Crowd Simulation And Collision Detection Based On Agent Method

With the speeding up of urbanization construction, increase of population’s density, there are more and more tall buildings, metro stations, pedestrian bridges, crossroads of transportation in cities. So the problems of public security and traffic are increasing rapidly. Especially when emergencies happen to the large buildings, subway stations, the pedestrian bridges, chaos and crowded behaviors appear easily:more seriously, it may lead to the crowd stampedes.In some degrees, these threaten the life safety of personnel. So it is very important to value the rationality of buildings, pedestrian bridges, public spaces. So how to deal with the public security problem effectively is essential.In recent years, with the development of computer graphics in games animation and virtual reality, Virtual reality technology has been widely used in the current application of information technology. Virtual reality system is mainly composed of animation design, three dimensional modeling and so on. Among them, the most important is the virtual environment and agent role perception of modeling. So the computer simulation, the three dimensional animation technology and Crowd simulation process of motion in reality have become hot spot in research at present.Based on OpenGL and OSG knowledge, this article mainly uses the Agent and combines Delta3D three-dimensional game development engine to build a crowd simulation platform of collision detection. In the process of the research, it uses navigation technology of seeds algorithm plan a path for an agent, uses Astar algorithm optimize the path. This paper makes models out of perceptions of Agent behavior. The platform tries to avoid collisions between he individual and the individual、individual and environment in the process of agent movement.Finally, based on simulation of crowd motion of the Delta3D open source three-dimensional game development engine in the environment of Visual studio2008, the results are tested under the two virtual scenes: small buildings and the pedestrian bridge of intersection in transportation.Test results can intuitively show the movement of people in the reality and the process. It is significant to virtual reality, system simulation and intelligent game development which is based on three-dimensional scenes. Moreover it also has certain guiding significance for the movement of people in reality.