Direction of arrival estimation for uncorrelated and coherent signals with uniform linear array

A novel direction of arrival (DOA) estimation method is proposed with uniform linear array when uncorrelated and coherent signals coexist. The DOAs of uncorrelated signals are first estimated using a new method of modified total least squares estimation of signal parameters via rotation invariance techniques. Afterwards, the contributions of uncorrelated signals are eliminated and a new matrix is constructed based on the spatial difference matrix. Then, the coherent signals can be resolved by performing the forward spatial smoothing scheme on the new matrix. The proposed method can resolve more signals than the array elements with good performance. Simulation results demonstrate the effectiveness and efficiency of the proposed method.     

DOA estimation for non-Gaussian signals using fourth-order cumulants

A novel method based on fourth-order cumulants (FOC) is proposed for direction of arrival (DOA) estimation with uniform linear array (ULA). The method can be applied in the situation that the non-Gaussian independent and coherent signals coexist with unknown coloured Gaussian noise. The method comprises two steps: the first step is to estimate the independent signals, and then they are eliminated; the second step is to resolve the coherent signals with the reconstructed FOC matrix of coherent signals. The proposed method can also be extended to the scenario when independent, partially correlated and coherent signals coexist, and the number of signals resolved by our method can exceed the number of array elements. Simulation results demonstrate the effectiveness and efficiency of our method.     

Method of direction-of-arrival estimation for uncorrelated, partially correlated and coherent sources

A new direction-of-arrival estimation method is proposed when uncorrelated, partially correlated and coherent sources coexist. These sources are estimated at two different stages. The uncorrelated and partially correlated sources are first estimated using conventional subspace methods. By exploiting the property of oblique projection, the contributions of uncorrelated and partially correlated sources are then eliminated from the covariance matrix and only the coherent sources remain. Finally, the coherent sources are estimated by the technique of spatial smoothing. The new method need not estimate the partially correlated sources repeatedly but can resolve more sources than the array elements. Simulation results demonstrate the effectiveness and efficiency of the new method.     

双目标相干源窄带信号波达方向估计算法

利用水下目标发出的窄带信号,能够实现目标方向的精确估计。针对均匀线列阵接收水下目标窄带信号的特点,提出一种相干源波达方向(Direction of Arrival,DOA)估计的快速算法。通过精确估计目标信号对应的空间频率,直接计算出单目标或双目标信号的波达方向。对该算法进行了理论推导,通过仿真对算法进行验证,并对影响DOA估计精度的因素进行了分析。仿真结果表明,该算法具有较高的估计精度,不受信号时间长度的影响,算法计算量小,便于实时处理,具有较好的工程应用前景。     

基于矩阵平滑改进的稳健Capon波束形成的功率估计

回顾了对角加载波束形成的研究历程,提出了一种改进稳健Capon波束形成（RCB）算法。该方法在计算过程中,加入了矩阵平滑及将加权向量向信干子空间投影的步骤,提高了对相干信号的分辨能力以及低信噪比条件下信号功率的估计能力。仿真结果表明该方法能够分辨相干信号,功率估计较为准确,并且在输入信噪比为-10dB的时候,降低旁瓣15dB～20dB：当信源估计个数大于实际个数的时候依然有效。     

基于协方差矩阵拟合的宽带无网格波达角估计

宽带波达角(Direction of Arrival,DOA)估计是声呐系统阵列信号处理中一个重要的研究方向。文章提出了一种基于相干子空间的改进稀疏与参数方法(Coherent Signal-subspace based Modified Sparse and Parameter Approach,C-MSPA),以实现高精度和高空间方位分辨能力的宽带DOA估计。算法利用聚焦矩阵将各子带上的采样协方差矩阵投影至聚焦频率上。完成聚焦后,文章基于频率选择的范德蒙分解理论对协方差矩阵拟合准则进行改进,使重构的协方差矩阵中包含的DOA信息严格限制在聚焦区域内,最终对重构的协方差矩阵进行范德蒙分解,得到DOA估计值。所提出的算法无需选取正则参数,同时避免了基不匹配问题。仿真和湖上实测数据分析结果表明,所提出的方法实现了高空间方位分辨能力且提高了DOA估计精度。     

一种适用于低信噪比条件的DOA估计方法

波达方向（Direction of Arrival,DOA）估计技术是信号处理中一个非常活跃的研究领域。但是无论传统的波束形成技术还是现代谱估计技术均适应于高信噪比的环境,当信噪比较低时,这些方法的波达方向（DOA）估计性能急剧下降。根据信号在时间上的强相关性和噪声在时间的弱相关性,提出了一种协方差矩阵的重构方法,该方法能够明显地提高协方差矩阵的信噪比。将新的协方差矩阵应用到最小方差无畸变响应（Minimum Variance Distortionless Re-sponse,MVDR）算法进行DOA估计,改善了传统MVDR算法在低信噪比条件下的DOA估计性能。计算机仿真和定向实验均表明在信噪比较低的环境中可以进行高精度的DOA估计。     

