The Development of an Intelligence-based Crowd Movement Model for the Simulation of Building Evacuation

Description

The required movement time for occupants to evacuate a building in the event of fire is crucial to ensure fire safety of the building. Here, a novel intelligent crowd-movement (iCM) model, based on Artificial Intelligence (AI) methodologies, is proposed. Specifically, artificial neural networks (ANN) models will be adopted in this project to capture evacuees' socio-psychological characteristics collected from a series of pre-arranged real-life evacuation experiments. The trained ANN models will be incorporated into the framework of the discrete element method to simulate the movements of individuals in the course of evacuation. As the proposed iCM model is formulated from the actual behaviour of the evacuees captured from real-life experiments, it will produce a realistic movement prediction. In addition, an existing automatic electronic drawing file reading system will be implemented in the iCM model for direct input of building geometry. This graphical user interface will increase the applicability and acceptability of the proposed iCM model to the building industry.