Closed PeterBowman closed 2 years ago
I wonder if the benefits of enforcing such a constraint are worth the effort to develop the full backend. Low accelerations/torques/currents values may still produce significant velocities.
Edit: see https://github.com/roboticslab-uc3m/kinematics-dynamics/issues/161.
I think the I2t driver protection is the way to go, also a firmware protection is far more reliable than any upstream slow-loop check.
Some CMC-controlled commands (i.e.
movl
andmovv
) make sure that the resulting velocity (obtained via solver methods) does not exceed certain limits. For instance:https://github.com/roboticslab-uc3m/kinematics-dynamics/blob/c187abdc0c9b2284d36df0f22cf64ecb238642b2/libraries/YarpPlugins/BasicCartesianControl/RateThreadImpl.cpp#L187-L195
Is it feasible to limit the resulting torque in the
gcmp
andforc
commands? How to obtain that in the first place? See https://github.com/roboticslab-uc3m/yarp-devices/issues/187#issuecomment-406342280.