Clarification on the function `dquat2tform.m`

[DanielePucci]

Hi @Ganimed @gabrielenava, it seems to me that the function

https://github.com/robotology/mex-wholebodymodel/blob/master/mex-wholebodymodel/matlab/utilities/%2BWBM/%2Butilities/dquat2tform.m

computes a sort of homogeneous transformation where the associated rotation matrix is computed not from a quaternion, but from a quaternion derivative:

https://github.com/robotology/mex-wholebodymodel/blob/master/mex-wholebodymodel/matlab/utilities/%2BWBM/%2Butilities/dquat2tform.m#L4

The meaning of this operation is not clear to me. Can you help me understand this?

Thanks.

[Ganimed]

Hi Daniele,

thanks for the advice! The function dquat2tform is a function which does not make any sense and is wrong. This function is a relict from the very beginning and I have overseen it. You can ignore it. In the next update it will be removed from the repository.

Only the opposite direction would make a sense (orientation to quaternion rate), like rotm2dquat or tform2dquat, if the angle velocity w is known.

But a function to "predict" the orientation is maybe useful and can be calculated like in the following link: http://math.stackexchange.com/questions/189185/quaternion-differentiation

[DanielePucci]

Hi @Ganimed, thanks for the clarification. For the time being, I would avoid to put some effort in defining functions for orientation prediction.

[gabrielenava]

Function removed with https://github.com/robotology/mex-wholebodymodel/commit/8da30955676ddeb4037e997df296128bf7cd024e. Closing this issue