Developing a sensor atom without polling

All topics on the Atoms in any Enterprise Dynamics Library.
Post Reply
ptis
Posts: 28
Joined: Wednesday 16 May, 2012 - 10:18

Developing a sensor atom without polling

Post by ptis »

Hello,

for a Hardware-in-the-Loop research project using Enterprise Dynamics (@ TU Dresden, Institut für Angwandte Informatik) I have to implement something like a sensor. So I tried polling to see if any other atom is within the dimensions of the sensor atom. But... well... thats not a really good way how to do it.

Are there any ideas how to solve that problem without polling and using dimensions? E.g. the Advanced Accumulating Conveyor has these sensor things which are really cool but limited to these conveyors (which also can't run backwards as I mentioned in an other thread).

Thanks
MarvinH
Posts: 93
Joined: Tuesday 25 January, 2011 - 11:07
Contact:

Re: Developing a sensor atom without polling

Post by MarvinH »

Hello!

Indeed polling should be avoided as much as possible.

Can you tell us something more about your model? Where are the "sensors" located? How do atoms move through your model? A description of your model could clarify this, posting the model itself (or the relevant part of it) would be even better.

In general, I think you should try to determine the events where the atoms "pass" the sensors and increase (and decrease) a value (global variable, cell value, attribute, etc) on these events. Then you could refer to this value to determine whether there is (at least) one atom that has passed the first "sensor" and not yet the second.

Maybe the idea I described above already gives you some ideas, otherwise please return to this topic.

Kind regards,

Marvin
ptis
Posts: 28
Joined: Wednesday 16 May, 2012 - 10:18

Re: Developing a sensor atom without polling

Post by ptis »

At the moment we are developing two sensor atoms to avoid the polling problem.

One is a polling based sensor which checks the positions und dimensions of all connected atoms. So we can simulate all kinds of unexpected behavior of the system, e.g. safety sensors around a robot cell.

The second one is a conveyor-related sensor. It's connected to a conveyor with some products in and checks the speed and dimensions to calculate the trigger time of the sensor. But that needs a recalculation on every change of the environment (e.g. new elements, other speed, "accidents") and isn't really reuseable (e.g. an end position sensor of a machine).

I hoped that some kind of visual-based events can be triggered. e.g. when the drawings (2D or 3D) intersect each other.
Post Reply