基于多传感器融合的汽车火灾报警系统设计与实现

发布时间：2018-03-24 11:48

本文选题：汽车火灾　切入点：D-S证据理论　出处：《中国矿业大学》2017年硕士论文

【摘要】：随着汽车在现代生活中的重要性越来越高,汽车火灾的问题也越发引起了人们的关注。近年来,汽车火灾数量在全国各地均有上升的趋势,火灾损失比例不断加大,损失与纠纷越来越多,也更加复杂。本文针对该问题进行了一系列研究并设计实现了汽车火灾报警系统来解决这一问题。论文主要工作如下:1)对汽车火灾发生机理、特点及过程进行了分析并在此基础上确定了汽车火灾探测特征量。之后,结合所依托项目提出了汽车火灾报警系统整体设计方案并对系统拟采用关键技术进行了论证,最后,给出了系统评价指标。2)结合汽车火灾报警系统特点对经典D-S证据理论进行了分析并在此基础上提出了经典D-S证据理论的两种改进方法:混合式结构改进与信任度分配改进;通过对传统汽车火灾报警系统的阀值比较模型进行改进提出了基于高斯函数的汽车火灾特征量基本概率赋值函数模型;具体应用中,首先在实际环境中采集了“斯柯达”小汽车车厢内燃烧数据作为研究的原始数据,然后,通过粗糙集理论对该数据进行了信息挖掘(删除冗余条件并确定各汽车火灾特征量的信任度权重),最后,分别运用经典D-S证据理论、单一混合结构改进D-S证据理论、单一信任度分配改进D-S证据理论和同时使用两者的改进D-S证据理论对原始数据进行了数据融合,结果表明这两种改进全部有效,且综合使用两者的改进D-S证据理论完全可以满足系统需求。3)基于模块化设计原则完成了系统硬件设计工作。首先,将系统硬件划分为五个独立模块,然后,依次介绍了各个模块的实现过程并对模块进行了组合展示。最后,通过对汽车火灾机理和烟气扩散的研究确定了系统硬件的安装位置。4)基于分层设计原则完成了系统软件设计工作。先介绍了软件总体设计方案,然后,依次介绍了各层次软件的具体实现过程及运行原理。5)根据汽车火灾发生机理及模拟火灾实验原则设计了汽车火灾模拟实验。文中对实验场景、远程显示界面和数据记录情况等进行了展示;通过统计得出系统性能指标数据并将其与第二章评价指标进行对比,最终得出了系统功能、性能方面的整体评价,并对不足之处做了分析。结果表明,本汽车火灾报警系统完全满足设计需求。
[Abstract]:With the increasing importance of automobile in modern life, the problem of automobile fire has attracted more and more attention. In recent years, the number of automobile fires has been increasing in all parts of the country, and the proportion of fire loss has been increasing. There are more and more losses and disputes. In this paper, a series of research on this problem is carried out and a car fire alarm system is designed to solve this problem. The main work of this paper is as follows: 1) the mechanism of automobile fire. The characteristics and process are analyzed and the characteristic quantity of automobile fire detection is determined. Then the overall design scheme of automobile fire alarm system based on the project is put forward and the key technology of the system is demonstrated. Based on the analysis of the classical D-S evidence theory, two improved methods of classical D-S evidence theory are put forward: hybrid structure improvement and trust distribution improvement; By improving the threshold comparison model of traditional automobile fire alarm system, the basic probability assignment function model of vehicle fire characteristic quantity based on Gao Si function is put forward. First of all, we collected the combustible data of Skoda car car as the original data of the research in the actual environment, and then, The data is mined by rough set theory (removing redundant conditions and determining the trust weight of each vehicle fire characteristic quantity. Finally, the classical D-S evidence theory and the single hybrid structure are used to improve the D-S evidence theory, respectively. The improved D-S evidence theory of single trust distribution and the improved D-S evidence theory of both are used to fuse the original data. The results show that the two improvements are all effective. And the improved D-S evidence theory can completely meet the system requirement. 3) based on the modular design principle, the hardware design of the system is completed. Firstly, the hardware of the system is divided into five independent modules, and then, the system hardware is divided into five independent modules. The realization process of each module is introduced in turn, and the module is combined and displayed. Finally, By studying the mechanism of automobile fire and smoke diffusion, the installation position of the hardware of the system is determined. 4) based on the principle of hierarchical design, the system software design is completed. First, the overall design scheme of the software is introduced, and then, The realization process and operation principle of each level software are introduced in turn. 5) according to the mechanism of automobile fire occurrence and the principle of simulating fire experiment, the simulation experiment of automobile fire is designed. The remote display interface and data record are displayed, and the system performance index data is obtained by statistics and compared with the second chapter evaluation index. Finally, the overall evaluation of system function and performance is obtained. The results show that the fire alarm system can meet the design requirements.
【学位授予单位】：中国矿业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U492.83

【相似文献】