Open hubernikus opened 4 months ago
Sorry, I missed this PR. Can you provide a bit more context on this change, i.e. what is the code that is failing without this change? Based on my experience with rpy
and similar euler transform, probably it is easier to just write a test snippet that check the consistency of a given existing method, rather then try to just check the convention by hand.
I propose to use extrinsic angle decoding of fixed transforms with
rpy
. This is according to the convention defined by ROS for rpy: https://wiki.ros.org/geometry2/RotationMethodsAnd also how other libraries such as MoveIt uses them: https://github.com/moveit/moveit_calibration/pull/47/commits/b7148d75c1a14ebf50335a3aaa051b140e6fdc0e
Additionally, I use the
Rotation
class form scipy, as that make the code much simpler. However, this implied that the result option is not possible for the back-conversion.