基于无线传感器网络的猪运动行为监测系统研究

发布时间：2018-02-22 08:33

本文关键词： 养猪无线传感器网络部署定位运动行为监测　出处：《华中农业大学》2014年博士论文　论文类型：学位论文

【摘要】：养猪产业是我国最重要的畜牧产业。随着现代养猪产业规模化发展,对养猪自动化管理水平提出了越来越高的要求。现有的RFID电子耳标技术通讯距离短、不具有扩展性,难以满足精准智能化猪养殖的迫切需求。因此将猪的环境、饮食、行为纳入到监测的范围,实现养猪过程的全过程、全方位地跟踪监测,能有效推动智能化猪饲养水平,提高养猪产业的管理效率,进而提高养猪场的经济收益。本文研究了将无线传感器网络(WSN)应用于养猪运动行为监测分析系统中的几个关键问题,包括圈养猪舍内无线传感器节点的边缘部署问题；猪舍内圈养猪的定位问题；猪舍内圈养猪的运动行为监测问题；基于WSN的养猪综合监测系统的架构和实体-关系(E-R)模型。研究结果如下： 1)评价圈养猪舍内无线传感器节点的边缘部署。在已有的覆盖率、覆盖效率、以及覆盖方差的基础上,提出了K重覆盖率、K重覆盖效率、K重全覆盖所需最小半径三个针对K重覆盖问题的评价指标,确立了猪舍内无线传感器节点部署方案的评价指标体系,并用该评价指标体系对4种可行性的部署方案进行了评价,得出最优部署策略,形成一套最优部署策略评价分析方法。 2)首次建立了K=1边缘覆盖问题中最少结点部署方案的计算方法。针对不同猪舍面积,不同节点通讯半径,以K=1全覆盖为目标,提出一套系统地以最少节点为目标的部署算法。在Adobe Dreamweaver软件开发平台及PHPnow互联网应用框架上,开发了猪舍无线网络节点部署边界条件及最小节点计算软件,对部署节点数进行计算验证；在Matlab软件平台上,开发了猪舍无线网络节点部署仿真分析软件,对部署方案进行仿真分析。 3)猪舍的三边直角加权质心算法的确立。对不同厂家的无线传感器节点进行研究和测试,运用曲线拟合,取得了0-80m和0-20m接收信号强度值(RSSI)与距离关系的最优模型；在三边加权质心算法的基础上,开发出针对猪舍的三边直角加权质心算法,并设计了该算法的具体运算步骤；10.43m×5.72m室内空间的测试验证表明,该算法可以达到平均误差1.346m,去掉两个特殊位置点后,平均定位误差可降低到1.0875m,每次定位时间间隔为10s。 4)研究了圈养猪运动行为的监测方法。设计了无线传感器节点的三轴加速度传感器硬件,以Z-stack协议栈为基础开发了加速度数据采集软件,并确定了与PC机通讯的通讯方法和通讯协议,实现了三轴加速度数据采集和通讯。 5)建立了神经网络模式识别分类算法。在神经网络模式识别算法的理论基础上构建了二层运动行为识别神经网络分类器,并设计了Matlab软件实现对四种运动行为(分别为走、跑、跳、静止)的识别。模拟试验证明,在85组训练样本试验时识别率达到100%,在439组大样本时,神经网络方法对走和静止两种状态的识别率达到100%,而跑和跳两种运动行为由于实际数据特征比较接近,其识别率分别达到99.1%和96.1%,整体识别率达98.9%。 6)圈养猪综合运动行为监测的研究。首次采用短时能量和短时过零率计算方法分析计算三轴加速度运动数据,设计了短时能量和短时过零率运动行为分类算法、FFT快速傅立叶运动行为分类算法和标准差运动行为分类算法。并基于VS2010平台开发了实时综合猪运动行为监测软件,通过实时模拟试验证明,采用短时能量分类法优于FFT快速傅立叶分类算法和标准差算法,对走、静止、跳和跑三大类运动行为进行识别分类时,识别率约为100%左右,区分跳、跑两种运动行为时,识别率约80%左右。 7)基于WSN的养猪综合精准监测系统框架的分析的设计。确立了基于WSN的养猪综合精准监测系统的框架结构和工作原理,并设计了网络化综合监测信息系统的实体-关系(E-R)模型。前期研究了染色竹条分级流水线的自动控制系统及高通量下水稻育种网络信息管理系统,在此研究的技术基础上,提出了基于WSN的养猪综合监测系统关键技术的研究。本文研究结果为无线传感器网络应用于养猪综合监测系统寻求了科学的解决方案,为后期进一步实现养猪综合精准监测系统提供了技术基础,并有着广阔的应用前景。
[Abstract]:The pig breeding industry is one of the most important animal husbandry industries in China . With the development of modern pig breeding industry , the demand for automatic management of pig is put forward . The existing RFID electronic ear tag technology has short communication distance and no expansibility , it is difficult to meet the urgent need of precision intelligent pig breeding . Therefore , the whole process of pig breeding process is realized , the monitoring can be carried out in all directions , the intelligent pig raising level can be effectively promoted , the management efficiency of the pig industry is improved , and the economic benefit of the pig farm is improved . This paper studies several key problems in the monitoring and analysis system of pig ' s movement behavior by applying wireless sensor network ( WSN ) to pig ' s movement behavior monitoring and analysis system , including the problem of edge deployment of pig house wireless sensor node , the problem of pig ' s movement behavior monitoring in pig house , and the structure and entity - relation ( E - R ) model of pig integrated monitoring system based on WSN . The results are as follows : 1 ) Based on the existing coverage , coverage efficiency and coverage variance , three evaluation indexes of K weight coverage , K weight covering efficiency and minimum radius required for K weight full coverage are put forward . The evaluation index system of wireless sensor node deployment scheme in pigsty is established , and four feasible deployment schemes are evaluated by using the evaluation index system , and an optimal deployment strategy evaluation method is formed . 