【作者】 黄玲；

【导师】 曾昭发；

【作者基本信息】 吉林大学 ， 地球探测与信息技术， 2010， 博士

【摘要】 本文首先分析了当前国内外探地雷达系统的系统参数和其探测性能,总结归纳探地雷达系统的类型及基本探测信号形式,并说明了常规探地雷达系统中存在的缺点和改善系统性能的技术难点。在此分析基础之上,作者提出了多输入多输出(Multiple-Input and Multiple-Output: MIMO)技术在探地雷达方法中应用的思想。本论文的主要内容是分析多输入多输出(Multiple-Input and Multiple-Output: MIMO)技术在探地雷达方法中的应用问题。首先从理论上论证了多输入多输出(MIMO)技术在探地雷达中应用的可行性和技术优势,给出了MIMO探地雷达的信号模型,然后通过建立数值模型及物理实验来里进行目标探测性能分析。研究表明,MIMO技术相对于常规探地雷达系统和常规探地雷达观测方法具有较高的目标探测精度和信号稳定性。本文提出了正交信号的多输入多输出(MIMO)探地雷达系统和基于似平面波(Plane Wave Like)多输入多输出观测方式的探地雷达系统,模拟分析两种系统的目标响应特征和目标成像效果。在基于正交信号的MIMO探地雷达系统的目标成像方面,本文提出基于目标回波到达角(DOA)估计的逆向成像方法。理论模拟分析及对比了最小二乘(DAS)、Capon、幅度相位估计(APES)等三种波束形成算法在目标DOA估计上的精度及最终的成像效果。在基于似平面波的多输入多输出观测方式的探地雷达系统(Plane Wave Like MIMO GPR: PWL-MIMO GPR)的方法及目标成像效果分析研究工作中,作者进行了不同形态目标体模型的模拟分析,有效地说明了多输入多输出观测方式能有效地提高信号的信噪比和稳定性。从总体来说,多输入多输出技术(MIMO)可以有效克服现有探地雷达系统存在工作频带有限、系统实现难度大等问题,提高了系统的目标检测能力,并降低系统设计难度。更多还原

【Abstract】 Ground Penetrating Radar (GPR) methods and techniques is playing an increasingly important role in production applications and cutting-edge scientific exploration and research, such as engineering and environmental geophysical exploration, engineering quality testing, polar and aerospace exploration. With increasing requirements of data quality and much of information on the production and scientific research, ground penetrating radar also shows a number of deficiencies. The first is that pulse radar system needs the transmitting antenna efficiently to launch narrow width time pulse with very high power, high-fidelity to achieve higher resolution and deeper probing depth, but the system’s hardware it is very difficult to realize that. Second, when sending and receiving signals who is a narrow time width pulse, it means that the signal is broadband, yet the system often faces design difficulties of nonlinear strong gain uneven, weak parameter setting for the areas of flexibility broadband signal generating, power amplifying, transceiver antenna power gain, radiation pattern, receiving signal conversion, data acquisition and storage. In addition, the pulse width the system transmitted signal becomes narrower, the lower the average power signal, so the smaller the depth of systems effective detection, and the signal to noise ratio (SNR) and stability of the received signal become worse, resulting in objective interpretation difficulties.GPR systems even can be achieved using ultra-wideband and narrow time width pulse for detection, but due to the large low-pass filtering and dispersion effect of background earth medium for electromagnetic wave, after the wide-band transmission signal is suffered the earth’s filtering effect, the receiving antenna can only to access limited bandwidth signal, and most of the signal band is affected by the dispersion effect, so wavelet shape is with distortion. Noise and interference signals play the dominant part in the high frequency part of received signal, which would cause a serious decline in performance of signal processing methods and difficultly to interpret the parameters of objects.Shifting in the data migration processing for ground penetrating radar, most data migration approach widely used in ground penetrating radar is assuming the radiation and reception of antenna is with uniform characteristics and ignores the target radar cross section (RCS) vary with the incident angle and reflection angle, and shows a fluctuations phenomenon. The direction of the antenna characteristics and the RCS Fluctuations of objects would make the profile of common offset observations to appear discontinuity on horizontal direction, and that ultimately affects the performance of data migration processing, also for imaging results.To address the above problems, the author introduces the multiple-input multiple-output (MIMO: Multile-Input and Multiple-Out) thinking to the GPR method to solve or avoid the above problems. The low-pass filtering of underground media is the role and nature of media properties, it will not be changed with vary detecting signal form and observation method. So to improve radar target detection, it will need to get more objects information from the limited, available received signal. For a sentence to say is that using low-frequency narrow-band signals to achieve high-resolution target detection. On the subject of study, the author mention using MIMO technology to enhance the detection resolution and probing depth of ground penetrating radar system, and also want to reduce radar hardware system design requirements, more ever to reduce the complexity data of processing and imaging methods. This paper introduces two multi-input multi-output (MIMO) thinking to the GPR method: the first is a co-location multiple-input multiple-output GPR radar system (MIMO-GPR) based on orthogonal; the other is a multiple-input multiple-output observation ground penetrating radar system who is working on plane