Autonomous navigation systems for emergency management in buildings

The evacuation of urban areas during an emergency is complex and challenging due to the dynamic conditions and ambiguity of information available to people in the affected area. Autonomous navigation systems can improve the outcome of such evacuations by providing up-to-date guidance and directions to people during the emergency. In this paper we present two distributed navigation systems deployed inside a confined space, such as a building, that use simple but effective communications to gather and disseminate information for the computation of evacuation paths. The first system is composed of a network of static decision nodes (DNs) positioned in the building, where DNs distributedly compute the best paths using local communication and computation, and each DN provides directions to people in its vicinity. The second system is composed of mobile communication nodes (CNs) carried by the people in the area. CNs form an opportunistic network in order to exchange information regarding the hazard and each CNs directs its user towards the safest/closest exit. Sensor nodes pre-deployed in the building monitor the environment and provide their measurements to both systems. We investigate the effect of failures of DNs on the evacuation outcome and study how the two systems can be used in conjunction to overcome such problems. A multi-agent simulation platform is used for the performance evaluation of our proposed systems in evacuation scenarios inside a three-floor building.