Streaming Collision Detection Using Programmable Graphics Hardware

Zw Fan,Hg Wan,Sm Gao
2003-01-01
Abstract:Real-time collision detection is required by most of computer graphics applications, and many collision detection algorithms exploiting graphics hardware capabilities have been proposed. Recent advances in programmable graphics processing units (GPU) make it possible to be used in general-purpose computations such as ray tracing, simulation of dynamic phenomena, and scientific computations. In this paper, we explore to solve the collision detection problem with programmable GPUs. An approach to mapping collision detection between two star-shaped objects to programmable GPU is presented. Specifically, the approach converts the problem into intersection computations between rays and triangles so as to match the parallel architecture of GPUs. Several optimization techniques have been incorporated to improve efficiency. Several scenarios have beet? tested and experimental results have been given and compared with other typical collision detection algorithms.
What problem does this paper attempt to address?