I’m facing a perplexing issue with MuJoCo’s solref parameters in a robotic arm force-control scenario, and I hope to clarify the contradiction I’ve observed to get targeted help.
Background
I’m simulating a scenario where:
What I Did
I tested two scenarios for setting the wall’s MuJoCo "solref" parameters—with drastically different results:
If I set the wall’s "solref" to my target values before starting the simulation, the behavior matches my expectation:
The arm’s end effector loses stable contact once the simulation runs. Oscillations occur (as intended), confirming the solref values themselves are effective at disrupting constant-force stability.
However, if I first let the simulation run normally (establishing stable constant-force contact for ~5 seconds), then use sim.setFloatArrayProperty() to adjust the wall’s "solref" to the exact same values as the static setup:
The simulation shows no change whatsoever.
The end effector continues to maintain stable constant-force contact, with no oscillations or force deviations—even though the parameter values are identical to the static scenario.
What I’ve Verified So Far
System Info
CoppeliaSim Version: 4.10
Operating System: Windows 11
Physics Engine: MuJoCo
I can’t figure out why the "solref" adjustment isn’t affecting the simulation. Has anyone encountered a similar issue, or does anyone know potential reasons?
Any suggestions or troubleshooting tips would be greatly appreciated. Thank you!