Control area is released with AGV still inside

Question

question

Patrick Zweekhorst asked Mar 14, '22 Matthew Gillespie commented Mar 31, '23

Control area is released with AGV still inside

We have made an example model to show a problem we have with some AGVs. This is the same model as you helped me on before (with the finalizeSpatialChanges before redirecting the AGV).
Somehow the AGV release the control area when it starts traveling again. I can’t really figure out how this is happening, maybe you have some ideas?

The problem can be seen at time 119.05.
You might also notice the small network above the current network. That was my first example model, but in that model, I could not reproduce the issue. When I copied part of the actual model, I could reproduce the problem.

Thank you already for your time.

Patrick

ControlArea_1.fsm

Software Version:

FlexSim 22.0.1

agv control area allocation

controlarea-1.fsm (47.7 KiB)

· 3

Jason Lightfoot ♦ commented · Mar 14, 2022 at 05:45 PM

It looks like it's happening in the redirect. Can you configure the Taskexecuter2 in the attached model so that it doesn't happen in first test network? (point the 3 labels on the TE to the sequence you were using)

controlarea-1_jl.fsm

0 ·

controlarea-1-jl.fsm (48.5 KiB)

Patrick Zweekhorst Jason Lightfoot ♦ commented · Mar 15, 2022 at 08:50 AM

Hi @Jason Lightfoot ,

It indeed looks to be going wrong with the redirect, but I have no clue why. Attached is one of the examples where we tried to create to problem in the other situation. In both situations the AGV will start traveling again at time 105. In the top part of the model the control area is still claimed. In the bottom part the control area is released when the redirect is executed.

Thank you for looking into this or referring it to one of the developers.
ControlArea_2.fsm

0 ·

controlarea-2.fsm (52.9 KiB)

Patrick Zweekhorst Patrick Zweekhorst commented · Mar 21, 2022 at 01:59 PM

Hi @Jordan Johnson ,

Do you know if there is someone able to look at this problem? Or to check if it is an bug with a workaround or that I am doing something wrong in the model logic.

Thank you for your help.

0 ·

Answer 1 · 2022-03-21T18:22:24Z

Phil BoBo answered Mar 21, '22 Matthew Gillespie commented Mar 31, '23

@Patrick Zweekhorst

I stepped through what is happening and will explain below with a suggested workaround. I'll add a case to the dev list with this example model for @anthony.johnson to determine how to address this issue as a bug or not.

Here is an image demonstrating the steps of what is happening with the travel paths when you redirect the agv in these two cases.

First, the current travel path is aborted, shown in red (step 1).

As part of aborting that travel path (step 2), it finds allocated objects that will be reallocated as part of the redirect. In the top situation, that is just CP5 and not the control area. In the bottom situation, that is CP182 and the control area.

In trimming the travel path after the preemption (step 3), it deallocates the objects from step 2 that are not on the new travel path, shown in purple. In the top situation, this correctly deallocates CP5. In the bottom situation, this deallocates the control area because it is not on the new travel path. This result is not what you wanted when you did the redirect(). If you had redirected to a control point that was beyond the control area, then it would not have been deallocated. (For example, if you move CP182 beyond the left edge of the control area.)

As a workaround, you could add another control area that will be allocated as part of the new path, such as this:

To get the overall traffic control that you want in this area, you may need to use more control areas to compensate for situations where the outer control area gets deallocated in an undesirable way during a redirection that preempts the agv's travel.

1647886239714.png (227.3 KiB)

1647886637109.png (20.0 KiB)

· 2

Patrick Zweekhorst commented · Mar 24, 2022 at 11:20 AM

Hi @Phil BoBo ,

Thank you for your clear and extensive explanation. The work around sound good and I only have this at a few places so it should definitely help me out.

Also looking forward to the bug fix.

Thanks again

Patrick

0 ·

Matthew Gillespie ♦♦ commented · Mar 31, 2023 at 09:56 PM

@Patrick Zweekhorst This issue is fixed in 23.0.6 released today.

0 ·

question