Simple simulation failed in ODE. Should I use Bullet instead

[princeward]

Hi all,

Greetings! My simulation is very easy: pushing a box in a plane (on the ground). Forces are applied in different positions on the box. So the box is supposed to translate and rotate on the plane. It's a research project, rather than a game.

However, the poor friction model in ODE drove me crazy. There are so many bugs there and I am unable to fix them. Just one quick example to show you how bad ODE is: suppose the friction is 5N. If a apply a force (6,3), which is 6N in X-axis and 3N in Y-axis, then the object will only move in X-axis due to the bad friction approximation model in ODE. There were many more problems followed when I tried to fix this. I gave up.

So I'm here to ask you guys, is Bullet capable of simulating this task? All I need is the accurate friction force as well as the torque generated by the friction when the box is rotating. If yes, I believe I will turn to use Bullet immediately.

Any reply is appreciated! Thank you!

[bwelch]

I *think* you should be able to do that pretty simply. Bullet can definitely apply the force the way you want, but the way it deals with friction is a little confusing to me. From what I know, it uses a scale from 0 to 1, where 0 is frictionless and 1 is complete friction. So figuring out how to get your 5N friction force could be kind of a pain. You could modify one of the demos pretty quickly to test it out, though.

[princeward]

Well, that's really a pain. In ODE you can directly assign the friction coefficient.

But if it's accurate then that's not a big problem. I'll try it out!