【作者】 李凤娟；

【导师】 杨晓萍；

【作者基本信息】 吉林大学 ， 通信与信息系统， 2015， 硕士

【摘要】 如今，火灾已经成为最常见和最普遍的灾害之一，它不仅威胁公众安全，也影响社会发展，而火灾探测技术对于预防火灾发生以及降低灾后损失十分有效。传统的消防系统功能相对简单，用于火灾判断的特征参数单一，容易出现误报和漏报的情况。将数据融合技术应用在火灾判断中，可以获得更多被测目标和环境的信息，可以进一步提高消防系统的智能性和准确性。早在1973年，美国就已经开始有了对数据融合技术的研究，1988年将其列为重点研发技术，1991年已经将该技术应用于军事战争。目前多传感器数据融合技术已经由分散研究转变成一个独立的研究领域，并且在信息电子学领域、计算机科学领域以及自动化领域都有着重要的影响。比较常见的数据融合方法有贝叶斯方法、D-S证据推理、人工神经网络以及模糊理论算法等。数据融合具有三层结构，针对这一特性，本文选择在不同的层次使用合适的算法。这样的做法可以利用算法各自的优势，弥补彼此的不足，进一步提高融合结果的准确性和可靠性，从而减少火灾判断的误报率。本文先是构建一个基于ZigBee技术的小型火灾特征数据采集系统，分别从硬件和软件设计两个方面进行详细的阐述，分析Z-stack协议栈的概念与应用方法。分别对该采集系统进行了相关的功能和性能上的测试，实验结果表明，该系统可以有效地采集温度、烟雾浓度和CO浓度三种火灾特征数据。然后提出将神经网络和模糊理论相结合的火灾探测算法应用在火灾探测系统中。详细地从理论上分析了这两种算法的原理以及优缺点，利用MATLAB仿真软件中的神经网络工具箱以及模糊推理工具箱，对两种算法分别进行了仿真和分析，得出将两种算法结合适用于火灾探测的结论。最后在VS2010开发平台上开发后台数据处理软件，分别实现串口数据传输、火灾探测算法以及对数据库访问三个功能。为系统提供了一个较好的人机交互平台，开发者可以观测到实时采集出来的火灾特征数据值，并可以看到系统对火灾判断后的结果，也可以访问数据库查看历史数据。更多还原

【Abstract】 At present, fire is the major one of the most often and common threat to publicsafety and social development in various disasters, so fire detection technology is theeffective means to prevent fire and reduce the fire losses. The traditional function ofthe fire protection system is relatively simple. And the characteristic parameters forfire judgment is single. The traditional system often miss some faults and producefalse positives. It will obtain a greater degree of information of the measured targetand the environment to apply the data fusion technology in fire detection.And it canfurther improve intelligence and accuracy of fire protection system.As early as1973, the United States has already begun the research on data fusiontechnology. In1988, it grow to be the focus of its research and development oftechnology. The technology has been applied to the military war in1991. Currently,multi-sensor data fusion technology has changed from sporadic dispersed researchinto an independent field of study. It even has an important effect in the field ofelectronics, computer and information science.Bias method, D-S evidential reasoning, neural network and fuzzy theoryalgorithm is relatively common data fusion algorithm. Data fusion has three layers ofstructure, according to this characteristic, this article use the appropriate algorithms atdifferent levels. So it can play the advantages of different algorithm and can make upfor each other’s deficiencies. It can further to improve the reliability and accuracy offusion results, so as to reduce the false alarm rate of fire judgment.This article first built a small fire characteristic data acquisition system based onZigBee technology. Respectively from hardware and software to realize the twoaspects in detail and analysis of the concept and the application method of Z-stackprotocol stack. The acquisition system has been tested and performance-relatedfunctions. Experimental results show the system can effectively collect temperature,smoke concentration and CO concentration parameters of fire.Then this article proposed a fire detection algorithm applied in intelligent firesystems. It combines neural network and fuzzy reasoning algorithm. The writertheoretically analyzed the advantages and disadvantages of these two algorithms, additionally using neural network toolbox and fuzzy reasoning tools boxes ofMATLAB software to simulate and analyze these two algorithms. The conclusionstells us that the combination of these two algorithms is suitable for fire detection.Finally, the writer develops a manage software on VS2010development platform.Respectively, to achieve function of serial data transmission, fire detection algorithmsand access to database. This manage software provides a good interactive platformthat can be observed parameters of fire features in real-time acquisition and see theresults of the system after the fire judgment. The user can also access the database toview historical data.更多还原