An IFC-based framework for semantic integration of BIM and mobile crowd sensing in real-time evacuation routing

Emergency evacuation in complex buildings benefits from tight coupling between digital building representations and real-time operational data. While Building Information Modeling (BIM) provides semantically rich indoor descriptions, many evacuation systems still treat sensing technologies and BIM as isolated components, resulting in static or weakly adaptive routing behavior. This paper proposes an IFC-based integration framework that establishes semantic interoperability between BIM entities and mobile crowd sensing (MCS) data to enable real-time, context-aware evacuation routing. IFC models are transformed into a dynamic decision graph in which spaces, exits, and sensors are continuously annotated with live crowd-sensed information. Occupant positions are estimated using Wi-Fi–based indoor localization supported by a two-stage crowdsensing strategy that accounts for both active and silent users. Evacuation routing is formulated as a multi-criteria shortest-path problem using an extended Dijkstra algorithm with dynamically updated weights reflecting distance to hazards, travel time, exit availability, and a route safety index derived from environmental conditions and crowd density. Weight prioritization is formalized using the Analytic Hierarchy Process and examined through sensitivity analysis. The proposed framework is implemented and evaluated in a four-story academic building under three emergency scenarios. Experimental results demonstrate improved evacuation effectiveness, reflected by higher in-time evacuation rates and reduced evacuation times under scenario-specific time constraints. In addition, evacuation times exhibit lower variability under constrained conditions, indicating stable routing behavior when egress options are limited. Overall, the findings confirm that semantic coupling of BIM and mobile crowd sensing enables digital-twin–like behavior for real-time emergency management and illustrates how engineering informatics can transform static building models into operational decision-support systems.