UR5e Force Mode Control

Hi,
I am currently working on some advanced impedance control technique and we are using the UR5e robot in force_mode. I would like to know what model does the force_mode uses to calculate the acceleration for the end-effector when some desired_wrench is given? Is it a unit mass model in the Cartesian directions? What are the relations with the motor torque?

Hi @udayanbanerjee15,

Force_mode is currently a closed source implementation that is trying to impact with the target force in the compliant axes. You can adjust gain and dampening to trim it for your need.

At URScript_Examples/admittance-control at main · UniversalRobots/URScript_Examples · GitHub there is an open source implementation that is based on the FT sensor and speedl. That will also make the robot speed scale when you are close to a singularity.

Ebbe

Hi Ebbe,

Thanks for your reply. As you can see in the below link that we have been able to achieve a decent impedance control in force_mode(). This works well for Cartesian position tracking as well (provided its a Cartesian point).

https://csciitd-my.sharepoint.com/:v:/g/personal/srz198482_iitd_ac_in/Efap8H8JFsxDncI8OSDjh_sBSnd5Rpimy0MkXtfIvx2Ehw?nav=eyJyZWZlcnJhbEluZm8iOnsicmVmZXJyYWxBcHAiOiJTdHJlYW1XZWJBcHAiLCJyZWZlcnJhbFZpZXciOiJTaGFyZURpYWxvZy1MaW5rIiwicmVmZXJyYWxBcHBQbGF0Zm9ybSI6IldlYiIsInJlZmVycmFsTW9kZSI6InZpZXcifX0%3D&e=CVWZhH

But our work requires tracking of some desired trajectory in the force_mode() . But unfortunately the mode shows up a bounded error in that case. So if you can share the model through which the accelerations are being calculated, it would be really helpful.

Regards,
Udayan

Hi Udayan,

It looks awesome!

I can forward your request. But I doubt the amount that will be shared.

On the sensing side there will always a little error. That is way there is implemented a “deadband” in the repository. I think you can also implement that on top of force_mode() and maybe also zero the FT sensor if there seems to be no activity for some time.

Ebbe