Emergency and Normal Navigation in Confined Spaces

Abstract: Emergency navigation algorithms direct evacuees to exits when disastrous events such as fire take place. Due to the spread of hazards, latency in information updating and unstable flows of civilians, emergency evacuation is absolutely a complex transshipment problem involving numerous sources and multiple destinations. Previous algorithms which commonly need either a full graph search or a convergence process suffer from high computational and communication overheads. This research report surveys the current emergency navigation algorithms and adapts the concept of Cognitive Packet Network (CPN) to the context of emergency evacuation. By using random neural networks, the CPN based algorithm can explore optimal routes rapidly and adaptively in a highly dynamic emergency environment with low expense. Simultaneously, in emergency situations there are typically different categories of evacuees such as people of different age groups. However, current algorithms only consider "normal" evacuees and do not meet the specific requirements of diverse evacuees. Our algorithms make use of the flexibility of CPN which can operate with different user-defined goals to customize appropriate paths for each category. The CPN algorithm is simulated in a graph based discrete-event simulator and Dijkstra's shortest path algorithm is taken as reference. The results show that the CPN algorithm reaches the performance of ideal path-finding algorithm and quality of service is improved by using specific goal functions for diverse categories of evacuees. Finally, we present a future plan for further research.