Emergency and Normal Navigation in Confined Spaces

My research interests lie in the field of emergency navigation and emergency management of confined places in both standard and emergency conditions. In conjunction, novel routing algorithms to realise energy efficiency in evacuation process are investigated. Disasters such as fire in buildings or vessels could cause a large amount of injuries and fatalities as well as enormous losses in properties. This problem aggravates in recent decades due to the urbanization in modern cities. To mitigate the loss during a disaster, fire drills are held to facilitate evacuees gain experience from simulated scenarios. However, the effectiveness of the fire drills is low due to drawbacks such as time-consuming and they can not model the actual situation in a disaster [35]. With the rapid development of the relevant technologies and the advent of low-cost wireless sensor networks, emergency response systems have become a significant component of large scale physical infrastructures [70]. Therefore, related navigation algorithms, asset allocation algorithms, reliable secure communication approaches [93], models and tools for emergency prediction and simulation have motivated considerable research. As the core of the emergency response systems, navigation algorithms aim to evacuate civilians to the exits safely and efficiently. However, due to the highly dynamic nature of an emergency environment, current state-of-the-art navigation algorithms for emergency still face a lot of challenges. For instance, most algorithms need a new convergence procedure or a full graph search when the hazard spreads. This process can be time-consuming due to the transmission interference and network congestion in communication [107]. Meanwhile, the communication overhead will increase exponentially with the scale of the building layout. Another problem is that most algorithms tend to find either the shortest paths or the safest paths but do not consider the impact of congestion, which is a vital factor in high occupancy rates [17]. Current congestion-aware algorithms can be divided into congestion-free algorithms and congestion-ease algorithms. Congestion-free algorithms which are based on the minimum cost network flow problem [6, 39] are impractical due to the high computational overhead and assumption of a non-spreading hazard. Congestion-ease algorithms employ queueing networks models [29, 110], flow-based models [132, 149] or potential-maintenance models [23, 24] to estimate the time delay on a path and redirect evacuees accordingly. However, the effect of current congestion located on a path is not taken into account. Furthermore,