群体运动仿真控制关键技术的研究与实现

发布时间：2018-01-08 17:10

  本文关键词：群体运动仿真控制关键技术的研究与实现　出处：《电子科技大学》2015年硕士论文　论文类型：学位论文

  更多相关文章： 社会力模型 A*算法 导航网格 相对速度障碍 最优碰撞避免

【摘要】：群体运动仿真控制技术是计算机数字媒体领域的重要技术,兴起于计算机技术高速发展的二十世纪末期,通过利用计算机知识、数物知识、设计知识等相关学科广泛的应用于科学、艺术、工程、商务、工业等多方面的领域,越来越多的行业开始将计算机群体运动仿真技术应用到他们的工作中,例如可以为城市建设规划和交通安全提供重要参考,还可以通过仿真灾难情况下的恐慌群体来制定合理有效的疏散方法等。本课题就是基于以上的研究背景下,详细分析了群体运动仿真的核心技术,在重点研究群体建模、路径规划和碰撞避免三个技术方向的基础上,实现了室内恐慌状态下的群体运动仿真系统。主要内容为:1、通过对环境建模和Agent群体建模技术的分析,实现了几何建模、空间建模的系统环境建模和感知模型、控制模型的系统群体建模。同时通过对社会力模型的详细分析,实现了加入社会力的群体建模,在此基础上分析了社会力方法的各个作用力之后,实现了一个加入刺激力的社会力模型的恐慌状态下群体建模。2、通过对A*算法的路径规划方法的分析,知道算法的计算效率和形成路径与网格的划分息息相关,在此基础上实现了一个基于三角形导航网格的A*算法来完成全局路径规划,通过划分三角形网格空间,减少了网格数量节约了存储资源,通过实验结果表明计算的效率更高,形成的路径更加平滑。3、通过对基于相对速度障碍的最优碰撞避免算法的分析,发现该算法在考虑Agent的碰撞避免时,采用的方案是让每个Agent承担碰撞避免的一半,互相避免碰撞,这样会导致算法比较死板,资源利用不合理,可以通过优先级的方式来确定每个Agent需要承担碰撞避免的多少,实现了一个基于优先级规则的相对障碍最优碰撞避免算法,优先级高的承担的碰撞避免任务少,低的承担的多,使算法更加智能化,运算速度更快。基于以上三点重点研究,实现了一个室内恐慌环境下的群体运动仿真系统,可以较为真实地表现出灾难情况下恐慌群体的运动特征。
[Abstract]:Mass motion simulation control technology is an important technology in the field of computer digital media. It arose in the end of 20th century when computer technology developed at a high speed. Through the use of computer knowledge, digital knowledge. Design knowledge and other related disciplines are widely used in the fields of science, art, engineering, commerce, industry and so on. More and more industries begin to apply the computer group motion simulation technology to their work. For example, it can provide important reference for urban construction planning and traffic safety, but also can make reasonable and effective evacuation methods through simulation of panic groups in disaster situations. This topic is based on the above research background. The core technology of group motion simulation is analyzed in detail. On the basis of focusing on group modeling, path planning and collision avoidance, this paper focuses on three technical directions: group modeling, path planning and collision avoidance. The simulation system of group motion under the condition of indoor panic is implemented. The main content is: 1. Through the analysis of environment modeling and Agent group modeling technology, the geometric modeling is realized. The system environment modeling and perception model of spatial modeling, the system group modeling of control model, and the group modeling with social force are realized through the detailed analysis of social force model. On this basis, after analyzing the various forces of the social force method, a social force model with stimulus force is implemented under the panic state. 2. The path planning method of the A * algorithm is analyzed. It is known that the computational efficiency and formation path of the algorithm are closely related to the mesh generation. On this basis, an A * algorithm based on triangular navigation mesh is implemented to complete the global path planning. By dividing triangular mesh space, the amount of mesh is reduced and the storage resources are saved. The experimental results show that the efficiency of calculation is higher and the path formed is more smooth. 3. Through the analysis of the optimal collision avoidance algorithm based on relative velocity obstacle, it is found that when the Agent collision avoidance is considered, the scheme adopted by the algorithm is to let each Agent bear half of the collision avoidance. Avoid collision each other, which will lead to a rigid algorithm, unreasonable use of resources, we can determine the number of each Agent need to bear collision avoidance through the way of priority. A relative obstacle optimal collision avoidance algorithm based on priority rule is implemented. The high-priority collision avoidance algorithm has fewer tasks and lower burden, which makes the algorithm more intelligent. Based on the above three key research, a group motion simulation system in indoor panic environment is implemented, which can truly show the movement characteristics of panic group in disaster.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.9
，

本文编号：1397960