Research On Collision Detection Algorithm In3D Game

Virtual reality, as a maturing field of scientific research, is not a new industry. In the filed,the main research contents focused on improving the efficiency of the algorithms and technicalperformance. With the popularity of the Internet, the industry of games has gradually developedto iconic core industry of VR in recent years. When the technology of modeling and drawingmap has some difficulties, the technology of collision detection is facing enormous challengeswith the increasing the requirements of high quality in games from players. There are tens ofthousands of triangular pieces behind the beautiful screen. The complexity of large-scale scenein game directly restricts the efficiency of collision detection. However, the speed of observingfor eyes is slower than the speed of refreshing on computer. So the real-time will play moreimportant role than the accurate in simulation. Therefore, in order to improve the efficiency andperformance of collision detection, the key will contact with the real-time in this paper. Thefollowing is the main content.1.Basic Analysis. Firstly, read amounts of papers to collect the solutions of collisiondetection; then, learn the related technologies to realize these algorithms. According of thecharacteristics and scopes of the algorithms, summarize the strengths and weaknesses. In stageof research practice, understand and master the basics of math involved in the collision detection.Only in this way, this paper will have solid foundation.2.Solutions. Choose different solutions to deal with two types of collision detection whichare mainly in games. On the basis that the systems of collision detection can be divided intoseveral stages for improving, use different technologies to meet the requirements of specialstages. Thereby, improve the efficiency and performance collision detection of the game. Thispart of the study focused on two aspects: Firstly, using the ray and SAP in the initial stage. In thecase that the objects of research are the dynamic hero and static world, the test mainly dependson ray and the BSP. And the rays come from the center of bounding box of hero. The BSP isduty for dividing the world. For the other case, deal with the dynamic objects use the AABBbounding box projection. Sort the projection into array by straight insertion sort. Then calculatethe region of overlap on three axes to test intersection, in order to remove some research objectsof impossible collision. Secondly, use the mix of PSO and the random method in the stage ofaccurate collision detection. To get the sampling point from the center of the bounding boxesthese may collide. On this basis, build search space and update the information of every particlein order to get the most optimal solution. Finally, primitive test for the objects these are theresults from last stage. In the progress, a lot of informations will be got and be provided to the response of collision.3.Experimental Analysis. Use OpenGL to achieve improved program in the paper, andrealize it again by Opcode. At the same time, analyze the improving of the efficiency andperformance before and after the algorithm. Experimental results show improvement programhave good performance in most cases, for example, improved efficiency and superiorperformance. The program will be used in the game engine. The test result proved that theprogram has operability and practicality.