A fast collision detection algorithm based on mixed hierarchical bounding volumes(BVH)is proposed

which combines the quality of compactness of the fixed direction hull(FDH)with the simplicity of AABB

absorbs the merits of image-based collision detection algorithm

and utilizes the computing capabilities of graphic hardware. In the preliminary period the FDH tree of the objects is established. Then the tree is traversed and the original detection is carried out simultaneously. Subsequently

the AABB of the object is established to set the view parameter

and the detection result could be obtained in real time when the convex pieces of the objects are drawn. The experimental results demonstrate that the proposed algorithm works well in the collision detection for complexly shaped rigid-body

and the speed of the collision detection is improved by about 17% and 20% compared to the traditional algorithms RAPID and RECODE.