2 ) The calculation method of minimum node deployment scheme in K = 1 edge coverage problem is established for the first time . For different pigsty area , different node communication radius and K = 1 full coverage , a set of deployment algorithm is proposed , which is based on minimum node . On the framework of Adobe Dreamweaver software development platform and PHPnow Internet application framework , the deployment boundary conditions and minimum node computing software are developed , and the deployment node number is calculated and verified . On the Matlab software platform , the deployment simulation analysis software of the wireless node of the pigsty is developed , and the deployment plan is simulated and analyzed . 3 ) The establishment of a three - sided right - angle weighted centroid algorithm for pigsty . The wireless sensor nodes of different manufacturers are studied and tested . Based on the three - side weighted centroid algorithm , a three - sided right - angle weighted centroid algorithm is developed for pigsty . The algorithm can achieve the average error of 1.346m , and the average positioning error can be reduced to 1.0875m after removing two special position points , and the time interval is 10s . 4 ) The monitoring method of the movement behavior of the loop pig is studied . The hardware of the triaxial acceleration sensor of the wireless sensor node is designed , the acceleration data acquisition software is developed based on the Z - stack protocol stack , and the communication method and the communication protocol for communication with the PC are determined , and the three - axis acceleration data acquisition and communication are realized . 5 ) The neural network pattern recognition classification algorithm is established . Based on the theory of neural network pattern recognition algorithm , two - layer motion behavior recognition neural network classifier is constructed , and Matlab software is designed to realize the recognition of four kinds of motion behaviors ( walking , running , jumping and rest ) . 6 ) The research on the monitoring of the comprehensive sports behavior of the pigs . The short - time energy and the short - time zero - crossing rate calculation method are used to analyze the three - axis acceleration motion data , the short - time energy and the short - time zero - crossing rate motion behavior classification algorithm , the FFT fast Fourier motion behavior classification algorithm and the standard deviation motion behavior classification algorithm are designed . The real - time comprehensive swine motion behavior monitoring software is developed based on the VS2010 platform . 7 ) The design of the framework of the pig comprehensive precision monitoring system based on WSN is designed . The frame structure and working principle of the pig integrated precision monitoring system based on WSN are established , and the entity - relation ( E - R ) model of the networked integrated monitoring information system is designed . This paper studies the automatic control system and the information management system of rice breeding network under high flux . Based on this research , the key technology of pig comprehensive monitoring system based on WSN is studied in this paper . The result is that wireless sensor network is applied to pig comprehensive monitoring system to seek scientific solution , which provides the technical basis for further realization of pig comprehensive precision monitoring system in the later stage , and has wide application prospect .

【学位授予单位】：华中农业大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TP212.9;TN929.5;S828

【参考文献】