Collision Detection for Cloth Based on Adaptive Enclosing Ellipsoids

Fast and accurate collision detection between cloth and models is essential in cloth simulation.This paper proposes a new method for collision detecting based on adaptive enclosing ellipsoids.We use the weighted average of the normal distance and the radial distance to represent the surface error between an ellipsoid and the model.Based on the surface error, a segmentation of the model is adaptively computed with the optimized K-mean clustering method for generating a union of tight enclosing ellipsoids.These ellipsoids are classified into two kinds according to the surface error.While the surface error of the ellipsoid is small enough, cloth grid points are detected with the ellipsoid instead of the model.Otherwise, cloth grid points should be detected with triangles of the model after being detected with the ellipsoid.Experiments demonstrate that this method can improve the detection efficiency and guarantee the detection precision.