Collision detection algorithm for deformable objects using particle-based method

Abstract

Most collision detection algorithms have been proposed based on hierarchical bounding representation. These techniques are notable to be used with deformable surfaces because their bounding representations have to be updated when surface deformation occurs which costs quite expensive. Therefore, this research proposed an alternative algorithm for collision detection among non-rigid deformable polygonal models using particle-based method. The basis of this algorithm is to set each particle as a sensor in a separated area to determine whether there exists a high possibility for collision. Surface partitioning is firstly applied by equally dividing vertices into several groups which can be approximated as the number of objects that can touch the surface. Each area is then assigned with a particle moving inside by the attractive forces from other particles on neighboring objects. If the collision occurs, their corresponding particles will also be collided. A tolerable parameter is properly set in order to determine the distance between two particles on the verge of collision