利用时空平均法改善DOA估计性能

在信号子空间类DOA估计算法中 ,自相关矩阵估计的优劣决定了DOA估计的性能 ,但是常规自相关矩阵估计方法是对阵列输出信号做时间平均。然而对于均匀线列阵这一特殊阵型 ,其阵列输出自相关矩阵的具有To eplitz结构。根据均匀线列阵信号的时空平稳特性 ,本文提出一种时空平均的方法来改善自相关矩阵的估计质量 ,即对阵列的自相关矩阵做时间和空间两次平均 ,从而提高了DOA估计的性参 ;仿真结果表明 :在低信噪比或低快拍数的条件下能够明显提高DOA估计性能     

协方差矩阵的相关法Toeplitz改进算法

提出一种基于均匀线形阵列的相关Toeplitz矩阵构造方法,建立了两种Toeplitz矩阵的构造方式,结合子空间法形成一种相干信源方位估计的高分辨方法。具体将接收阵列各阵元与参考阵元输出信号做相关,获得一组相关向量,应用相关Toeplitz矩阵构造算法构造阵列输出的Toeplitz矩阵,从而得到去相干的相应接收阵列的协方差矩阵,最后再应用高分辨子空间估计方法完成相干信源方位估计。仿真结果表明：所采用的相关Toeplitz矩阵构造算法达到了很好的去相干效果,并将Toeplitz近似化方法改进成为无偏估计方法,显著提高了相干信源方位估计的精度,并使Toeplitz矩阵构造的计算量减少到原方法的1/M。     

Generalized Array Architecture with Multiple Sub-Arrays and Hole-Repair Algorithm for DOA Estimation

Arranging multiple identical sub-arrays in a special way can enhance degrees of freedom (DOFs) and obtain a hole-free difference co-array (DCA). In this paper, by adjusting the interval of adjacent sub-arrays, a kind of generalized array architecture with larger aperture is proposed. Although some holes may exist in the DCA of the proposed array, they are distributed uniformly. Utilizing the partial continuity of the DCA, an extended covariance matrix can be constructed. Singular value decomposition (SVD) is used to obtain an extended signal sub-space, by which the direction-of-arrival (DOA) estimation algorithm for quasi-stationary signals is given. In order to eliminating angle ambiguity caused by the holes of DCA, the estimation of signal parameters via rotational invariance techniques (ESPRIT) is used to construct a matrix that includes all angle information. Utilizing this matrix, a secondary extended signal sub-space can be obtained. This signal sub-space is corresponding to a hole-free DCA. Then, dealing with the further extended signal sub-space by multiple signal classification (MUSIC) algorithm, the unambiguous DOAs of all incident signals can be estimated. Some simulation results are shown to prove the improved performance of proposed generalized array architecture in DOA estimation and the effectiveness of corresponding hole-repair algorithm in eliminating angle ambiguity.     

CSM在基于Lamb波的结构损伤检测中的应用

在基于Lamb波的结构损伤检测中往往不可避免地遇到结构中多个损伤引起的反射波信号相干问题而对检测精度造成影响。考虑到阵列信号处理中的空间谱估计可以对信号源进行辨别和定位,尤其是在信号的波达方向(Direction of Arrival, DOA)估计上的优越性,将其应用于基于Lamb波的结构损伤检测中以获得准确的结构损伤信息。于是,该文采用空间谱估计中相干信号子空间方法(Coherent Signal Subspace Method, CSM)解决了相干信号问题。以一个损伤结构的仿真分析为例,验证了CSM方法在基于Lamb波的结构损伤检测中处理相干信号的可行性。结果表明,无论是结构损伤和边界反射引起的信号相干或两个结构损伤引起的信号相干,依然可以通过该方法获得准确的结构损伤信息。     

基于空间平滑处理的稳健宽带高分辨方位估计算法

实际系统中阵列误差导致的模型失配问题一直是高分辨方位估计技术走向实用化的一个瓶颈。相对于窄带信号,宽带信号阵列模型误差更加复杂和多样,通过建立宽带阵列信号误差模型,得出了基于空间平滑的稳健的宽带高分辨算法是一种抑制阵列误差的有效方法的结论,这种基于空间平滑的宽带高分辨算法,通过空间平滑后对各频率处不同子阵的互谱密度矩阵求和,一定程度上减弱了阵列误差的影响,具有较高的稳健性。本文结合消声水池试验,分析并验证了其效果。     

基于L型稀疏阵列的二维波达方向估计

针对传统的二维波达方向估计方法估计的信源数少、计算复杂度大的问题,提出了一种基于L型稀疏阵列的二维波达方向估计方法,利用安放在L型阵列上的互偶阵列来形成虚拟阵列,从而增加阵列自由度;通过将虚拟阵列分成若干个等间隔稀疏子阵列来压缩角度搜索区间,从而降低计算复杂度;通过MUSIC搜索得到谱峰,并利用最大似然准则筛选和匹配入射信号的俯仰角和方位角。该方法使可估计的信源数超过实际的物理阵元数,不仅提高估计精度且大大降低了计算复杂度。MATLAB仿真结果验证了该方法的有效性。     

