一种服装仿真的并行碰撞检测算法

ollision detection has been one of the major interests in 3D physical simulation. A data partition algorithm is proposed in this paper based on pre-existing cloth motion data. The algorithm includes the preprocessing stage and collision detection stage. In the preprocessing stage, the simulation models are partitioned into several collision clusters according to the inner triangles' motion distinction extracted from the pre-existing motion data, followed by the hierarchical bounding box tree construction for each collision cluster. In the collision detection stage, the collision detection process is parallel implemented utilizing every single collision cluster after the parallel update of collision clusters' bounding boxes. Experiments results confirm that the method can efficiently improve the collision detection efficiency.