【作者】 杨勇；

【导师】 袁乃昌；

【作者基本信息】 国防科学技术大学 ， 电子科学与技术， 2011， 博士

【摘要】 靶场破片初速和速度衰减系数的测定是战斗部静爆试验中比较重要及复杂的项目,本文针对战斗部破片速度参数的测量问题,提出采用毫米波阵列雷达实现对位于阵列近场的破片目标速度参数的测量与估计。从空间谱估计技术、阵列的优化设置、阵列模型误差的校正、破片目标速度参数估计算法以及毫米波破片速度测量系统实现等五个方面开展了研究。论文首先阐述了研究背景和意义,回顾了靶场测速系统、毫米波连续波雷达和空间谱估计技术的发展现状,详细地比较和分析了目前国内外靶场破片测速方法的优缺点,指出了其在测量破片速度参数中存在的问题。推导了远场和近场窄带信号数学模型,介绍了两种子空间类空间谱估计算法、最大似然估计算法以及空间谱估计的Cramér-Rao界(CRB)。针对阵列的优化设置问题,提出了一种以提高空间谱估计性能为目标的阵列优化设置方法,与传统的以峰值旁瓣电平为目标的优化方法相比,提高了优化阵列的空间谱估计性能。通过计算机仿真对该方法进行了验证。针对阵列模型误差的校正问题,提出了一种阵列位置与通道不一致误差的校正算法。为忽略阵元间的互耦效应,简化阵列误差的校正,本文特意使用阵元间距较大的阵列。文中根据通道相位误差较大和较小两种情况分别给出了对应的校正算法。通道相位误差较大时,将阵列向两侧旋转相同的角度,可计算获得辅助源的方位,从而可估计阵元位置误差及通道幅相误差。而通道相位误差较小时,直接以未校正的阵列垂线附近辅助源的方位估计值作为其真实值,估计阵元位置误差和通道幅相误差。对两种情况分别进行了仿真。提出了基于毫米波阵列雷达的破片速度参数估计算法。针对各个破片目标相对于天线阵元多普勒频率的不同,提出按多普勒频率分离出各个目标对应的回波信号,然后分别估计各个目标的位置,由目标相对于每个接收阵元的多普勒频率和目标的位置估计出目标的速度,最后由不同时刻目标的速度得到目标的初速和速度衰减系数的估计值。介绍了几种用于频率估计的算法,并提出了相应的目标信号分离方法。文中推导了频率估计以及目标位置估计的CRB,并分别对频率估计和位置估计性能与CRB进行仿真比较,根据比较结果选择ML算法作为最终的频率估计算法和信号分离算法。最后对初速及速度衰减系数估计进行了仿真,仿真结果表明所提算法能够准确地估计目标速度参数。设计了毫米波破片速度参数测量系统。说明了测量系统的原理、系统的结构以及各个部分的原理。使用回波信号源对该系统进行了暗室测试实验,包括静止情况下对信源位置进行估计以及运动情况下对信源不同时刻的速度进行估计,获得了满意的实验结果。更多还原

【Abstract】 The measurement of start velocity and coefficient of velocity attenuation of fragments from static explosion test of fragment warhead is an important and complicated test item at shooting range. To realize the velocity measurement of shooting range fragments, a methodology using millimeter-wave array radar to obtain the velocity parameter measurement and estimation of near-field moving fragments is presented in this dissertation. Five main subjects have been researched, which are the spatial specrum estimation techniques, the optimization of linear array, array manifold calibration, the velocity parameter estimation algorithm of moving fragments and the realization of the millimeter-wave velocity measurement system for shooting range fragments.Firstly, the research background and significance is introduced, and the shooting range velocity measuring system, the development of millimeter-wave continuous wave radar and spatial specrum estimation techniques are reviewed. The advantages and disadvantages of the common shooting range velocity measurement methods at home and abroad are compared and analyzed in detail. The problems of these methods in measuring velocity parameters of the fragments are pointed out.The mathematical models of far-field and near-field narrowband signal are deduced. Two subspace-based spatial spectrum estimation methods, maximum likelihood method and the Cramér-Rao bound for spatial spectrum are introduced.In order to solve the optimization of linear array, a novel array optimization method is presented, which takes the improvement of the direction-of-arrival angle(DOA) estimation performance as the optimization goal. Compared with the conventional method based on the peak side lobe level (PSLL), the method presented performs more efficiently. Computer simulations verify this conclusion.The calibration methods for unequal gain and phase responses, and the small errors in element positions are researched. In order to neglect the mutual coupling and simplify the array model calibration, the physical separations of the array elements in use exceed a few wavelengths designedly. Two calibration methods are presented, which calibrate the array with large channel phase errors and small channel phase errors, respectively. When the channel phase errors are large, the DOAs of the pilot sources are calculated through rotating the array twice with the same angle, and the gain/phase perturbation matrix and element position errors can be estimated. When the channel phase errors are small, the uncalibrated DOA of the pilot source, which is located at the orthogonal direction of the array boresight, is taken for the actual DOA. Then the array model errors are estimated.A fragment velocity parameter estimation algorithm is presented, which is based on the millimeter-wave array radar. According to the different Doppler frequencies of moving fragments, the echo signals of the fragments are separated. Then the location of these fragments are estimated, respectively. Using the Doppler frequencies of each fragement relative to the array elements, the fragement velocity estimation can be obtained. Finally, the start velocity and coefficient of velocity attenuation of fragment are estimated according to the velocities of different times. Several frequency estimation algorithms along with signal separation methods are introduced. The Cramér-Rao bounds of frequency estimation and fragement location estimation are also deduced. According to the simulation results, use ML method to estimate the frequency and separate the echo signals finally. The simulation results of start velocity and coefficient of velocity attenuation estimations prove the validity of the proposed algorithm.A millimeter-wave fragment velocity parameter measurement system is designed. The measurement principle of this system is introduced. The system architecture and design of the components are also described. Finally, satisfied results are obtained through the darkroom test experiments, including the location estimation of fixed sources and the velocity estimation of moving source.更多还原