声矢量阵相干源方位估计算法

基于声矢量阵的质点速度场平滑（ParticleVelocityFieldSmoothing,PVFS）算法是一种有效的解相干算法,但是当存在大量相干源时,该算法性能急剧下降甚至失效。在PVFS算法的基础上,提出了矩阵平方平滑（MatrixSquareSmoothing,MSS）-PVFS算法,该算法是对PVFS算法的改进,通过对PVFS算法构造的数据协方差矩阵进行平方、矩阵分块以及矩阵块间交叉相乘等数学运算,增强了PVFS算法解相干的能力,并大大增加了其分辨相干源的数目。计算机仿真结果表明,MSS—PVFS算法的效果与空间平滑（SpatialSmoothing,SS）-PVFS算法大致相同,但在低信噪比条件下,该方法具有更高的DOA估计精度。     

Subspace‐based techniques for 2‐d DOA estimation with uniform circular array under local scattering

Abstract

In mobile communication systems, local scatterers in the vicinity of the sources cause angular spreading of radiating signals as seen from a base station antenna array. Thus, the base station antenna array is typically situated on the roof of a high building away from potential multipath reflectors. The uniform circular array (UCA) geometry provides 360° azimuthal coverage and also provides information on source elevation angles. We consider the problem of two‐dimensional (azimuth and elevation) direction‐of‐arrival (DOA) estimation with UCA. In the multipath scenario, the base station antenna can receive many coherent signals that cause the array manifold to be different from the conventional array manifold model. Herein, parameters of the spatial signature in the presence of local scattering are presented which apply to UCA. Then, we present a fast searching technique to improve the efficiency of the MUSIC algorithm for two‐dimensional DOA estimation. The fast signal subspace‐based estimation method utilizes the ESPRIT algorithm and then adopts sequential one‐dimensional searching to save computational cost. Several simulation results are included for illustration and comparison.     

盲源分离用于DOA估计研究

方位估计是水下目标定位、跟踪的前提。提出了一种复数域奇异值分解盲源分离方法,估计出了阵列流形,并利用阵列流形表现出的相位关系完成了目标方位估计。通过仿真实验与MUSIC高分辨方位估计方法和相关文献的方法进行了对比,结果表明该方法能很好地实现目标方位的实时估计,且性能更佳。     

基于均匀圆阵的信号二维方向角和多普勒频率的盲估计

针对均匀圆形阵列 ,在时空旋转因子的基础上 ,推导出四阶累积量矩阵 ,采用波达方向矩阵法 ,实现了信号的方位角、俯仰角和多普勒频率的同时估计。此估计算法利用所有阵元信息估计信号波达方向 ,提高了方向估计精度 ,解决了“频率兼并”问题 ,使有效应用范围扩大。利用了四阶累积量的盲高斯性 ,有效地抑制了高斯白或有色噪声的影响 ,仿真结果表明了此算法的有效性     

基于空间平滑的多波束测深声呐相干分布源方位估计

在多波束测深声呐的工作环境中,若海底反向散射信号不满足点源假设,方位估计精度将严重下降,而基于分布源模型的方位估计算法可以适应这种环境。目前,大多数分布源算法要求分布源之间不相干,有人提出采用Toeplitz方法估计相干分布源,但该方法精度不高并且忽略了角度扩展参数。为解决多波束测深声呐相干分布源的方位估计问题,提出了基于空间平滑的广义MUSIC方法,公式推导证明了算法的有效性,通过计算机仿真给出算法方位估计的精度以及不同信噪比条件下的性能,最后采用多波束测深系统的实验数据对算法进行了验证。     

基于模型稀疏表示的单快拍波达方向鲁棒估计

基于传感器阵列输出模型的稀疏重构,研究了利用单快拍数据进行波达方向（Direction of Arrival,DOA）估计的问题。考虑到在实际应用中,目标信号个数远小于传感器阵元数,目标信号DOA相对于空间来说也是稀疏的,将传统的传感器阵列输出模型进行稀疏化表示,得到阵列输出数据的稀疏表示模型,研究了一种基于l₁-范数最小化的单快拍DOA估计算法（L1-Min）。该算法将稀疏参数求解问题转化为二阶锥规划（Second-Order Cone Programming,SOCP）问题的一般形式,并在二阶锥规划的框架下求解,同时分析了算法中正则化参数的选取依据。L1-Min算法对小样本、相干多径信号、目标信号角度间隔小等非理想条件都具有较好的鲁棒性。仿真实验验证了算法的有效性。     

基于插值矩阵技术与二维空间平滑法的近场相干声源识别研究

在研究插值矩阵技术与二维空间平滑法两种方法的基础上,将两者进行结合,提出了一种对近场相干声源进行识别的方法。首先通过插值矩阵技术把实际阵列接收到的近场信号转换为形式上的远场信号,然后通过二维空间平滑法求出阵列的平滑协方差矩阵,再由平滑协方差矩阵得到给定方向向量对应的信号功率,实现二维相干声源的识别。算例表明,该方法比传统的波束形成能更准确地定位出声源的位置。