vision system is normally integrated with the robot either using some sort of fieldbus or EthernetKRL. if your simulation is done in KUKA OfflinePC, you can get both of them work.
of you are using OfficeLite, fieldbuses are not supported but you can still get vision system integrated using EthernetKRL.
but... it really does not matter how your vision processes or deliver data, it is the test results that matter. and those can be simulated too - for example entered though file or hard coded or loaded from an array...
why?
if you are not really using vision system but rather SIMULATING it... well... you need to replace relevant program parts with something that SIMULATES interaction with the vision system. this is something you need to make.
example for VT3.1
INI
PTP HOME Vel= 100% DEFAULT
PTP P1 Vel=100% PDAT1 Tool[1]: Gripper BASE[3]
IF SIM_VT_OFF THEN
; do things as usual ...
VT.TRIGGER 2Dfix_relative
VT.WAITFORRESULT 2Dfix_relative
VT.SETTARGETBASE CalibrationBase=Base[1]:CalibBase TargetBase=Base[2]:PartBase
ELSE
; run simulation ...
MsgNotify("Simulating vision, Set $FLAG[1] to trigger it")
$FLAG[1]=FALSE
WAIT FOR $FLAG[1]
MsgNotify("Simulating vision, trigger sensed")
WAIT SEC 3 ; give yourself time if needed
BASE_DATA[2] = {X ..., Y ... }
ENDIF
PTP P1 Vel=100% PDAT1 Tool[1]:Gripper BASE[2]:PartBase
PTP P2 Vel=100% PDAT2 Tool[1]:Gripper BASE[2]:PartBase
PTP HOME Vel= 100% DEFAULTAnd would you believe it ... no KONI or camera or anything needed. things got SIMULATED.
so you want to simulate something but do not know how things work and what part exactly you need to simulate.
The simulation software can stream the images to a specific port.
it can? you have this part working?
or you mean it could (eventually... if someone knew what to send, when, in what format, to what port etc.). that would be simulation of a camera.
or you man just sharing folder and letting external application drop a file?
KUKA Office PC should be capturing this stream and make the KUKA KRL library think the stream is from a real camera and process further.
As i recall VisionTech is an option from KUKA and it is really only a front end. It waits for trigger, then sends it to Baumer camera connected to KONI port. Which then returns image to KRC. Then the image is processed by VisionPro running on KRC. Then result of the VisionPro is communicated to robot program through EthernetKRL. This is probably done using Cognex SDK (Vidi).
this is meant to be done in programming language like C#. same language could be used to communicate with the robot (sense trigger, return results). normally communication would be done through EthernetKRL but you can use whatever you are familiar with.
But there are of course other ways. One approach may be to check if visionpro supports command line parameters. That makes front end easy. The only thing left is to send value to KRC.
Another way is for more stubborn applications. This works even if the target app does not expose API or have SDK. It means using windows to interact with target app. But depending on target app user interface, this can be quite clunky. like a wizard.
