【作者】 尤国红；

【导师】 邱天爽；

【作者基本信息】 大连理工大学 ， 信号与信息处理， 2013， 博士

【摘要】 波达方向估计和时间延迟估计是无线被动定位技术中的重要手段,作为现代信号处理中信号参数估计与提取的重要组成部分,被广泛应用在雷达、声纳、无线通信系统定位、无线电监测和军事等领域。近年来,循环平稳信号的波达方向估计得到了越来越多的重视,但目前的算法大多都是基于二阶循环统计量的,这些算法的性能在α稳定分布噪声环境下会出现严重的退化,为此本文重点研究了脉冲噪声环境下循环平稳信号的波达方向估计问题。另一方面,时间延迟估计发展至今虽然出现了很多性能优良的估计方法,但针对射频窄带信号的时延估计问题仍然是无线定位技术中的难点,并且中频频差的存在也是影响时延估计精度的主要因素,因此本文还对基于频差补偿的射频窄带信号的时延估计方法进行了深入的研究。本文主要创新工作如下：(1)针对均匀线阵下宽带循环平稳信号DOA估计困难问题,将分数低阶循环统计量与谱相关信号子空间拟合算法(SC-SSF)相结合,提出了基于分数低阶循环相关的SC-SSF算法,实现了脉冲噪声环境下宽带循环平稳信号的DOA估计。在此基础上,通过构造新的扩展数据模型,提出了一种扩展的分数低阶循环MUSIC算法,该算法具有较高的估计精度。同时,为了进一步提高算法的估计性能,将相位分数低阶矩与循环统计量相结合,定义了相位分数p阶循环相关的概念,进一步提出相位分数p阶循环MUSIC算法及其快速根值MUSIC算法,算法具有较高分辨估计性能。(2)针对α稳定分布环境下均匀圆阵的循环平稳信号的DOA估计问题,在分析与验证分数低阶循环相关的相移特性的基础上,提出了基于均匀圆阵的韧性分数低阶循环MUSIC算法,实现了α稳定分布噪声下方位角和俯仰角的同时估计。同时,本文还提出了一种基于圆阵的扩展循环MUSIC算法,该算法通过构造扩展的数据矩阵,充分利用循环相关和共轭循环相关的信息,突破了传统算法要求阵元数大于信号源数的局限性,提高了方位角与俯仰角的估计精度,是一种具有较高分辨的DOA估计算法。(3)针对中频偏差对时延估计的影响,提出了两种基于频差补偿的时延估计方法。针对采用相位谱直接对窄带射频信号进行时延估计存在的局限性,利用平方倍频的方法估计中频偏差,采用CZT变换增加时延估计的有效频域点数,提出了基于频差补偿的相位谱时延估计方法,提高了时延估计的精度。为了解决相近信号源的时延估计问题,将频差存在时的时延估计模型与MUSIC算法估计模型相比拟,提出了一种基于频差补偿的MUSIC频率检测时延估计方法,该方法充分利用了MUSIC算法的具有较高分辨性能,提高了时延估计的性能。更多还原

【Abstract】 Directions-of-arrival (DOA) estimation and time delay estimation are important means of wireless passive position technology. They are also important parts of parameters estimation and extraction in modern signal processing and have been widely used in radar, sonar, wireless communication positioning system, radio monitoring, military and other fields. In recent years, DOA estimation of cyclostationary signal has been obtained more and more attention, but most of the current algorithms are based on the second-order cyclic statistics. The performance of these algorithms will degenerate seriously in α-stable distribution noise environments. This desertation mainly studies DOA estimation of cyclostationary signal under impulsive noise environment. On the other hand, as the time delay estimation develops up to now, many excellent methods of the time delay estimation have been proposed. But the problem of narrowband RF signal delay estimation is still a difficult work in wireless location technology, and the existion of intermediate frequency difference (IFD) will affect the accuracy of time delay estimation. So this desertation mainly study the new methods of the time delay estimation based on frequency difference compensation (FDC). The main researches and conclusions are listed as follows:(1) Since DOA estimation of broadband cyclostationary signal based on ULA is a difficult work, by combining the fractional lower order cyclic statistics with SC-SSF algorithm, a spectral correlation signal subspace fitting algorithm (SC-SSF) based on the fractional lower order cyclic correlation is proposed. Then the DOA estimation of broadband cyclostationarity signal in impulsive noise is realized. Furthermore, an extended fractional lower order cyclic MUSIC algorithm is proposed by constructing a new extended data model. The algorithm has higher estimation accuracy. Meanwhile, in order to improve the estimation performance, a novel phase fraction pth-order cyclic correlation by fusing the phase fractional lower order statistics and cyclic statistics is defined. Then a novel extended phase fraction pth-order cyclic MUSIC algorithm (EX-POC-MUSIC) and phase fraction pth-order cyclic Root-MUSIC (EX-POC-RMUSIC) algorithm are proposed. Both algorithms have high-resolution estimation performance.(2) In order to solve the DOA estimation problem of cyclostationary signal based on uniform circular array (UCA) under α-stable distribution environment, after analysing and verifing the phase shift characteristics of the fractional lower order cyclic correlation, we propose a fractional lower order cyclic MUSIC algorithm based on UCA, which can estimate both azimuth and elevation angles accurately even in impulsive noise. Meanwhile, a novel extended cyclic MUSIC algorithm based on UCA is developed in this paper. The algorithm exploits simultaneously the information contained in both cyclic correlation matrix and cyclic conjugate correlation matrix of the extended array data vector, and breaks through the limitation of the classic MUSIC algorithm in which the number of source signal can not be greater than the number of element in the receiving array. The algorithm improves the estimation accuracy of the azimuth and elevation angle, which is a high-resolution DOA estimation algorithm.(3) As the intermediate frequency difference (IFD) impacts the delay estimation, two time delay estimation (TDE) algorithms based on the frequency difference compensation (FDC) are proposed. Since time-delay estimation using the phase spectrum for narrowband radio frequency (RF) signal exists limitations, frequency doubling methods is used to estimate the intermediate frequencier (IF) of both receicved signals, and chirp Z transform (CZT) is applied to increasing the available frequency bins within a limited bandwidth. Then, a phase spectrum TDE method based on FDC is proposed, which can improve the accuracy of TDE. In order to solve the time delay estimation problem of the closely spaced signals, the model of TDE with IFD and the model of MUSIC algorithm are compared, and then an effective time delay estimation method based on FDC for narrowband RF signals is proposed. This method exploits the high-resolution performance of MUSIC algorithm, which improves the performance of TDE.更多还原