【作者】 张翼；

【导师】 庄钊文； 黎湘；

【作者基本信息】 国防科学技术大学 ， 信息与通信工程， 2009， 博士

【摘要】 人体的探测在安全防护、灾害救援、军事医学等领域有着广泛需求,其中利用电磁波对人体微动的提取与辨识研究日益受到重视。本论文针对人体步态下的肢体摆动对电磁波的调制作用、人体微多普勒特征分析、步态参数的估计、姿态参数的估计、运动模式的辨识等问题开展研究。论文首先综述了人体微动雷达特征研究的发展现状,提出了需要研究的问题并对本论文研究内容进行了规划。针对人体微动对电磁波的调制作用问题,在基于类人型机器人和虚拟人研究的基础上,提出了适用于人体雷达特征研究的人体线结构模型,分析了各肢体的运动学轨迹;在此基础上推衍出了人体行走状态雷达回波的解析表达式;运用时频分析手段提取了行人雷达回波中的微多普勒特征,研究了人体运动参数与微多普勒特征的关系,特别对人体规范行走中关键步态的微多普勒特征进行了研究。针对估计人体行走参数的问题,深入研究了van Dorp P.提出的特征参数提取方法和人体步态参数估计方法,并应用于实测数据,提取了人体微动特征,估计了步态参数;研究了人体上肢姿态参数与多普勒特征的内在关系,提出了从行人上肢微多普勒特征中提取特征参数的方法,并用于估计上肢姿态参数;提出了基于函数逼近论原理的规范行走中人体下肢姿态参数的估计方法及实现途径。针对人体运动辨识问题,提出了基于微多普勒特征分析和步态参数估计的分辨方法。首先分析了人体跑步的时空特性,将之与人体规范行走的时空特性进行分析与比较,在基于规范行走与跑步状态下人体微多普勒特征分析的基础上,提出了可用于进行运动辨识的特征;在对大量实测数据分析与实验的基础上,提出了以支撑矢量机作为分类器的人体运动分辨方法;深入研究了Otero M.提出的人体行走时频图分析方法,提出了一种单人和多人行走的分辨方法,探索了人和动物的分辨问题。为了支撑论文上述研究工作,设计并研制了人体微动雷达测量系统,设计完成了相关数据测量实验并分析了结果。首先对雷达测量系统要求进行了剖析,设计了雷达系统的组成结构,着重分析了雷达回波采集系统的硬件设计与实现,而后提出了人体步态雷达测量的实验方案,完成了大量的实验数据采集,并对该系统存在的问题作出了说明,提出了下一步的改进目标。最后,对论文工作和研究方向的发展趋势、应用前景进行了总结,指出了需要进一步研究和解决的问题。更多还原

【Abstract】 Due to the complicated battlefield circumstance and the requirement of safety protection, disaster rescue, and military medical science, the study on the radar feature of human micro-motion is more and more highly regarded. This paper just studies the radar feature of human walking mode, the modulating function of limbs’swinging movement on the radar waves when walking, micro-Doppler feature analysis of human body, estimation of pace parameter, estimation of posture parameter, and identification of movement mode, etc.The paper first described the present situation of radar feature study of human micro-motion, laid out the content of this paper and raised the issues that need research.As for the modulating function of human micro-motion on radar wave, the paper raised human body linear structure model that is suitable for human body radar feature study based on the study of human-alike robot and virtual human, analyzed the kinematic track of each limb, deduced the radar echo of human walking and gave the resolution expression. Then, the paper extracted the micro-Doppler feature of walking human radar echo by applying time-frequency analysis means, studied the relationship between human motion parameter and micro-Doppler feature, and especially studied the micro-Doppler feature of key paces of human walking.As for extraction of human walking feature, the paper deeply studied the methods of feature parameter extraction and human pace parameter estimation raised by P. van Dorp, and applied them in field-measured data and thus extracted human micro-motion feature and estimated pace parameter; the paper also studied the inner relationship between human upper limb posture parameter and micro-Doppler feature,raised the method of extracting feature parameter from the micro-Doppler feature of the upper limbs of walking human and applied it to estimate upper limb posture parameter; then the paper raised the method and approach of estimating the posture parameter of lower limbs of walking human based on function approximation theory.As for human motion identification, the paper raised the distinguishing method based on micro-Doppler feature analysis and pace parameter estimation. The author firstly listed the time and space features of human running and compared them with those of normal human walking. Then, based on analysis of human micro-Doppler features during walking and during running, the paper raised the features that can be used in motion identification; based on analysis and experiment of plenty of field-measured data, the paper raised a human motion distinguishing method that applies the support vector machine as classifier. The paper also deeply studied the time-frequency chart analysis method for human walking raised by Michael Otero, raised a method for distinguishing single human walking and more-than-one-person walking, and explored the method of distinguishing human beings and animals.Based on the study on the basis of a great amount of field-measured data, the paper introduced the establishment of radar measurement system and the analysis of experiment result. Firstly, the author analyzed the design requirement of radar measurement system, designed the composition structure of radar system, and stressed the hardware design and realization of radar echo collecting system; then the author raised the experiment plan of human pace radar measurement experiment and collected plenty of experiment data. The author also listed the problems that exist in the system and raised the improvement target in future research.At last, the author summarized the developing trend and application prospect of the dissertation study and pointed out the problems that need further study and resolution.更多还原