Ask Your Question

What is the proper way to obtain the fixed-axis covariance matrix from quaternions?

asked 2014-03-11 01:45:17 -0600

updated 2014-05-04 13:12:45 -0600

REP 103 defines the standard for rotation representation as quaternions, but the covariance matrix associated to it is given in terms of fixed axis rotations. In my work, I use quaternions directly for attitude estimation and obtain covariances in terms of the quaternion parameters.

How can I convert these covariances to the convention used by ROS, so they can be published appropriately?

PS: I have searched ROS Answers and found a few discussions briefly addressing this topic, but usually referring to external documents or additional packages. I think it would be useful to have a single definitive reference for this here, so that implementations don't depend on the reader's interpretation of what is expected.

edit retag flag offensive close merge delete


@tfoote@William, any chance you guys could help me out with this one?

georgebrindeiro gravatar image georgebrindeiro  ( 2014-04-27 07:13:59 -0600 )edit

1 Answer

Sort by ยป oldest newest most voted

answered 2014-05-04 15:29:42 -0600

Chad Rockey gravatar image

This answer links to a thesis that discusses how to convert the covariances.

edit flag offensive delete link more


thanks, Chad. your answer was actually the second link in my post. I just wanted to make sure that this method corresponds with all the ROS conventions. as far as I'm concerned, this is the kind of info that should be part of the standards. I would go as far as including conversion functions :)

georgebrindeiro gravatar image georgebrindeiro  ( 2014-05-22 11:26:31 -0600 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools



Asked: 2014-03-11 01:45:17 -0600

Seen: 670 times

Last updated: May 04 '14