Porting of DCM for use in Quadcopter

[Leo Barnes]

Hello!

I have been reading up on the firmware in the UDB and find the use of
the DCM to be especially interesting. I have just recently started
sketching on a quadcopter, but since I find the process more fun than
the end-result and work as a programmer (specializing in image
analysis/compression), I would like to design my own flight control
software. My first step is to build an IMU from a 3-axis gyro and a 3-
axis accelerometer (I know that I won't be able to eliminate yaw drift
without some kind of additional sensor, but this is a later problem)
which seems to be very similar to what you have done in your DCMDraft2
document. I would like to take the specific dynamics and kinematics of
a quadcopter into account however.

Do you know if there is anyone else working on this? Always good with
additional insights of how other people have done it, especially since
it was some time since I took classes in control theory/dynamics/
kinematics.

Best regards,
//Leo

[William Premerlani]

Hi Leo,

 

If you are planning on taking the full dynamics and kinematics of the quadcopter into account, you definitely want to read Robert Mahoney's papers on the subject, for starters. He treats the full rotational dynamics of certain types of aircraft.

 

So far, we (UDB team) have been exploring how far we can push the kinematic representation, and simple linear feedback control, for airplanes, without any knowledge of the specifics of the dynamics, other than some general assumptions.

 

My interest has been mainly fixed wing aircraft, so I have not been following what others have been doing with quadcopters. There have been several postings on diydrones that have used either the DCM algorithm or Kalman filtering, but I do not know how far into the dynamics they have gotten.

 

There is also a helicopter group that you might want to join. They have made some progress in applying the DCM algorithm to heli stabilization. Progress has been hampered by the discovery that most gyros are vulnerable to vibration. We have recently discovered that Invensense gyros are likely to be suitable for heli and quadcopter application, and have been testing out some prototype boards.

 

One of the things I my list of things to do is to apply some of Mahoney's control concepts to fixed wing. So far, our work has been focused on applying the nonlinear DCM algorithm to attitude estimation, but using only simple linear feedback for control. Mahoney has some ideas for taking fuller advantage of the rotational properties of DCM in the controls, I intend to look into that.

 

Certainly, I would be interested in hearing about your work.

 

Best regards,

Bill

 

--
To unsubscribe, reply using "remove me" as the subject.