【作者】 贺峰；

【导师】 周智敏； 黄晓涛；

【作者基本信息】 国防科学技术大学 ， 信息与通信工程， 2011， 博士

【摘要】 人体目标探测在反恐、灾后搜救等场合具有重要应用,基于雷达的人体目标探测技术是近年来人体目标探测的研究热点和前沿课题。本文的研究围绕运动人体目标雷达探测理论与技术展开,重点研究运动人体目标雷达回波模型、宽带/超宽带雷达运动人体目标特征提取、超宽带穿墙雷达的干扰与杂波抑制、实测数据处理与分析等关键理论与技术。首先研究了运动人体目标雷达回波建模。将人体目标运动简化为典型运动形式的复合运动,基于线性刚体结构建立和完善了原地踏步状态和行进状态人体目标模型,上述模型具有与真实人体目标运动状态吻合度高、参数调整方便、思路简单等特点,可推广应用于多种雷达信号形式。基于连续波信号得到了运动人体目标雷达回波,从理论上分析了运动人体目标的瞬时频率特征和瞬时运动特征,利用仿真工具分析了不同探测条件下雷达回波特性,为分析运动人体目标特征提取奠定了理论基础。其次研究了宽带/超宽带雷达运动人体目标特征提取。以摆动目标为简化模型,研究了原地踏步目标特征提取。基于频率步进连续波信号得到摆动目标宽带/超宽带雷达回波,提出了利用脉冲对消、高通滤波或自适应时频表示技术进行微动信号分离的处理方法;根据摆动目标时间-距离像和时间-频率像变化规律,提出了两种基于时间-距离像的摆动目标特征提取改进算法,提高了摆动周期的估计精度,提出了基于幅度阈值检测从目标时间-频率像直接提取摆动特征的算法,并研究了时频变换方法对特征提取性能的影响;本文研究的摆动目标特征提取方法同样适用于旋转等周期性运动目标。以平动和摆动构成的复杂运动目标为简化模型,研究了行进运动目标特征提取。基于频率步进连续波信号得到复杂运动目标宽带/超宽带雷达回波,根据复杂运动目标时间-距离像和时间-频率像变化规律,提出了基于距离和差值序列的复杂运动目标特征提取新方法,该方法解决了不同摆动幅度的估计问题,算法实现简单、参数提取精度高;提出了基于运动补偿的复杂运动目标特征提取新方法,该方法解决了多散射点目标趋势项的估计与处理问题,算法实时性好、参数估计精度高;本文研究的复杂运动目标特征提取方法可推广应用于由一般平动和周期性运动构成的复杂运动目标。最后基于实测数据研究了超宽带穿墙雷达人体目标检测与特征提取等问题。研究了超宽带冲激信号雷达回波特性,介绍了所研制的超宽带穿墙雷达的结构和基本组成。干扰与杂波抑制是人体目标检测的关键,针对耦合的时变性,提出了一种基于频域幅度时延对齐的耦合抑制新算法,在一定程度上解决了时变耦合抑制问题;针对杂波的复杂性和目标运动状态的不规律性,提出了基于自适应背景相消的杂波抑制新算法,该算法对杂波环境和目标运动类型适应性强,从理论上将脉冲对消法、背景相消法等常规杂波抑制方法统一于自适应背景相消算法,完善了穿墙雷达杂波抑制理论和方法。开展了不同探测条件下多种运动类型的人体目标穿墙探测实验,利用实测数据处理结果验证了干扰与杂波抑制算法以及运动人体目标特征提取算法的有效性和超宽带穿墙雷达对人体目标的优良探测性能。指出了系统存在的主要问题,为系统的进一步改进提供了参考。论文提出的运动人体目标雷达回波模型、宽带/超宽带雷达运动人体目标特征提取方法和超宽带穿墙雷达系统中关键信号处理算法将丰富人体目标雷达探测理论与特征提取技术,对穿墙雷达的研制具有较强的理论意义与应用价值。更多还原

