How to increase rmp lula performance on a custom 6d robot

Hi all,
i’m working on the creation of gym environment where a custom 6d robot arm is imported following the isaac sim workflow ( from urdf to the rmp lula motion policy).
Actually since the robot trajectory impacts on the simulation of an gym environment where the robot is awaited to try some picking movement, i realize that the actual parameter of the rmp are not the perfect for increasing the cycle time of my robot.
Starting from the franka file, changing parameter in the target_rmp section i reached out to increase the robot speed, but the behaviour evolves completly unstable.
There are some expert that can help me in understanding the problem?
Thank you so much in advance for the help.

Hi @vincenzodepaola94

Detailed information about each RMPflow parameter is documented here: Motion Generation — isaacsim 2022.2.1 documentation

This page also includes a tuning guide for creating stable performance:

In general, RMPflow is not designed for creating time-optimal trajectories, so there will exist a tension between speed and stability. For a truly deep dive into the inner workings of RMPflow, this paper details the motivation and implementation quite fully: [1811.07049] RMPflow: A Computational Graph for Automatic Motion Policy Generation

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.