Research On Collision Detection Algorithm Based On Particle System In The Virtual Simulation

Collision detection is used to determine the objects in a given time within the same time one or more pairs occupy the same area. It is one of the problems that can not be avoided in the field of robot motion planning, computer simulation, virtual reality, games, etc. With the computer hardware and software and network technologies become more sophisticated, especially the rapid development of technology of computer animation, simulation, virtual reality, people are eager to true simulation of the real world. One of the key technologies is the much-needed real-time collision detection.From parallel technology, mesh simplification, and particle swarm optimization, collision detection and three cloud-based computing point of view, analysis, design implementation and verification of a new collision detection algorithm based on particle system, how to use the graphics hardware high computing performance, programmability, and multi-processor parallel computing power, and speed up the collision detection process to improve the collision response techniques are useful exploratory study.Parallel technology of large-scale particle collision detection algorithm is used in the virtual scene. In this study, in parallel to handle large-scale particle system objects collision detection, parallel algorithms, divide and conquer strategy to establish the balance of each object in the environment bounding box tree, formed a task tree by traversing every two bounding box tree traversal, and traversal of the tree of all the tasks equally distributed to each processor, and then using the pipeline technique in the parallel algorithm, to accelerate the collision detection algorithm to traverse the task by dividing the process tree, the application of multi-threading technology in the process which can quickly detect a collision.The proposed collision detection algorithm based on mesh simplification and particle swarm optimization. Algorithm using the framework of the bounding box and bounding sphere model of mixed-level bounding box data structure, the first Garland grid to simplify the technology complex scenes, the error in the preprocessing stage to allow simplified within narrow the search space of particles, using particle swarm optimization algorithm for fast search. Experiments show that this algorithm greatly improves the efficiency of collision detection, to achieve real-time detection.Proposed model based on cloud computing collision detection optimization algorithm. MapReduce is a key technology for cloud computing, it is a programming model that Google has developed a simple and abstract to achieve distributed computing tasks, you can simplify the data processing tasks on large clusters consisting of general-purpose machine. The idea is to be performed, disassembled into the Map and Reduce the way to parallel computing, or the effect of the distributed processing, The cutting data, task scheduling, junction point communications and hearts of fault-tolerant features by MapReduce be done automatically. In MapReduce, the user only needs to define the Map process and Reduce process, you can achieve the task parallelization.