Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Could obi implement this?
#1
[/url]https://www.youtube.com/watch?v=DYWq_CjsSVw
[url=https://www.youtube.com/watch?v=DYWq_CjsSVw]
I wonder whether obi rod could implement this in the above vedio. Jump straight to 58 seconds.

There are some key features which I think should be pointed out first:
1.  The black line in the vedio is made of metal in reality. It is highly resistant to stretching and twisting but easy to be bended. Most important of all, it is very stable for there is almost no jittering while moving(I have tried the obi rod but it jitters a lot while moving).

2.The blood vessel which around the black line could collide with the black line. The vessel animates periodically so it can be seen as a rigidbody rather than a softbody because the animation is known prior. 

So does obi rod could implement this? I have implemented use finite methods,but the collision force is hard to compute.
Reply
#2
(25-12-2019, 03:41 AM)yyxymlove Wrote: [/url]https://www.youtube.com/watch?v=DYWq_CjsSVw
[url=https://www.youtube.com/watch?v=DYWq_CjsSVw]
I wonder whether obi rod could implement this in the above vedio. Jump straight to 58 seconds.

There are some key features which I think should be pointed out first:
1.  The black line in the vedio is made of metal in reality. It is highly resistant to stretching and twisting but easy to be bended. Most important of all, it is very stable for there is almost no jittering while moving(I have tried the obi rod but it jitters a lot while moving).

2.The blood vessel which around the black line could collide with the black line. The vessel animates periodically so it can be seen as a rigidbody rather than a softbody because the animation is known prior. 

So does obi rod could implement this? I have implemented use finite methods,but the collision force is hard to compute.

Nope, can't even get close. There's many reasons why it cannot:

- Obi is a particle-based engine. As such, the only collision primitive it can use are particles. A vessel is a hollow cylinder, and although you could approximate its shape using lots of particles, this is both expensive and inaccurate. Vessels could be best modeled as a custom constraint on the distance from rope particles to a curve, but this is an extremely specific use case so it hasn't been implemented.

- Obi uses a discrete representation of materials. Mass is lumped in particles, so behavior changes depending on the topology of the object. FEM is much better suited to this problem because behavior won't change significantly depending on guide wire/vessel topology.

- Obi is an iterative engine. Because of this, the simulation is not guaranteed to converge unless lots of iterations or a very short timestep are used. The error will show up as spurious compliance (which is actually a good thing, since a traditional velocity-based solver would just blow up). In practice this means that material compliance depends on your time budget, so it's hard to get very resistant materials with a low budget. This is usually not a problem since cloth, ropes, fluid and most softbodies deform quite easily, but other materials such as metal do not.
Reply
#3
(26-12-2019, 10:58 AM)josemendez Wrote: Nope, can't even get close. There's many reasons why it cannot:

- Obi is a particle-based engine. As such, the only collision primitive it can use are particles. A vessel is a hollow cylinder, and although you could approximate its shape using lots of particles, this is both expensive and inaccurate. Vessels could be best modeled as a custom constraint on the distance from rope particles to a curve, but this is an extremely specific use case so it hasn't been implemented.

- Obi uses a discrete representation of materials. Mass is lumped in particles, so behavior changes depending on the topology of the object. FEM is much better suited to this problem because behavior won't change significantly depending on guide wire/vessel topology.

- Obi is an iterative engine. Because of this, the simulation is not guaranteed to converge unless lots of iterations or a very short timestep are used. The error will show up as spurious compliance (which is actually a good thing, since a traditional velocity-based solver would just blow up). In practice this means that material compliance depends on your time budget, so it's hard to get very resistant materials with a low budget. This is usually not a problem since cloth, ropes, fluid and most softbodies deform quite easily, but other materials such as metal do not.

Thanks for the detail ! What a pity i can't use the obi to do it! By the way, could you know if there is a fem plugin for unity? Or any other suggestions to implement this other than FEM?
Reply