60 likes | 212 Views
Collision Detection Optimization. By: Iskandar Ng Eng Guan. Optimization Approaches. Bounding Volume: Sphere Box (AABB, OBB) K-DOP (Discrete Orientation Polytope ) Convex Hull Spatial Partitioning: Octree (1->8 1 ->8 2 ->8 3 ->8 n leaves)
E N D
Collision Detection Optimization By: Iskandar Ng Eng Guan
Optimization Approaches • Bounding Volume: • Sphere • Box (AABB, OBB) • K-DOP (Discrete Orientation Polytope) • Convex Hull • Spatial Partitioning: • Octree(1->81->82->83->8n leaves) • BSP tree (1->21->22->23->2nleaves) • Hash Function
Bounding Volume Sphere 1. Bounding Sphere 2. Sphere Intersection AABB Collision when: d < r1+r2
Spatial Partitioning Octree • Represent the 3D volume with a cube. • Subdivide cube by half in each direction: We get 8 sub-cubes. • Repeatedly subdivide each sub-cube the same way.
Conclusion • Design Factor: • Geometry representation • Bounding Volume Simplification • Type of queries • Simulation parameter • Rigid, • # objects, • Interpenetrate • Ease of implementation • Ensure deliverable project