A semantic framework for on-site evacuation routing based on awareness of obstacle accessibility

The dynamic nature of construction sites increases the complexity of evacuation routing in emergency situation. The existing emergency response in construction site adopts manual management for monitoring which requires a lot of manpower and is prone to error. The Building Information Modeling (BIM) based evacuation routing methods is usually implemented by as-designed model without reflection of real scene. Therefore, this paper proposes an obstacle-aware framework for construction site to efficiently generate evacuation routing based on BIM, Computer Vision and Semantic Web, maintaining the consistency of model and real scene. The methodology is tested in a real construction site. Experiment shows that the awareness of obstacle accessibility supported by the developed semantic model can improve the accuracy of evacuation routing. Comparing with the existing evacuation routing methods, the proposed obstacle-aware framework improves the accuracy of evacuation routing and further provides a reference for other applications that need environmental awareness.