【Abstract】 Detection of human target is an important issue in antiterrorism and search-and-rescue situations after emergencies, and detection technique of human target based on radar has been a research hotspot for a few years. In this paper, scientific researches have been made on radar detection theory and techniques of locomotor human targets. The main attention have been paid to key theory and techniques of radar echo model and wide band (WB) or ultra-wide band (UWB) radar signature extraction of locomotor human targets, interference and clutter suppression algorithms of UWB through the wall radar (TWR), and real data processing and analysis of experimental radar system.Firstly, radar echo model of locomotor human target is analyzed. The movement of human target is abstracted as composite motions composed of typical movements, and the model of human marking time and the model of human walking are established and improved based on linear-rigid structure, respectively. The characteristic of models discussed in this paper are that models fit well with real human’s movement, parameters of models can be adjusted conveniently, and the process of modeling is simple. Therefore, these two models can be applied to various radar signals. The radar echoes of locomotor human targets are acquired under the condition of continuous wave (CW) signal, instantaneous Doppler signatures and instantaneous movement signatures of targets are analyzed in theory, and the characteristic of radar echoes under different detection conditions are discussed by simulations in order to establish the theoretical foundation of signature extraction of locomotor human targets.Secondly, WB/UWB radar signature extraction of locomotor human targets is analyzed. First of all, signature extraction of human marking time is discussed based on the simplified model of vibrating targets. WB/UWB radar echoes of vibrating target are obtained based on stepped frequency continuous wave (SFCW) signal, and a micro-motion signal separation method is brought forward with pulse canceller, high-pass filter (notch filter), or adaptive time-frequency representation(ATFR)algorithm. Then after the periodic laws of time-range profile and time-frequency profile of the target being analyzed, two improved micro-motion signature extraction algorithms based on time-range profile are carried out to improve the precision of the period estimation of vibrating target, a micro-motion signature extraction algorithm based on amplitude threshold detection is carried out to extract signatures directly from time-frequency profile, and the influence of time-frequency transforms onto the micro-motion signature extraction algorithm based on amplitude threshold detection is investigated. Studies indicate that micro-motion signature extraction algorithms put forward in this paper can be applied to other periodic moving targets such as rotary targets.Signature extraction of pedestrian is discussed based on the simplified model of a complex motion with translation and vibration. WB/UWB radar echoes of a target with complex motion are obtained based on SFCW signal first, and then after the characteristic of the target’s time-range profile being analyzed, a novel signature extraction method based on summation and difference sequence of ranges is proposed. The method can be utilized to deal with the problem of parameter estimation of the target with different vibrating amplitudes, and has specialties of easy implementation and high resolution of estimation results. A novel signature extraction method based on motion compensation is carried out after the characteristic of the target’s time-frequency profile being analyzed. The method can be utilized to deal with the problem of trend extraction and processing of multi-scatterers, and has specialties of good real time processing characteristic and high resolution of estimation results. Studies indicate that signature extraction methods discussed in this paper can be extended to analyze other targets with complex motions composed of general translation and other periodic motions.Lastly, human target detection and signature extraction based on real experimental data of UWB-TWR are investigated. The characteristic of UWB impulse radar echoes is discussed first, and then the structure and basic configurations of the UWB-TWR developed by NUDT are presented. Interference suppression and clutter suppression are the key factors to human target detection, and a novel coupling suppression algorithm based on the compensation of amplitude and delay within frequency domain is put forward, which can partly deal with the problem of time-variant coupling suppression. A novel clutter suppression algorithm based on adaptive background subtraction is brought forward in order to reduce clutter, and the algorithm can be applied to complex clutter environments and multiple movements of targets. In theory, the clutter suppression algorithm discussed in this paper can be regarded as common form for conventional clutter suppression algorithms such as pulse canceller, and background subtraction, and enriches the theory and methods of through the wall radar clutter suppression. TWR experiments are performed under different conditions of detection and movements. After data processing, experimental results validate interference and clutter suppression algorithms, signature extraction algorithms of locomotor human targets, and the good detection performance of the radar system. Some important issues are pointed out in order to improve the system further at last.Radar echo model of locomotor human targets, WB/UWB radar signature extraction of locomotor human targets, and key signal proceesing algorithms of UWB-TWR discussed in this paper can enrich the radar detection theory and signature techniques of human target, and be much valuable in theory and in application for development of TWR.更多还原