google-deepmind / google-deepmind/mujoco
friction between cable and geometry (eg capstan)
- Dominant language
- C++
- Stars
- 15.2k
- Forks
- 1.8k
- Avg merge
- 10d 16h
- Merged PRs (30d)
- 25
Description
### The feature, motivation and pitch
Hi MuJoCo team,
I’d like to request support for capstan friction in spatial tendons, which is important for cable-driven robotic hands. A simple model (T_high = T_low · exp(mu · theta)) would already be very useful.
Right now it’s hard to implement this in user code because spatial tendon wrapping does not expose wrap geometry (e.g., theta or entry/exit tangent points), and the tendon model effectively exposes only a single scalar tendon force—so users end up splitting tendons and manually approximating theta.
Could you either (1) add an optional built-in capstan friction model for cylinder/sphere wrapping, or (2) expose wrap geometry outputs (theta + entry/exit points) via API/sensors to enable robust plugins?
Thanks for considering,
### Alternatives
_No response_
### Additional context
_No response_
Contributor guide
Assessment
This issue has not been assessed yet.