On 3/26/14 11:48 , David Raila wrote:
>
>       I have a machine with 6 axes and non-trivial kinematics, and
> currently I am only currently using inverse kinematics.
> The constraint is such that some joints must be moved exclusively of
> others for the machine to be deterministic,
> i.e. if joints that need to be exclusive are moved simultaneously, the
> position of the machine becomes unstable.
> For example the machine can move X,Y,Z simultaneously, but motion of any
> of A,B,C must be exclusive, i.e. only
> one of A,B,C can be in motion at a time, and during that motion X,Y,Z
> must be fixed.

Hi David, that sounds like a neat machine!  Got any pictures or videos?


>       One way to accomplish this is to only give the machine G-codes that
> conform to the constraints by splitting
> XYZABC motions into multiple serialized motions - say a sequence of XYZ
> then A, then B, then C, and this will
> work for the applications of this particular machine, but I'd prefer a
> more generalized solution where it is possible
> to have 'smooth', or nearly smooth motion in 6 axes.
>
>       I am wondering if this can be accomplished by building the
> constraints into the kinematics code, which would
> do something along the lines of a 'round robin' of the conflicting joint
> positions, where the returned joint positions
> would only contain non-conflicting positions in a round-robin fashion.
> If the machine cycle is fast enough, the motion
> should be nearly smooth while avoiding conflicting joint moves that
> would be unstable.

That seems like it could work.

As you know, when the trajectory planner sees a move like "G0 X10 A45", 
it's going to emit a sequence of waypoints that smoothly move both X and 
A simultaneously to their programmed endpoints.

Maybe you could increase your ferror limits and then teach your kins to 
round-robin the axes, for example by moving just the axis that's the 
farthest from its current target, on each cycle of the motion controller.

I'm not sure how that would interact with machine accel limits etc, but 
i dont see any fundamental reason why it wouldnt work.


-- 
Sebastian Kuzminsky

------------------------------------------------------------------------------
Learn Graph Databases - Download FREE O'Reilly Book
"Graph Databases" is the definitive new guide to graph databases and their
applications. Written by three acclaimed leaders in the field,
this first edition is now available. Download your free book today!
http://p.sf.net/sfu/13534_NeoTech
_______________________________________________
Emc-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-developers

Reply via email to