An architecture for safe evacuation route recommendation in smart spaces

In this paper, we treat pedestrian evacuation in emergency scenarios of networked smart spaces. Personal safety may be jeopardized due to natural catastrophes (e.g., hurricanes, earthquakes, etc.) and/or adversarial actions of intentional enemies. During evacuation, the severity of emergency may increase causing partial or complete blockage of some evacuation routes. Thus, it is of the highest importance to (re)route evacuees based on updated real-time structure safety conditions. In this paper, we propose a multi-agent based architecture for dynamic route safety optimization in large smart space evacuation. The objective of the model is to ensure that the smart space network gets evacuated securely while aptly responding to unpredictable contingencies in the network safety.