基于动作识别的矿井人员定位系统的设计

发布时间：2018-02-02 02:12

本文关键词： 矿井人员定位小波包支持向量机动作识别虚拟仪器　出处：《吉林大学》2017年硕士论文　论文类型：学位论文

【摘要】：煤炭是世界各国的支柱型能源,我国属于煤炭大国,煤炭产量高居世界前列,但国内中小型煤矿普遍存在井下环境复杂、条件差且巷道狭窄等安全隐患问题,并且井上管理人员不能及时了解井下作业人员的运行轨迹及实时位置,一旦发生矿井事故导致人员被困,救援队伍会因不能及时掌握井下被困人员位置而难以展开工作,对救援工作是十分不利的。本文基于国内现有矿井人员定位系统定位精度低及成本高昂等问题,提出了一种基于动作识别的矿井人员定位系统,该系统的工作机制如下:井下人员需佩戴数据采集单元,将采集到的数据信号通过蓝牙模块传输给上位机软件系统,上位机软件系统采用小波包特征提取及支持向量机(SVM,Support Vector Machine)分类算法确定人员的动作类型(包括走步、跑步、站立),再依据Y轴加速度信号数据的波峰数得到人员动作的数量,事先输入井下人员的步行步幅及跑步步幅,从而通过计算确定井下人员的运行距离,同时根据电子罗盘判定人员是否转弯,确定出井下人员的运行轨迹,并结合射频识别(RFID,Radio Frequency Identification)作为误差校正时使用,最后比对井下平面图确定人员位置。具体的研究内容分为如下几个方面:首先,研究人员步行的周期性状态及其特点,由此得出数据采集单元的最佳佩戴位置,并对采集到的数据信号进行观察分析,确定采用小波包特征提取方法来处理信号数据,提取人员动作信号的特征向量。其次,对系统中的关键技术进行了简单介绍,并对动作信号进行时频分析,确定了系统核心算法采用小波包特征提取及支持向量机分类的方法,为系统提供了理论基础。再次,本系统采用支持向量机分类器来区分人员的动作类别,其分类依据为动作信号的能量特征向量,因为人员在做不同动作时会产生不同的振动能量。将通过小波包特征提取到的能量特征向量送入SVM中进行数据训练和分类,以此来实现人员动作类型的判定。最后,设计了一套矿井人员定位系统,由数据采集单元及上位机软件系统两大部分组成。数据采集单元由三轴陀螺仪、电子罗盘、蓝牙通信模块、三轴向加速度计及单片机最小系统组成,完成对人员动作信号的采集传输功能;上位机软件系统从虚拟仪器的角度出发,采用Lab VIEW及MATLAB语言编写,完成对信号数据存储、算法应用、数据处理及井下人员定位等功能。
[Abstract]:Coal is the pillar energy of all countries in the world. Our country belongs to the big coal country, and the coal output is high in the world. However, the underground environment is complex, the condition is poor and the roadway is narrow and so on. And the well management personnel can not understand the operation trajectory and real-time location of underground operators in time, once the mine accidents lead to people trapped. Rescue team can not grasp the location of trapped underground workers in time and difficult to start work, which is very unfavorable to rescue work. This paper based on the problems of low positioning accuracy and high cost of the existing mine personnel positioning system in China. A mine personnel positioning system based on action recognition is proposed. The working mechanism of the system is as follows: underground personnel need to wear data acquisition unit. The collected data signal is transmitted to the upper computer software system through Bluetooth module. The upper computer software system adopts wavelet packet feature extraction and support vector machine (SVM). The Support Vector Machine classification algorithm determines the movement type of the person (including walking, running, standing). Then according to the Y-axis acceleration signal data of the number of peaks to get the number of personnel action, and input in advance the walking and running stride of the underground personnel, so as to determine the operating distance of the underground personnel by calculation. At the same time, according to the electronic compass to determine whether the personnel turn, determine the trajectory of the underground personnel, and combined with RFID RFID. The Radio Frequency Identification is used as an error correction. Finally, compared with the underground plan to determine the location of personnel. The specific research content can be divided into the following aspects: first, the periodic state and characteristics of the walking of the researchers, from which the data acquisition unit of the best wearing position. And the collected data signal is observed and analyzed, the wavelet packet feature extraction method is adopted to process the signal data and extract the characteristic vector of the human action signal. Secondly. The key technology of the system is introduced briefly, and the time-frequency analysis of the action signal is carried out, and the method of wavelet packet feature extraction and support vector machine classification for the core algorithm of the system is determined. Third, the support vector machine classifier is used to distinguish the category of action, which is based on the energy feature vector of the action signal. Because people will produce different vibration energy when they do different actions, the energy feature vector extracted by wavelet packet feature will be sent into SVM for data training and classification. Finally, a mine personnel positioning system is designed, which consists of two parts: data acquisition unit and PC software system. The data acquisition unit is composed of three-axis gyroscope. Electronic compass, Bluetooth communication module, triaxial accelerometer and the minimum system of single chip microcomputer, complete the collection and transmission of human action signal; From the point of view of virtual instrument, the upper computer software system uses Lab VIEW and MATLAB language to complete the functions of signal data storage, algorithm application, data processing and downhole personnel positioning.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD76

【参考文献】