大规模人群行为仿真模型的研究与实现
发布时间：2019-04-27 02:48
【摘要】： 随着计算机图形学与虚拟现实技术的发展,计算机动画的创作备受关注,作为计算机动画中的热点问题,群体动画在多个技术领域中得到了广泛应用,如建筑规划、数字化展示、影视动漫、电脑游戏等行业。从当前研究的情况来看,群体动画技术存在着比较突出的问题:一是目前采用的群体太过单一,基本上都是由一个或几个模板变换而来的,这样的群体过于单一,没有真实的人群的感觉,二是以当前的计算机硬件水平来说,要实现个体数量众多的群体动画仿真建模,使用直接绘制的方法很难达到实时性的要求。这些问题促使人们更加重视对群体动画建模技术的研究。 本文在收集并分析群体行为常用建模方法基础上,所做主要工作如下: (1)建立了几种群体动画行为仿真的关键模型 本文中群体行为建模的关键算法,包括感知模型的建模、决策模型的建模、行为规划模型的建模,对它们进行分析介绍之后,进行了实验,并分析了其优缺点。其中的感知模型分为静态环境和动态环境两种感知模型,决策模型的建模中,采用有限状态机来建立。 (2)提出了对基于势能场的群体行为建模技术的改进方法 该方法以连续动力学为基础,把空间划分为均匀的网格,使全局导航和局部碰撞避免整合到一个框架中,组成一个动态势能场,作为人群前进的驱动力,而当人群分布不均匀,有些网格不需要参与运算时,引入了静态势能场。这种方法能够减少所有网格都参与计算时的计算量,提高系统的效率。 (3)研究了群体行为一致性问题 主要介绍了受多个领域的研究人员关注的群体行为一致性问题方面的知识介绍了5种情况下群体行为一致性的条件和相应的数学表示方法和各自取得的成就,在这些成果的基础上对群体行为一致性的应用领域做了展望。
[Abstract]:With the development of computer graphics and virtual reality technology, the creation of computer animation has attracted much attention. As a hot issue in computer animation, group animation has been widely used in many technical fields, such as architectural planning. Digital display, film and television animation, computer games and other industries. From the current research situation, the group animation technology has a more prominent problem: first, the current group is too single, basically from one or more template transformation, such a group is too single, There is no real crowd feeling. Second, in the current level of computer hardware, it is very difficult to use direct rendering method to achieve the real-time requirements in order to realize the simulation modeling of group animation with a large number of individuals. These problems make people pay more attention to the research of group animation modeling technology. On the basis of collecting and analyzing the commonly used modeling methods of group behavior, the main work is as follows: (1) several key models of group animation behavior simulation are established in this paper, and the key algorithms of group behavior modeling are presented in this paper. It includes the modeling of perceptual model, the modeling of decision model and the modeling of behavior planning model. After analyzing and introducing them, the experiments are carried out, and the advantages and disadvantages of these models are analyzed. The perceptual model is divided into two perceptual models: static environment and dynamic environment. In the modeling of decision-making model, finite state machine is used to build the model. (2) an improved method for modeling population behavior based on potential field is proposed. Based on the continuous dynamics, the space is divided into uniform meshes, so that global navigation and local collision can be avoided in a single framework. A dynamic potential field is formed as the driving force for the advance of the population. When the distribution of the population is not uniform and some grids do not need to participate in the calculation, the static potential field is introduced. This method can reduce the computational complexity of all grids and improve the efficiency of the system. (3) this paper studies the problem of group behavior consistency, mainly introduces the knowledge of group behavior consistency problem concerned by researchers in many fields, and introduces the conditions of group behavior consistency in five cases. And corresponding mathematical representations and achievements, On the basis of these results, the application fields of group behavior consistency are prospected.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2010
【分类号】：TP391.41
本文编号：2466627
[Abstract]:With the development of computer graphics and virtual reality technology, the creation of computer animation has attracted much attention. As a hot issue in computer animation, group animation has been widely used in many technical fields, such as architectural planning. Digital display, film and television animation, computer games and other industries. From the current research situation, the group animation technology has a more prominent problem: first, the current group is too single, basically from one or more template transformation, such a group is too single, There is no real crowd feeling. Second, in the current level of computer hardware, it is very difficult to use direct rendering method to achieve the real-time requirements in order to realize the simulation modeling of group animation with a large number of individuals. These problems make people pay more attention to the research of group animation modeling technology. On the basis of collecting and analyzing the commonly used modeling methods of group behavior, the main work is as follows: (1) several key models of group animation behavior simulation are established in this paper, and the key algorithms of group behavior modeling are presented in this paper. It includes the modeling of perceptual model, the modeling of decision model and the modeling of behavior planning model. After analyzing and introducing them, the experiments are carried out, and the advantages and disadvantages of these models are analyzed. The perceptual model is divided into two perceptual models: static environment and dynamic environment. In the modeling of decision-making model, finite state machine is used to build the model. (2) an improved method for modeling population behavior based on potential field is proposed. Based on the continuous dynamics, the space is divided into uniform meshes, so that global navigation and local collision can be avoided in a single framework. A dynamic potential field is formed as the driving force for the advance of the population. When the distribution of the population is not uniform and some grids do not need to participate in the calculation, the static potential field is introduced. This method can reduce the computational complexity of all grids and improve the efficiency of the system. (3) this paper studies the problem of group behavior consistency, mainly introduces the knowledge of group behavior consistency problem concerned by researchers in many fields, and introduces the conditions of group behavior consistency in five cases. And corresponding mathematical representations and achievements, On the basis of these results, the application fields of group behavior consistency are prospected.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2010
【分类号】：TP391.41
【引证文献】
相关硕士学位论文 前1条
1 刘立栋;监控视频中大规模群体系统模型的研究[D];哈尔滨理工大学;2012年
，本文编号：2466627
本文链接：https://www.wllwen.com/wenyilunwen/dongmansheji/2466627.html
