Friction coefficient set in the xml does not match the simulation exactly

[raisimtech]

Friction coefficient set in the xml does not match the simulation exactly

[jhwangbo]

I could reproduce this error. I quickly looked at it but couldn't find the issue. There are a few approximations for numerical stability. I'll check which one is causing this discrepancy. This issue is known (https://leggedrobotics.github.io/SimBenchmark/rolling/index.html) but I didn't spend time to investigate it. I'll take a deeper look at it next week.

BTW, the first argument for setExternalForce is the local index. and local index for all SingleBodyObject is zero. The method will actually ignore the first argument. So you don't have to get the index to apply the force.

[weixuanzhang]

BTW, the first argument for setExternalForce is the local index. and local index for all SingleBodyObject is zero. The method will actually ignore the first argument. So you don't have to get the index to apply the force.

Thanks for the heads-up!

[jhwangbo]

This is due to the model used in RaiSim. Unfortunately, there is no perfect model for contacts. They all have pros and cons. I can point you to the literature on this matter but this will be a lot of reading. The bottom line is that this is an expected result.