An indoor crowd simulation using a 2D-3D hybrid data model

Recent LBS-related technologies tend to extend to indoor spaces using localization sensors such as RFID. In order to implement real time evacuation applications, at least two problems must be resolved in advance; first, proper indoor data models and implementation methods that can accommodate evacuees positioning and routing computations should be available, second, evacuation simulations also need to be performed using the same indoor databases for consistent integration. However, none of these have been suggested explicitly as of now. Although some 3D modeling studies have dealt with topological structures, they are mainly focused on outer building volumes and it is difficult to incorporate such theoretical topology into indoor spaces due to complexity and computational limitations. In this study, we suggest an alternative method to build a 3D indoor model with less cost. It is a 2D-3D hybrid data model that combines the 2D topology constructed from CAD floor plans and the 3D visualization functionality. We show the process to build the proposed model in a spatial DBMS and visualize in 2D and 3D. Also, we illustrate a test CA (cellular automata)-based 3D crowd simulation using our model.