wave like mode (Plane Wave Like MIMO GPR: PWL-MIMO GPR).Applicationco-locationbasing on orthogonal signal, the first work is to collect and analysis the charge of multiple-input multiple-output (MIMO) technology in wireless communication technology, sounding radar technologies, voice processing, medical imaging and other areas of testing innovative technologies and research results, particularly for the using of narrow-band orthogonal signal in MIMO technology of these research areas. And then I summarized up the advantages of MIMO technology, and the comparisons of these advantages with the problems of conventional pulse ground penetrating radar were analyzed, and then looked for a ideas to solve conventional pulse ground penetrating radar problems. The author analyzed the application feasibility of MIMO technology introduced to GPR method, and proposed an idea of MIMO technology applied in ground penetrating radar, basing narrow-band orthogonal signal. And the author illustrated a feasible method to implement MIMO technology in ground penetrating radar.About the MIMO-GPR systems analysis, the paper firstly discusses the basic assumptions of MIMO technology applied in GPR system. The signal model of MIMO-GPR is presented based on orthogonal signal. The targets imaging method for MIMO-GPR in this paper is basing on the direction of arrival (DOA) estimation and reverse direction imaging. hree beam forming arithmetic were used to estimate targets DOA. They are least squares (LS, DAS), Capon, amplitude and phase estimation (APES). Through the numerical simulation of the three DOA estimations, the author anglicized the relation between the DOA estimated accuracy and antenna element spacing, number of antenna elements and signal to noise ratio of received signal. After analysis of this research, we can see that Capon beam forming arithmetic has a high capacity of DOA estimation, the amplitude and phase estimation (APES) and the least square (LS, DAS) followed. At the targets direction, the target angular resolution reflection coefficient form the least squares (LS, DAS) has little to do with the received signal noise ratio, but at non-targets direction, it has a greater side-lobe and shock. Capon and APES take a high relationship between the angle estimated resolution and signal to noise (SNR) in the targets direction. In non-targets direction, Capon and APES for DOA estimation accuracy will reduce with the signal to noise ratio (SNR) vary down, and high side-lobe phenomenon will appeared on both sides of the target direction by with similar oscillation likely least squares estimation result. From the research of target Imaging by reverse direction method, the conclusions can be drawn: the higher the resolution of targets arrival of angle DOA estimation, the smaller of side-lobes, and then the target imaging got accuracy and higher resolution; more testing points which have greater angle difference relative to the target point, and relative inclination of the measuring points relative to target can cover a greater different angle, the more accurate imaging of the target location estimation could be. Reverse direction targets imaging based on DOA estimation only needs a small number of antenna array measurement points along on the surface of the ground, but can be imaging the targets within a great area down side or not of the antenna array, which improve working efficiency. In general, the co-location MIMO radar (MIMO-GPR) system basing on orthogonal signal can effectively avoid the most problem of conventional ground penetrating radar system, and can achieve a good result of targets imaging.For system design problems of high complexity of MIMO-GPR who is Basing on orthogonal signals, the paper also proposed to use Switch-MIMO radar systems to achieve the basic MIMO-GPR system. The author proved that Switch-MIMO system and the basic MIMO system work with the equivalence signal model by analyzing their signal model, the research effectively reduces the complexity of MIMO system design and implementation.In the second part of this study, the author proposed a plane wave like signal MIMO radar system. MIMO radar systems basing on plane like wave signal could use conventional pulse radar signals as its detecting signal, and use the hardware system as its launching and receiving equipment. The MIMO ground penetrating radar system extended the antenna to be array and excitant-observe signal simultaneously on the antenna array. All transmitting antennas transmit nearly the same waveform signal, and they synthesized to be a transmitting beam which is approximate a plane wave. Before the study of antenna array multiple-input multiple-output (MIMO) observation method, the author firstly use finite difference time domain (FDTD) method to analyze the polarization, amplitude and phase characteristics of target echo