Simulation slider for PTP move - strange mapping

Started by frenezulo, July 26, 2014, 05:12:03 PM

Previous topic - Next topic

frenezulo

Hi Johannes (and others),
I have a question regarding the simulation slider and PTP move. I have several positions and when I want to simulate the movement between the positions (by script), I don't know what values should I prescribe for each position.
For example, I have 3 positions (and START and END position), so I made one slider with the extent 0-4 and remapped it to 0-1 extent. The positions change in these values: 0.41, 2.01, 2.81, 3.61. It doesn't make sence for me! Is it somehow possible to change positions in linear manner (by slider or other component)?

Best regards,
Tomas

Johannes @ Robots in Architecture

Hello Tomas,

You're right, there are some special things to consider regarding the structure of the simulation slider.
If you've got three positions, you actually end up with actually six robot positions...
1 E6AXIS as defined in the menu as the start position
1 PTP position that is a duplicate of the first Cartesian position
3 robot positions that you created using KUKA|prc
1 E6AXIS as defined in the menu as the end position

This is done because the movement from start position to first position can be quite long and as such using LINear movements can be problematic (singularities etc.).
And there is one more thing to consider: While the values are evenly spaced along the slider, they visualization "switches" when it moves from one are to the next. See the graphic below:



I hope the slider values now make sense, I have to admit that I never quite noticed the non-normalized behaviour as I commonly work with more positions where the deviation isn't really noticeable any more. A way to circumvent this is just to use the List Item component and only stream one individual position to the KUKA|prc core component.

Best,
Johannes

frenezulo

Hello Johannes,
thanks for the explanation, it makes more sence to me now.

In the meantime, it came to my mind, that posting one position from some list at time would be the easiest way to overcome the described problem. So this is the same solution as you suggested :)

Thanks,
Tomas