中文摘要：

针对商场等大型场所火灾监控存在的火情态势估计不明、监控节点孤立、预警准确率低、火源定位不准确等问题,提出一种基于无线传感器网络的大型场所火灾检测与定位算法。采用人工神经网络对火灾检测节点数据进行学习训练和火情估计,并经过模糊推理决策,给出火灾报警信息。利用受限空间烟雾扩散模型、气体温度扩散模型实现火源定位,并准确估计火情,以提供消防人员灭火决策。从单节点神经网络火灾报警仿真实验得出单节点火灾检测概率较高,同时从多节点火情定位实验得出,火情的覆盖范围描绘准确,火源定位精度较好。理论分析和实验结果表明火灾检测和定位算法是可行的。

英文摘要：

To solve the low accuracy of fire situation assessment,fire alarm and fire source localization during fire detection in lager room for regular fire monitoring system due to fire monitoring node isolation,a fire detection and localization algorithm for large place based on wireless sensor network is proposed.The artificial neural network is firstly trained by fire data from wireless sensor nodes,and the artificial neural network with fuzzy decision is then used to estimate and decide the fire situation.The limited space smoke diffusion model and the gas temperature diffusion model are applied to locate fire source and measure the coverage of fire,and it is possible to assess fire situation more accurately in large room.It can provide more information for firefighters to take measures.The simulation results show that not only single node fire alarm based on neural network has high fire alarm probability but also total fire detect wireless sensor network has high accuracy on fire situation assessment and fire source localization.Theory analysis and experimental results show that the fire detection and localization algorithm is feasible.