Simulation Method Of Large-scale Crowd Fire Evacuation In Virtual Reality Based On Chain Navigation Grid

The simulation of large-scale crowd evacuation in virtual reality is of great significance for emergency response.However,the existing large-scale crowd evacuation simulation methods are difficult to be applied to virtual reality scenes.Therefore,it is necessary to find a suitable simulation method according to the characteristics of evacuation simulation in virtual reality(VR)and improve the rendering efficiency of large-scale crowd.This paper presents a chain navigation grid for virtual reality large-scale crowd evacuation simulation and a vertex raster rendering method to improve the rendering efficiency of dense crowd.The method proposed in this paper is verified and analyzed by constructing fire evacuation scene,evacuation crowd model and developing prototype system.The main research work are as follows:(1)Chain navigation grid construction algorithm based on improved cellular automata.Firstly,the environmental familiarity factor and exit patency factor are introduced,and an improved cellular automata simulation method for crowd evacuation is designed to make the cellular automata evacuation simulation more accurate and realistic;Then,based on the improved cellular automata,a chain navigation grid construction algorithm is proposed,and the path point interpolation method and multi-floor connection problem of chain navigation grid are discussed.Finally,the relevant parameters of the chain navigation grid are measured,and the chain navigation grid for large-scale crowd evacuation simulation is realized.(2)Optimization method of large-scale crowd visual rendering based on vertex raster.Firstly,a general 3D model animation is constructed according to the model characteristics and physiological characteristics of the evacuated population,and the model and bone animation are transformed into texture and raster files that can be processed directly by GPU.Then,the fire evacuation scene is modeled,and the fire temperature and smoke are simulated by FDS software.Finally,combined with the above methods,an efficient visualization method for large-scale crowd evacuation is designed.(3)Prototype system development and test result analysis.The prototype system is developed based on unity virtual reality engine,oculus quest2 and other software and hardware;According to different test purposes,two test sites were selected to analyze the efficiency of evacuation simulation and crowd visualization.The results show that the efficiency of chain evacuation navigation grid evacuation simulation is improved by 6%.Compared with the traditional crowd rendering method,vertex raster rendering method improves the real-time maximum rendering crowd by 2 to 31 times.