-
Notifications
You must be signed in to change notification settings - Fork 218
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Gazebo support for KUKA KR10 #125
base: melodic-devel
Are you sure you want to change the base?
Gazebo support for KUKA KR10 #125
Conversation
Reverted changes to adhere to indigo convention, changed to default KUKA colors, and added dummy inertial values |
I realise that getting anything but approximates is difficult/impossible, but the current dummy values give all links a mass of 10 Kg. That seems a bit much. There are a few ways to estimate masses and inertias from meshes, such as using Meshlab or using SolidWorks. Another alternative could be to use approximations based on primitive shapes (such as is done in the We could consider adding similar macros as the one in |
Related #128. |
The mass values of the UR look like accurate values (beside the first one). I can only make estimates on the mass. Using a macro for cylindrical inertia estimates seems like a good idea. I will try to implement it in the near future. |
So i added a common_robotparameters.xacro in kuka_resources where we could store all "identified" robot parameters such as mass, link lengths in DH notation, cog etc What do you think? |
Hi, thanks for the PR. Adding the |
I'm also not sure about using DH for this. Afaik, KUKA does not use DH, so it seems strange to start doing that in this repository. |
My ideas were:
I will the put the parameters to the respective macro file and use the notation from the KUKA manual instead (i will do this in 2-3 weeks). |
@cschindlbeck wrote:
Ok, that could make sense. This would connect to the discussion @BrettHemes and I had in #105. re: reduce errors: we're still using text-based identifiers with no robust way of verifying that we've used the correct one other than looking at the model in RViz or manual inspection of the xacro macro text. It feels a bit like 'fixing' one side but introducing another possible source of errors.
Yes, it's a common convention, but as I wrote earlier, afaik KUKA doesn't use it. That means we'd have to rely on contributors to derive the DH parameters for us, which doesn't seem like it improves things (compared to how we do it now in the xacro macros).
No problem. |
I have put all code into the kr10r1100sixx_macro.xacro. |
@cschindlbeck wrote:
I'm not sure which link names you are referring to?
And I completely agree :) see my earlier #125 (comment):
|
These:
as in here |
Ah. I'm not too hung up on particular names in this case. It might make sense to move those Jade+
It would be good to give the properties a bit more of a descriptive name though, I'm not really a fan of single-character variable names. I understand these come from DH, but since we're not using that (and consequently our frame orientations don't align with it) I'm not sure using DH names for these lengths still makes sense? |
I shortened the names and made them more descriptive. |
Experimental support for Gazebo KUKA KR10.
Currently on kinetic and i have my custom colors applied