Help using SOFA-GYM

[TNZK-0103]

Dear SOFA Community,

I am currently conducting research using SOFA, and I would like to ask for your advice regarding the use of SOFA-GYM for motion planning.

My research focuses on spoon manipulation using a soft actuators attached to a robotic hand, and in particular on exploring how the fingers should be moved to achieve stable and effective manipulation.
As one possible approach, I am considering the use of SOFA-GYM to design motion strategies through reinforcement learning.

More specifically, I am envisioning an approach in which a soft robotic hand model, together with contact and friction interactions in SOFA, is formulated as a Gym environment, and manipulation stability and task success are used as evaluation criteria for learning.

I would greatly appreciate your opinion on whether SOFA-GYM would be a suitable framework for this type of problem.
If there are any related examples, best practices, or technical considerations that I should be aware of, I would be very grateful for your guidance.

Thank you very much for your time and consideration.
I look forward to your response.

Kind regards,

[hugtalbot]

Dear @TNZK-0103

The SofaGym project is currently in standby but it will be resumed in 2026. New engineers will start working on RL with SOFA.

So far, the SofaGym project is compatible with SOFA v23.06 release, which now starts being old.
We will announce the restart of this work, since we are well aware how much it is expected from the community 👍

Thank you for your feedback.
Note for @alxbilger

[TNZK-0103]

Thank you for your reply and I am looking forward to using SOFA-GYM in 2026.

[hugtalbot]

Thank you for your understanding @TNZK-0103