Collision Detection

I need to devise a new algorith for collision detection using matrix representations of both the objects and the environment. Does anyone know of any research done in this area? And if so, is it still a popular method or not?


Hi there

What do you mean by matrix representations of objects and environment? Could it be voxel-based data?


I mean, we have a big matrix to store the information for all the objects in our world model. for example, the matrix below represents a triangle in 2D (1’s represent objects).


Therefore, instead of detecting collision through ray-tracing, collision between two planes, and other specific types, I use the info in the matrix representation to detect collision.

I am working on this topic for my research. I want to see if anyone knows about any work being done in this area.



You have probably tried this, but just in case, google shows a few interesting papers when you search for

real time volumetric rendering
voxel based collision detection

A lot of this sort of data comes from medical imaging, like CT scans. A few years ago I wrote a simulator for x-ray imaging and had to use volumetric data. If for every voxel (cubic element in your matrix rep) you have either 0 or 1, then defining a collision should be strightforward, and the only thing you have to worry about is performance. However, quite often you have a real number, related to how dense is the material occupying the voxel in question. It’s more tricky to define a collision here, especially if the data also contains noise.


Thanks for answering my question. It did help me a lot. I will post some more questions, if I encounter something new.

Thanks again.


The collision detection is a main topic in robotic research area. They uses a lot of matrix computation (like the matrix used in Opengl calculation) to determine if the robot is in contact with an objet.

You can look at some good paper on this topic on internet. However, somes basic matrix operation are needed to perform these calculation but it’s not too rought.

I hope that this will help you.