method”. The biggest strength of this algorithm is it’s ability to calculate a control torque without needing to know the systems constant parameters like inertias, etc. By taking both, the difference between real and desired trajectory and, via an “adaptive law”, a term corresponding to the difference between it’s real and “guessed” constant parameters into the control law, this method is capable of “learning” the system’s parameters while following the trajectory, with increasing accuracy with time.
It’s mostly used in robotics for trajectory control. An evaluation can be found in the paper I added in "result"
The Code
...will come soon
.
Result
Here: Documentation on Lyapunov's direct Methode for Attitude Control you can find a short paper which contains the main results and evaluation. It also contains the equations of motions, careful: may still be bugs in there.
There is a 3D animation of the Satellite available (AVI ~6.6MB) for those who don't believe the graphs: http://protoforge.org/files/File.php?id=43&
Here are some results from an example, using a initial tumbling situation (q represents angles, qd angle velocities):
q = [0, 68.75, -68.75] [deg]
qd = [0, 0.57, 2.86] [deg]
after t = 550 sec, an "outside" error was simulated, by adding
qd_error = [0, 12.61, 8.59] [deg]
to the system.
Note that both latitude and longitude motion are accelerated for presentation purpose
I chose the scaling parameters in a way that no torque bigger than 8E-5 [Nm] occurred.
There are still a few bugs in the code, but it should be good enough to show that the algorithm works fine.
Desired q, relative to nadir
Real q, relative to nadir
Torque
