I am working on an autonomous 4-legged robot simulation with a research team using NVIDIA Isaac Sim and I want to implement it using the Reinforcement Learning. I am trying to incorporate a cost function that involves minimizing torque and maximizing speed. However, I am concerned about the risk of joint damage due to excessive torque.
To address this, I would like to monitor the torque being applied to the robot’s joints during the simulation. I want to know how to monitor the torque, I’ve been searching for this problem a lot lately and I noticed that I could add a torque sensor, but I don’t know how to do so, or where should I put it, so I need your guidance to do that.
Thank you for your time and assistance.
Hi @meryanraad ,
You mean actuator effort? There is currently no direct way to get this. But you can estimate it quite well using
tau = stiffness * (targetPos - currentPos) + damping (targetVel - currentVel)
for a joint drive, where all variables above are scalars.
The articulation force sensors are deprecated and will be removed in a future release, so I would advise against using them.
We are also working on a joint force feature where we report the force transmitted from parent to child link as a full 6d spatial force, which will include the dof consraint forces such as drives and friction. But this is not yet released.
Philipp
1 Like
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.