ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question
0

Is it possible to simulate cntx% of fanuc with ros ?

asked 2022-05-03 04:11:35 -0500

Lemon gravatar image

As you may know, fanuc does not provide any insights on the inner work of cnt. I need to simulate exact fanuc robots with their different movement types (specially joint) and with their different options (fine, cntx%). Is it possible with ros ? Did someone tried to reverse engineer the cntx% algorithm of fanuc ? Thanks.

edit retag flag offensive close merge delete

1 Answer

Sort by » oldest newest most voted
1

answered 2022-05-03 11:06:22 -0500

gvdhoorn gravatar image

updated 2022-05-03 11:09:28 -0500

No, not to my knowledge.

All ROS integrations I've seen (or worked on) simply use whatever Fanuc provides.

They do not circumvent or replace.

So they are as dependent and 'encumbered' by the CNT-system as regular TP programs.

Additionally, there's nothing specific to Fanuc "in ROS".

I don't believe you can ask "Is it possible to simulate [..] of Fanuc with ROS", as the available simulators are all generic tools.

edit flag offensive delete link more

Comments

Following that, how is it possible to plan collision-free trajectory without simulating cnt ? Maybe cnt of fanuc is similar to all robots ? It's probably just a bézier curve.

Lemon gravatar image Lemon  ( 2022-05-04 01:47:08 -0500 )edit

how is it possible to plan collision-free trajectory without simulating cnt ?

By assuming the robot will do a reasonable job of following the trajectory faithfully, and by including padding in the collision checks.

But if you're asking: "how can we know the robot will execute a given trajectory perfectly?", then the answer would be: we can't.

That's the biggest problem with using TP to command motions on Fanucs. It's an unavoidable trade-off between precision and motion smoothness.

But there are other ways to control Fanuc robots which don't suffer from this. J519 being one.

gvdhoorn gravatar image gvdhoorn  ( 2022-05-04 02:41:59 -0500 )edit

The stream motion option. I will check it out. Thanks for the answers.

Lemon gravatar image Lemon  ( 2022-05-04 03:34:32 -0500 )edit

Fanuc support said it's replaced by RMI and RMI uses TP .. weird

Lemon gravatar image Lemon  ( 2022-05-04 03:49:39 -0500 )edit

RMI does not replace J519.

Fanuc just doesn't want to sell you J519.

They then try and tell customers asking for J519 that RMI is "the replacement".

J519 is like ABB EGM and KUKA RSI.

RMI is a drip-feed for TP programs (with all the limitations (and then some) that TP programs have).

J519 is a 125 Hz joint and Cartesian position control, hard real-time external interface over UDP.

RMI is a JSON-over-TCP/IP RPC protocol, which happens to also let you start TP-backed motions.

Those are most definitely not the same things.

gvdhoorn gravatar image gvdhoorn  ( 2022-05-04 04:35:37 -0500 )edit

Could you please mark the question as answered btw if you feel it has been answered?

Your original question was about "using ROS to simulate Fanuc CNT behaviour", which I believe we addressed.

Your follow-up comments are technically different questions.

gvdhoorn gravatar image gvdhoorn  ( 2022-05-04 04:36:59 -0500 )edit

oh yeah sorry i'm new here. It's weird that fanuc don't want to sell us the J519.

Lemon gravatar image Lemon  ( 2022-05-04 04:44:08 -0500 )edit

It's weird that fanuc don't want to sell us the J519.

That would be something you'd have to take up with them.

You're (most likely) in Europe, and Fanuc EU is very difficult about it.

If you'd have been in the US, there would be much less of a problem.

If this is a research project, you might be able to get it as a trial, but I would be surprised.

gvdhoorn gravatar image gvdhoorn  ( 2022-05-04 04:47:00 -0500 )edit

Question Tools

Stats

Asked: 2022-05-03 04:11:35 -0500

Seen: 115 times

Last updated: May 03 '22