数字波束形成的信号分析

【作者】 李鸣；

【导师】 牟善祥；

【作者基本信息】 南京理工大学 ， 信号与信息处理， 2003， 硕士

【摘要】 数字波束形成是未来移动通信系统中的研究热点，本文着重对数字波束形成中的信号分析研究。 首先结合具体的数字波束形成接收机系统，研究了阵列接收信号处理方法，分析波束形成实现过程。针对级联接收机系统给出性能参数的取值分析，并对信号处理过程进行仿真。 接下来对数字波束形成中的误差归纳总结，分别对误差进行推导，建立误差的模型，给出误差与数字波束形成性能之间的关系。 讨论循环平稳信号的特性，并以调频和调幅信号的循环平稳特性示之。然后对噪声的循环相关函数对循环平稳信号的性能进行分析。 介绍基于信号的循环平稳特性的CAB算法，并推导其迭代计算方法，对CAB算法的优劣之处进行分析。 最后提出多种基于CAB的自适应算法： 采用子阵平滑技术改善CAB性能。在CAB算法中引入传统波束形成中的信号子空间的方法，并对其进行改进，改进后的信号子空间对导引矢量进行修正，并有良好的波束保形性能。基于角度先验知识的投影变换对CAB的影响，减小噪声子空间的影响，并对投影算子给出修正的方法。对CAB算法引入遗忘因子，消弱取样数据对周期平稳相关矩阵误差敏感的不利影响，并对周期平稳相关矩阵误差进行分析。提出多级结构的CAB算法。更多还原

【Abstract】 Digital beamforming is the research focus of mobile communication system nowadays. This dissertation makes a study of the signal analysis in digital beamforming.With respect to the actual adaptive beamforming receiver system, the received array signal processing method is analyzed, then the principle of optimum weight value is deduced. After that, the processing of beamforming is pointed out. Aiming at the cascaded receiver system, we analyze the performance parameters and simulate the course.Then we conclude the errors in beamforming. Origin of the errors are deduced respectively and the model of error are established, and relation between them and performance of beamforming are presented.The cyclostationary of signal is introduced. AM and FM signals are demonstrated, then we study the influence of cyclostationary functions of noise to performance of beamforming. CAB algorithm and its operation is deduced, and its limitation is studied.At last, we propose some algorithms based on CAB algorithm.Then in terms of small snapshots, the performance of CAB algorithm is depressed. A method based on sub-array combined with correlation function is introduced to CAB algorithm. At a cost of aperture, the number of snapshots is equivalent to reduplication.With introducing the theory of eigen-space, ESB method in CAB algorithm is studied. The principle of the new method is analyzed.The influence caused by noise sub-space is analyzed and a new method is presented, which can realize reducing noise affect. Through the rough estimation of DOA, weight value is projected in signal plus interference sub-space in order to improve the performance.Forget factor is introduced to CAB algorithm. With analysis on cyclostationary frequency error, a method with forget factor is proposed to maintain the performance in the case of CFfi.Because of the great calculation of inverse correlation matrix in CAB algorithm, real-time control is difficult. We adopt multi-stages technique in CAB algorithm, the dimension of correlation matrix can decrease enormously.更多还原