基于Khatri-Rao子空间和传播算子的宽带声源波达方向估计算法

本文提出基于Khatri-Rao子空间和传播算子的宽带声源波达方向估计算法。该算法将声源不同频率处的协方差矩阵变换重排为一个高维矩阵,然后利用传播算子方法估计宽带声源波达方向。该算法计算复杂度介于聚焦Khatri-Rao子空间和相干子空间算法之间。仿真和实验结果表明,该算法在降低计算量的同时,估计误差与聚焦Khatri-Rao子空间算法相近,远小于相干子空间算法。     

近似窄带假设下的最小方差无失真响应波束形成

最小方差无失真响应波束形成算法在应用于语音等宽带信号时,依赖窄带假设可以在频域各个子带分别进行滤波。窄带假设下语音信号协方差矩阵是秩-1矩阵,而实际中窄带信号模型只是实际信号模型的一种近似,同时由于存在统计量估计误差,估计的语音信号协方差矩阵的秩一般大于1。提出利用语音协方差矩阵和噪声协方差矩阵的广义主特征向量来估计相对传递函数,用于重构语音信号协方差矩阵为秩-1矩阵。在REVERB数据集以及CHiME-4数据集上进行实验验证,最小方差无失真响应波束形成算法经过语音协方差矩阵低秩近似后,对估计误差的鲁棒性提高,输出信噪比分别提升平均0.8 dB和1.4 dB,同时提升了语音识别准确率。     

浅海波导环境失配下的协方差矩阵拟合稳健匹配场声源功率估计

实际浅海波导中环境噪声为相干噪声,最小方差匹配场声源功率估计方法能在相干噪声背景下准确估计声源辐射功率,但该方法受环境不确定性影响较大;此外,由于最小方差匹配场声源功率估计方法使用信号幅度作为中间量估计声源功率,信号幅度估计误差会二次放大并传递到声源功率估计结果中。本文提出一种协方差矩阵拟合稳健最小方差匹配场声源功率估计方法,该方法引入信道传递函数不确定集,结合协方差矩阵拟合思想将声源功率估计问题建模为在信道传递函数不确定集约束下对函数取极值的问题,使用Lagrange乘子法求解该问题得到信道传递函数估计值和声源辐射功率估计值。环境失配影响声源辐射功率估计性能的根本原因在于信道传递函数偏差较大,协方差矩阵拟合稳健匹配场声源功率估计方法有效减小了环境失配时信道传递函数的偏差,从而显著提升环境失配稳健性。此外,该方法使用权值直接估计声源功率,无需使用信号幅度作为中间量,避免了估计误差的传递。仿真验证了协方差矩阵拟合稳健匹配场声源功率估计方法的环境失配稳健性。     

扩散噪声下协方差矩阵重构的语音分离与降噪

针对传统多通道语音分离算法在扩散噪声下性能下降的问题,提出了一种用于语音分离及降噪的空间协方差模型及参数估计方法。该方法将扩散噪声视为独立声源,利用由导向矢量重构的空间协方差矩阵建模目标声源的空间特性,并通过空间协方差分析方法估计用于语音分离的多通道维纳滤波器。同时,还提出了一种联合该方法的后置滤波器参数框架,为输出信号降噪和失真的折中提供了更多选择。在扩散噪声下的单目标和多目标实验中,所提方法的语音提取和分离性能都优于对比算法,联合参数的后置滤波器可提供更为符合人们要求的降噪语音,验证了所提模型与参数估计方法的有效性。     

密布式多输入多输出声呐阵列目标波达方向估计

针对低信噪比条件下多输入多输出声呐受对称噪声分量影响导致测向性能降低的情况,提出了一种基于协方差矩阵重构方法的波达方向估计算法。首先,将噪声场分为对称噪声和非对称噪声两部分,利用协方差矩阵虚部与对称信号无关的性质,去掉协方差矩阵的实部来降低对称噪声对目标波达方向估计精度的影响,采用降维转换方法和矩阵虚部置换原理重构协方差矩阵的实部,避免了双频谱的干扰。然后利用Toeplitz方法对重构的协方差矩阵进行解相干修正,通过奇异值分解获得噪声子空间,最后对目标的波达方向进行估计,可实现微弱信号的准确测向。理论分析和实验结果表明,该方法明显抑制了对称噪声,提高了目标的波达方向估计性能,具有运算速度快、自由度高和目标分辨力强的特点。     

采用四阶累积量多重矩阵重构的低信噪比相干声源波达方向估计

针对相干声源子空间能量扩散且协方差矩阵欠秩难以有效估计波达方向(DOA)的问题,提出了一种采用高阶矩阵变换的估计方法-四阶累积量多重矩阵重构(FOC-MMR)。该方法首先对阵列声压数据分帧进行短时傅里叶变换,然后对四阶累积量扩展的高阶协方差矩阵进行奇异值分解(SVD)得到高阶噪声特征向量,保证该噪声特征向量与扩展后的高阶阵列流形矢量正交匹配,最终实现相干信号的DOA估计。相干单频矩形脉冲信号仿真结果表明,将FOC-MMR方法应用于均匀线阵(ULA, M=4),在信噪比SNR≥-15 dB时,相干信号(θ₁=-20°和θ₂=20°)的均方根误差保持在1.5°以内;在SNR=10 dB时,可正确分辨的两相干信号方位间隔Δθ可以低至5°。在相干脉冲声源实验中,通过混入SNR=5 dB高斯白噪声,验证了FOC-MMR算法在应用于由多个ULA组成的矩形面阵时,其分辨邻近声源和抑制高斯噪声的能力较高。FOC-MMR算法通过对声压阵列数据扩展得到满秩的高阶协方差矩阵,不仅解决了由信号相干造成的噪声和信号特征向量之间能量扩散的问题,还实现了以较高的测向精度和...     

