Re: tetra transformations

[Fred Proctor <frederick.proctor-at-nist.gov>]

Parallel kinematics fans,

Till Franitza wrote:

> Hallo Brian and emc list!
> I have coded your trafo in matlab and i coded another trafo in matlab
> that uses the quaternion representation directly. They BOTH work well.
> Yours is quite tricky. I have never seen that way to do it.
> I coded a third trafo that doesn`t work that good. I have them at home
> and will send them to you tomorrow.

Sounds good. The Matlab versions won't do me any good; can you dump the
code in C form?

> What was the problem with your
> trafo? As far as i have tested it, it converges quite stably. And in
> the cases it didn`t, there was a singularity. I did not test very
> much, but is it possible that you made a forward trafo in situations
> where there simply is no solution? If the algorithm detects that, what
> is supposed to happen? I think the drives should not move and the
> worker should get a warning message. Then he can go back in machine
> coordinates and go somewhere else.

I recall that the problem was with the convergence of the forward
kinematics. Brian Register wrote the forward kins so I don't know the
details. I'd like to try your version that uses the quaternion
representation directly and see if that works on the minitetra.

--Fred