GPS with Position Control

[Cristian Beltran]

Dear Chelsea and Sergey,

Thanks for publishing this project. 

I am a graduate student at Osaka University and I have been trying to implement GPS on the robot NextAge (7-DoF arm including waist) using position control. I assume that the action from the policy is the next joint angles of the robot arm. I have been testing this by learning a trajectory with an end effector target pose. However, it does not seem to work, at the last policy the robot is still unable to reach the final end effector pose. 

The algorithm used is BADMM.

So my question, is it possible to learn a policy using this BADMM algorithm and using position control? or is it requires to use torque control?

Best Regards,

Cristian Beltran 

[Chelsea Finn]

We haven't experimented extensively with position control.

I would generally recommend using MDGPS rather than BADMM, as it is easier to tune.

You might consider taking a look at this pull request, which uses position control on a simulated UR robot: https://github.com/cbfinn/gps/pull/49

--
You received this message because you are subscribed to the Google Groups "gps-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gps-help+unsubscribe@googlegroups.com.
To post to this group, send email to gps-...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gps-help/27a170d6-b335-4b77-917f-3d5c5e6d1783%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.