Bounding Volume Hierarchy for Avatar Collision Detection: Design Considerations

One of the most important aspects in interactive virtual environment applications is maintaining interactive frame rates while providing realistic and believable experience. Due to this fact, collision detection for interactive virtual environment applications acts as an important part of the system. However, due to its computationally intensive nature, collision detection can become a bottleneck to the system. Avatar can be seen as the user representation in the virtual world. User interaction and manipulation of avatar, and interaction between avatar and virtual objects made it impossible to anticipate collision. In order to minimize the collision test, a number of approaches have been proposed. One of the examples is by using a multiple-phase collision detection; where usually the earlier phase is carried out to cull away pairs of objects that do not possibly collide and the second phase is conducted to verify collision. Quite a number of proposed techniques and algorithms were introduced to accelerate the collision detection process. One of the most efficient and widely used in exact collision detection is bounding volume hierarchy. However, less attention has been directed on bounding volume hierarchy for human-like avatar. Therefore, the focus of this study is on bounding volume hierarchy to be used for collision detection between avatar and avatar, as well as avatar and virtual objects in a virtual environment application. Since bounding volume hierarchy is the main focus of this study and hierarchical representation is suitable for human-like avatar, various considerations in designing, constructing and updating the hierarchy is discussed. A proposed framework is presented along with a future plan for this study.