摘要：Fire hazard monitoring and evacuation for building environments is a novel application area for the deployment of wireless sensor networks. In large room such a marketplace, it is crucial for firefighters to know the fire hazard situation and position, and further make plan to exterminate fire hazard. Otherwise, there are many shortcomings in conventional fire hazard monitoring system such as unknown fire hazard status and unknown fire hazard potion for the fire monitoring nodes are isolated in the large place. In this situation, the fire hazard information from these separated fire hazard monitoring nodes should be fused to deduce fire hazard situation. To solve the shortcomings such as low accuracy of fire situation assessment and fire localization during fire detection a in conventional fire hazard monitoring system, this paper proposes a fire detection and localization algorithm for the large place based on wireless sensor network. In wireless sensor network, the artificial neural network is firstly trained by fire data from the wireless sensor nodes, and the artificial neural network with fuzzy decision is then used to estimate and decide the fire situation. The smoke diffusion model and gas temperature diffusion model in limited space are applied to locate fire hazard source and measure the coverage of fire, so it is possible to assess fire situation more accurately in large room, which can provide more information for firefighters to take measures to exterminate fire hazard. Theory analysis and experimental results show that the fire detection and localization algorithm proposed is feasible.
