【作者】 马静；

【导师】 张晓彤；

【作者基本信息】 北京科技大学 ， 计算机科学与技术， 2016， 博士

【摘要】 当前,急需一种有效的、可用于地表下通信与定位的技术解决震后、难后人员通信、精确定位、搜救等难题。然而,由土壤、沙石、水等复杂成分构成的地下通信介质严重制约现有高频电磁通信算法和天线尺寸,使得传统电磁波通信技术不再适合。对于建立地下无线通信系统,磁感技术是一个非常好的选择。比起电磁波通信技术,近场低频磁感通信技术具有无多径效应,衰减速度相对较慢,无大延迟,小尺寸天线等优点,更适用于透地通信以及透地定位。然而,通信介质的复杂性,通信范围的局限性和透地定位的精准性等难点问题依旧存在。本论文就以上难题提出基于电小天线的地表下近场磁感应通信在不同部署条件下的传播模型及应用：通过对综合考虑地下环境的点对点通信模型的研究,掌握其基本传播特性；通过对基于磁中继技术的二维、三维天线阵列通信模型的研究,提出长距离通信策略；通过对基于三维天线阵列模型定位技术的研究,提出对地表下磁感通信重要应用——高精度定位的解决方案。本文主要创新性工作如下：1)综合考虑土壤成分和噪声等环境因素,提出无磁中继点对点通信模型并验证。传统近场磁感通信模型利用理想环境下天线间互感模型进行搭建,没有考虑地质因素和噪声因素。本文综合考虑地质因素中含水量比重对土壤电导率的影响,噪声因素中环境噪声和热噪声对接收电压强度的影响,增加磁感通信中互感公式的电导率因子、接收电压种类两项因素,提出地表下近场磁感通信无磁中继点对点通信理论模型。同时,设计并验证通信装置,利用通过校验的通信装置在不同尺寸线圈、不同含水量、不同角度和不同距离下进行多次实验,对磁场强度、接收功率、路径衰减等传播特性进行研究。2)利用磁中继技术,提出二维天线阵列通信模型并研究其传播特性。相较于增加线圈尺寸的方法,采用磁中继技术具有天线尺寸小、通信距离长的优点,特别适合地表下长距离通信和天线易埋设的应用场景。本文分别就强、弱耦合状态下直线型天线阵列的接收功率进行研究,提出强耦合状态下天线的最优摆放间距以及弱耦合状态下的最优通信频率的算法。在此基础上,进一步提出基于六边形天线阵列的传播模型,在色散方程、带宽、群速度、能量密度和路径衰减五种传播特性下与传统四边形天线阵列进行对比。仿真实验结果表明,在二维空间下,相较于传统四边形阵列,本文提出的六边形天线阵列带宽提高15.4%,能量密度增加了3倍。3)提出基于磁中继技术的三维天线阵列通信模型,研究其传播特性；并利用该模型提出了基于模式识别与分类的三维定位算法。三维六边形天线阵列的传播特性在带宽、群速度和能量密度等方面都有很大的提高。特别的,比起三维四边形阵列,其带宽提高了15%,能量密度增加了3.6倍。基于此天线阵列,本文归纳出两种常见井下部署场景,综合考虑收发端频率波动、信道噪声强度波动、土壤含水量波动、摆放角度的不确定性五项随机因素,提出基于支持向量机模式识别与分类算法的透地定位技术,克服传统定位技术在包含信道随机因素、通信系统的随机因素和场景部署的不确定性的情况下的不适用性。结果表明,本文提出的方法不但给出了不确定因素和随机性存在时,定位精度超过96.5%的方法,而且表明利用机器学习算法将磁感通信扩展到恶劣环境的适用性。更多还原

【Abstract】 Recently, due to the urge requirement of human accuracy localization and survival signal propagation after earthquake and mine disaster, underground communication and localization technology are currently a fairly active and promising research field. However, owing to the complexity of underground propagation media consist of soil, rock, and water at varying ratios, the original electromagnetic technology does not operate well because of the limitation on traditional algorithm and antenna size. Magneto-induction (MI) technique is an excellent candidate to establish underground wireless links. Compared with electromagnetic (EM) communication technology, the near field low frequency MI communication technology is more suitable for through-the-earth communication and localization due to the advantages of no multipath effect, relatively slow attenuation, no large delay and small size antenna. However, the complexity of the communication medium, the limitations of the communication range and the precision of the through-the-earth localization are still difficult problems. On the above problems, this paper proposes the propagation model and application of near field MI communication under different deployment conditions based on the electric small antenna:through the research on point-to-point model, the basic propagation characteristics are showed; through the research on two- and three- dimensional antenna array communication model based on magnetic relay technology, long distance communication strategy are proposed; through the research on localization based on three-dimensional antenna array model, the important application—high precision localization solution is proposed. The major innovations are as follows:Propose a channel model and study its propagation characteristics by comprehensive considering geological and environmental factors. Traditional near-field MI communication model is built by mutual inductance between antennas under ideal environment without taking geological and environmental factors into account. Through considering the effect of moisture content on soil electrical conductivity and the influence of ambient noise and thermal noise on the received voltage intensity, we build underground near-field MI point-to-point communication model. Meanwhile, design and verify the communication device.By use of this device, we do many experiments in different sizes of coils, different moisture, different angles and different distances to study the propagation characteristics of magnetic field strength, receiving power and path loss, and so on.Propose a two-dimensional antenna array and study its propagation characteristics by use of magnetic relay technology. Large size of coil could get long distance communication, but small coil is more suitable for burying. In this paper, the received power of linear antenna array is studied under the condition of strong and weak coupling, and the optimal communication relay distance and frequency are given under the condition of strong coupling and weak coupling, respectively. On this basis, the hexagonal antenna array is put forward. Five kinds of propagation characteristics in the dispersion equation, bandwidth, group velocity, energy density and path loss are compared with traditional quadrilateral antenna array. The results show that the hexagonal antenna array proposed in this paper is superior to the conventional quadrilateral array in the five propagation characteristics. Furthermore, the bandwidth of the new array increased 15.4% and the energy density of the new array increased 3 times.Propose a three-dimensional (3D) antenna array and study its propagation characteristics; by use of this model, propose a 3D through-the-earth localization algorithm based on pattern recognition and classification. The propagation characteristics of three-dimensional hexagon antenna array are greatly improved in bandwidth, group velocity and energy density. Especially, the bandwidth of the new array increased 15% and the energy density of the new array increased 3.6 times. Based on this array, we generalize two deployment scenarios and propose a generic approach combining support vector machine (SVM) and classification by comprehensive consideration of five random factors, including the frequency fluctuation of the transceiver, the fluctuation of the channel noise intensity, the fluctuation of soil water content, and the uncertainty of the angle of display, to overcome the disadvantage of traditional localization method in harsh environments with abundant sources of random effects. Our results not only show that the proposed system provides a localization accuracy of over 95% when uncertainties and random effects are present, but also suggest that the applicability of magnetic-induction communication can be extended to harsh environments by machine learning algorithms.更多还原