Research And Application Of Collision Detection For Virtual Simulation

With the progress of technology,VR is widely used in manufacturing,military,education,health and entertainment.As one of the core issues in VR,collision detection has attracted more and more attention.Traditional algorithms may have situations of penetration and omission,which greatly weakens the authenticity of simulation.To improve the performance of virtual simulation system,an anti-penetration algorithm and a collision prediction algorithm are proposed.The main work of this paper is as follows.(1)Main collision detection technologies are studied,and causes of penetration and omission in collision detection are analyzed.To meet different requirements of collision detection,a 6-dimensional classification method is proposed.The basic principle of the hierarchical bounding box is elaborated in detail.(2)An anti-penetration algorithm based on OBB is proposed.To get triangles of collision area,a selection method of triangles based on OBB is designed.(3)A collision prediction algorithm based on AABB trees is proposed.By using unit quaternion,the relative motion in any direction between models is transformed into the relative motion along X axis.The hierarchical structure is used to quickly search the closest distance between models.(4)An experimental platform based on Unity3D is built.Taking two cases with penetration as the research object,the feasibility and effectiveness of our algorithms are verified.The results show that algorithms can solve problems of collision penetration in the experiment of virtual robot sucker grabbing object and vernier caliper measurement.