Switch UR10, UR10e and UR16e to effort controllers

[ipa-kut]

This fix addresses #521 which in turn addresses #386, and was worked on during WRID2020

Based on a rough understanding of this comment, only the three URs mentioned in the title have been updated.

PID values for the effort controllers were taken from here

After making these changes, the gazebo simulation of the robot was observed to start in a "weird" pose :
UR16e:

UR10 and UR10e:

However, the controllers initially seem to initially work well, and the robots could be jogged using the RQT Joint trajectory controller plugin

[fmauch]

As this PR is branching of melodic-devel-staging none of the models use the updated cog from #504.

Eventually, this should be changed for all 7 robot variants, though.

[ipa-kut]

I will change this for the remaining URs now.

As this PR is branching of melodic-devel-staging none of the models use the updated cog from #504.

Would it be better to wait until #504 is merged, to rebase this and test again?

[fmauch]

I will change this for the remaining URs now.

As this PR is branching of melodic-devel-staging none of the models use the updated cog from #504.

Would it be better to wait until #504 is merged, to rebase this and test again?

I think, this can be implemented independently from #504 but for fine tuning it might be interesting whether it makes a difference whether we use the cylinder approximation for the inertia or we use the matrices provided by the robot controller. So, testing and fine-tuning this might require a rebase on #504. However, as the results from those tests will also impact #504, we should not wait for that to be merged, but develop them in parallel. I think, this won't be too much of an issue as there are no conflicts between those two.

[ipa-nhg]

Tested with #526 and it works.

Great contribution @ipa-kut, thanks! 👍

[RobertWilbrandt]

The changes from this should now be merged through #619. Thanks @ipa-kut for getting this started!