Research And Realization Of Large-scale Simulation Model Of Crowd Behavior

With the development of Computer Graphics and Virtual Reality, Computer animation remains a major concern. As a hot topic in the reign of Computer animation, group animation has been widly applied in number of technique fields, such as construction planning, digital displaying, television animation, and computer game, etc. However, there still are a couple of issues existing to be dealt with in the felid of group-animation, which was judged from current research. The two of the prominent issues are:One is the adopted groups are much too single at present. Most of groups were converted from one or several models, which are too single to the feeling of the real human beings;The other is under the current condition of computer hardware, it's difficult to satisfy the real-time requirement in the means of drawing directly to realize modeling of group-animation with many numbers of individuals.Based on the collection and analysis of common modeling on group behavior, the main works in this paper are as follows:(1) Build several key models on understanding group animation behavior.The primary algorithms of modeling group behavior proposed in this paper include creating perception model, decision-making model, and behavior planning model. The experiments are followed by the analysis of these models. Then advantages and disadvantages of those models are listed in details as well. In the perception model, there are two models illustrated, namely static environment perception model and dynamic environment perception model. And finite-state machine is employed to build decision-making model.(2) Propose an improved method of modeling group behavior on the basis of Potential FieldThe method, based on continuous Dynamics, divides the space into uniform grid, and integrates the global navigation and partial collision to create a dynamic potential field which is used as people's drive force. And static potential field is introduced when people are not distributed uniformly, and some grids are not included in the calculation. This method can avoid all of grids participating calculation, as well as reduce calculation so as to prompt the system's efficiency.(3) Research the coherence problem of group behaviorThis part is mainly used to introduce coherence knowledge of group behavior, which is attracting many researchers'attention from a couple of fields. Here we use mathematical method to describe the coherence problem of group behavior. We also introduce the different coherence requirement of group behavior under five different conditions, their corresponding mathematic methods and their achievements respectively. Furthermore, we prospect some application reign of coherence problem of group behavior on the basis of those achievements.