signal under single input (SI) and multiple input (MI) cases. This research provides a basis theoretical for the array observation study.Through numerical simulation of various targets model, we can see that multiple-input multiple-output (MIMO) observed mode has a higher signal noise ratio of targets observing response, and may increases targets response dynamic range, relative to conventional common offset observation way by two antennaes of ground penetrating radar system. The multiple-input multiple-output observations method effectively reduces the effect of data profile by radar cross section (RCS) fluctuations, radiating and receiving pattern antenna varying with observed angle, so that the anomalies on observed profile will become more stable. When we integrate the multiple-input multiple-output (MIMO) array signal along the time axis, we can formed a the horizontal target abnormal line or horizontal abnormal distribution cross-section, which can be directly interpret the target horizontal distribution characters of subsurface targets under the antenna array. In addition, the process of integration of the received signal along time domain axis can be achieved in the analog signal processing stage, which means that you can reduce the design difficulty of A / D conversion, data acquisition and data storage module, which simplifies system designed and effective reduces system manufacturing costs. Compared to equivalent sampling mode of conventional pulse ground penetrating radar system, multi-input multi-output observations method effectively improve the measurement speed and signal stability, but also to speed up the interpretation of distribution patterns of the target. Finally, the using of narrow-band harmonic continuous wave signal as multiple-input and multiple output detecting signals was proposed. Although at this signal mode, targets detection results slightly lower than the pulse signal detection mode, but it effectively reduces the system’s hardware design and bandwidth requirements, and increases system transmitting power, increases signal dynamic range.Finally the author make a physical experiment to analyze the target response under multiple-input and multiple-output observation method, this work proved that multiple-input multiple-output observation can take the feasibility and advantages in the actual target detection. The author built a multi-channel, high accuracy experimental stepped-frequency radar system (SFGPR), which could simultaneous measure data in time-frequency domain. The establishment of high-precision stepped-frequency ground penetrating radar (SFGPR) basing on E5071B Vector Network Analyzer proves a good testing foundation for high-precision ground penetrating radar research and analyzing multi-parameter of magnetic wave propagation in subsurface. On the use of a range of different offsets of common-offset data to synthesize multiple-input multiple-output (MIMO) equivalent observations, we can see that MIMO observation methods can improve the target response signal to noise ratio, suppressing target multiple waves, thereby increasing the target explain precision. That is after analysis of a lot of model with simple and complex shape targets under multiple-input and multiple output observation.This thesis mainly analyzes the application idea of MIMO technology intruded to ground penetrating radar method. After demonstrating the application feasibility and superiority of MIMO technology applied in ground penetrating radar, we make a lot of numerical and physical simulation to analyze system characters. This paper focuses on two types of multiple-input multiple-output (MIMO) radar system. The first is a co-located ground penetrating radar system (MIMO-GPR) basing on orthogonal signal. The idea of its signal model is originally coming from MIMO technology application in the sounding radar research. This paper proposed a target inverse direction imaging for MIMO-GPR basing on DOA estimation. The other type of MIMO ground penetrating radar system proposed in the thesis is a plane wave like system using multiple-input and multiple-output(MIMO) observation method (Plane Wave Like MIMO GPR: PWL-MIMO GPR). Its signal waveform can be similar to conventional ground penetrating Radar systems. With analysis a lot of numerical result under building model with different shape targets, the author summed up the characteristics for different direction of polarization of target echo response, this research work provides a theoretical basis for target testing and interpretation under multiple-input and multiple-output observation method. Numerical simulation results demonstrate that the multiple-input multiple-output (MIMO) radar can not only improve the target detection capability and can effectively reduce the difficulty of system design and improve the system detection capability.更多还原