JMonkey and collisions

Hi all,

to begin with, I’m sorry if this is not the right place for this discussion. I kindly ask the moderator to move the topic if necessary. :slight_smile:

I’m totally new to JM and I’ve found it quite an interesting project. Currently, I’m involved in an ongoing Java project which will provide a visual representation of 3D entities moving and colliding in a 3D space. In this project, the entities are represented as polytrees and discretely moved (time steps), V-clip collision detection algorithm is exploited with a binary search to find the first time of contact (ftc)- if any - along every timestep. If a collision occurs between “compatible” entities, they can bind. As an output of the simulation we have sort of “snapshots” of the 3D environment at the end of the timesteps or at the moment in which a collision occurs.

Searching for possible engines/visualization toolkits to show this output, I’ve found JM which seems to be a complete solution also to manage collisions (and other stuff). Since V-Clip is old and tends to behave strangely in certain cases, I’m interested in replacing it with something newer.

  1. Can JM be suited for these purpose? In particular the TriMesh object as a replace for our polytree entities?
  2. Do you know other Java-implemented algorithms that work better than v-clip? Is there something that can be used from JM apart from TriMesh methods?
  3. Do you think that the project can be totally coded into JM? How much “knowledge” can we associate to every 3D entity?
  4. Our current architecture exploit clustering via an internal middleware we coded. How much flexible is JM to be tailored over such middleware?

    Sorry if the information I provided are partial. If you need further details, I’m here. Thanks in advance for any idea or feedback.