Research On The Collision Detection Algorithms In Virtual Environment

It is a challenge to detect collision in domains of virtual reality, animation, robotic, etc. It has become the basic issue that whether a contact or penetration has happened in virtual environment or not. Many articles have given a mount of valuable and practicable study on the problem. But while many novel domains come forth and the requirements for the reality of scene and real-time of interactive become more and more critical, it is important to improve the performance of collision detection to ensure the reality of virtual environment.After giving a comprehensive understanding and intensive analysis for all kinds of algorithms, the paper presents, implements and experiments a group of novel collision detection algorithms in three perspectives, aiming at handling problems of the existing collision detection algorithms.An algorithm that detects collision between two static convex polyhedra is presented. The method makes use of the principle that the value of an arbitrary point in the convex polyhedral can be presented by the values of its vertexes. The problem of detecting collision between two convex polyhedral is converted into a constraint nonlinear programming problem that aims at computing the minimum distance between two convex polyhedra. According to the value of minimum distance, the algorithm can determine whether two convex polyhedra have collided or not. Another, the approach is able to obtain the penetrate distance besides detection collision. The efficiency of the algorithm is illustrated by concrete instances. Basic genetic algorithm is used to solve the problem. The method deals with constraint conditions and initializes variant genetic operators firstly; then, genetic algorithm is applied to the distance model and compared with the result of traditional method. The results have shown that genetic algorithm has a lower execution time but a higher precision. Also, the results indicate that it is feasible and efficient to apply genetic algorithm to such kind of distance model.The storage structure of AABBs hierarchies is improved. The optimized AABBs hierarchies is used to detect collision between deformable objects. The algorithm makes use of the interactive detection between the basic geometry bodies in bounding box and optimizes the storage structure of bounding volume by removing leaf node storage information in the perspective of space complexity. Instances have illustrated the efficiency of the method.Most collision detection algorithms try to decrease the number of the intersection triangles. The algorithm presented in this paper simplifies intersection detection between two triangles from two aspects. By combining the characteristic of DirectX, algorithm makes use of the distance between a point and a triangle. According to whether the distance goes through a triangle or not, the algorithm decreases the steps of intersection collision between triangles .The excusive speed is accelerated and instances are used to illustrate the feasibility and efficiency. The concrete situation of interaction is obtained by judging the overlapping regional which is on the line of two triangles intersection. The language of JAVA and VRML are used to describe the algorithm that is well suitable for collision detection in the network environment.