Key Lightweighting Technologies of Web3D for Virtual Training of Metro Station Fire Evacuation

Abstract

It would be a horrible scene when a fire breaks out in any underground metro station, which could result in high human casualties. Therefore, the training of fire evacuation is highly necessary for the public. Among various current training methods, online fire evacuation training is a very popular one, owing to its low-cost and convenience. However, due to the huge size of data of both the representation of the metro station and the fire scenario, the limited bandwidth of network, and the weak rendering ability of web browser, the online fire evacuation training simulation usually runs extremely slow or is not able to run at all. This paper proposes a new virtual reality online training system for metro station fire evacuation. Firstly, a method based on semantic and voxelization component checking is introduced for light-weighting the large-scale metro station static scene, and the BIM (Building Information Modeling) data can be reduced by as much as 10 times. Next, we propose a smoke redundant-removing and normalization method, which is used for substantially light-weighting the dynamic FDS (Fire Dynamics Simulator) smoking data (as much as 200 times in our experiments). With the above methods, the metro station and smoke data can transmit through the internet quickly, and real-time rendering can be achieved on web pages by using a multi-thread mechanism. Finally, we present our eACO (evacuation based on adaptive Ant Colony Optimization) algorithm which can be used for the planning of mass fire evacuation. A prototype system based on eACO is implemented for VR (Virtual Realization) the fire evacuation training on Web, with which the user just needs to surf the internet (without loading and installing plug-ins) and take part in the fire evacuation training. The experimental results demonstrate that the proposed solution is feasible for online training in metro station fire evacuation. The technologies developed are also suitable for fire simulation and evacuation training in other urban infrastructures.