How to transform the grasp coordinates to the camera coordinate #7
Description
I encountered an issue during deployment. My current approach is as follows: I first transform the rotation provided by the model into the camera coordinate system, and then I map both the translation and rotation into the robot base coordinate system.
When I try to convert both translation and rotation from the model's output to the robot base coordinate system, the translation works fine, but the rotation always seems incorrect. I attempted to right-multiply the model's rotation by the transformation matrix [0,0,1], [0,1,0], [-1,0,0] to align it with the camera coordinate system. However, it failed.