0
\$\begingroup\$

I'd like to build something like rigid origami in 3D space. So I need some mesh where each vertex is a particle and the set of the edges makes a chains/fabric where each edge is a "fixed length" spring and can drag the linked vertices.

I have no big experience with such a physical simulation, I have just made some experiments with verlet physics and the typical "cloth simulation". This is nice but each edge is deformed, streched or shrunken. And this is not good when you talk about rigid origami simulation.

Is there a way to make the verlet physics spring as a fixed length edges? Or maybe some different simulation engine I don't know? Inverse kinematic? Something else? I've seen in this question that Box2D can handle fixed length joints but it is in 2D only.

I'm not asking for a language-specific problem but just for information I'm using Java, OpenGL (JOGL), and toxiclibs.

\$\endgroup\$
1
\$\begingroup\$

Creating a such origami simulation is very complicated. In my opinion, to manage to create a such simulation, strong knowledge of geometry and math in general is required.

I would advise you to try to approach to this problem from the side of geometry rather than from physics side, because it is totally geometric problem. Adding physics to it (at least at first stages) would cause only problems.

I understand, that there is a temptation to just create a physically correct rectangle of paper like body, and physics engine will handle everything somehow. However this way would require a complex physics simulation. I do not know any game physics engines, that could do that. It would require a dynamic mesh generation, complex self collision handling, simulation of thing deformable plate, etc.

As for your original question about fixed length edges. If edges would be fixed length and rigid, the problem became more about kinematic, in other words - geometric problem.

\$\endgroup\$
  • \$\begingroup\$ hi thank you very much. I'm still investigating and geometry/kinematics is the way to go, you're right. what do you think about using instead of soft bodies engine (like the verlet physics) a rigid body dynamic? is this kind of engine more about kinematics? (I'm referring to jbullet for example) \$\endgroup\$ – nkint Oct 16 '14 at 12:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.