噪声情况下采用稀疏非负矩阵分解与深度吸引子网络的人声分离算法

为实现噪声情况下的人声分离,提出了一种采用稀疏非负矩阵分解与深度吸引子网络的单通道人声分离算法。首先,通过训练得到人声与噪声的字典矩阵,将其作为先验信息从带噪混合语音中分离出人声与噪声的系数矩阵;然后,根据人声系数矩阵中不同的声源成分在嵌入空间中的相似性不同,使用深度吸引子网络将其分离为各声源语音的系数矩阵;最后,使用分离得到的各语音系数矩阵与人声的字典矩阵重构干净的分离语音。在不同噪声情况下的实验结果表明,本文算法能够在抑制背景噪声的同时提高分离语音的整体质量,优于结合声噪人声分离模型的对比算法。      

车载场景结合盲源分离与多说话人状态判决的语音抽取

在车载分布式传声器阵列场景中,结合盲源分离TRINICON (Triple-N ICA for convolutive mixtures)算法与多说话人状态判决实现期望语音抽取。根据分布式传声器阵列与声源的相对位置关系,设计特定的盲源分离初始化条件以保证输出通道与声源的映射关系;根据分布式传声器阵列的频响特点,设计特征矢量来进行多说话人判决,并将判决结果引入TRINICON算法参数迭代过程。在使用实际车载录音数据的仿真评测中,所提方法在不同信噪比下有较高的鲁棒性,可有效提升TRINICON算法的收敛速度和语音信号的信扰比,且可以确保准确的通道映射。评测结果表明该方法可以在车载场景中有效抽取出期望语音,为车载复杂场景下的声信息提取提供了一种可靠且收敛快速的解决方法。     

基于声源方位信息和非线性时频掩蔽的语音盲提取算法

针对欠定卷积混合的语音信号模型,提出一种基于声源方位信息和非线性时频掩蔽的语音盲提取算法。首先对低频段混合语音信号进行时频分析估计瞬时相对时延(ITD)并采用势函数聚类分析方法估计出声源个数及其ITD,接着锁定目标提取准确的目标语音方位信息,最后利用独立语音在时频域上的近似W一分离正交性,采用非线性时频掩蔽的方法提取目标语音。仿真实验表明,该方法能锁定任意感兴趣目标方位,能有效提取目标语音,文中实验条件下信噪比增益平均达9.5 dB。     

基于干扰协方差矩阵重构的恒定束宽鲁棒自适应波束形成

针对宽带波束形成中的恒定束宽波束响应优化设计问题与鲁棒性问题展开研究。首先,提出一种基于相位补偿的恒定束宽全局优化设计方法,通过对阵列流形向量进行相位补偿来设计恒定束宽波束,与现有的一些方法相比,该方法不仅能获得全局最优解,而且物理实现简单。同时,还提出一种基于协方差矩阵重构的鲁棒自适应宽带波束形成算法。该算法采用Capon估计器估计样本数据的空间一频率谱密度函数,然后对期望信号波达方向之外的角度区间进行积分来重构干扰加噪声协方差矩阵,最后利用重构的协方差矩阵设计自适应波束形成器权系数。该波束形成器设计问题被表述成凸优化问题求解。仿真结果表明,在整个输入信噪比范围内,该算法几乎都能获得接近理想值的输出信干噪比。     

联合深度神经网络和凸优化的单通道语音增强算法

噪声估计的准确性直接影响语音增强算法的好坏,为提升当前语音增强算法的噪声抑制效果,有效求解无约束优化问题,提出一种联合深度神经网络(DNN)和凸优化的时频掩蔽优化算法进行单通道语音增强。首先,提取带噪语音的能量谱作为DNN的输入特征;接着,将噪声与带噪语音的频带内互相关系数(ICC Factor)作为DNN的训练目标;然后,利用DNN模型得到的互相关系数构造凸优化的目标函数;最后,联合DNN和凸优化,利用新混合共轭梯度法迭代处理初始掩蔽,通过新的掩蔽合成增强语音。仿真实验表明,在不同背景噪声的低信噪比下,相比改进前,新的掩蔽使增强语音获得了更好的对数谱距离(LSD)、主观语音质量(PESQ)、短时客观可懂度(STOI)和分段信噪比(segSNR)指标,提升了语音的整体质量并且可以有效抑制噪声。     

Weak signal direction of arrival estimation for colocated multiple-input multiple-output sonar array

In order to suppress the influence of symmetrical noise component on multiple-input multiple-output(MIMO)sonar's direction of arrival(DOA)estimation under the condition of low signal-to-noise ratio,we propose a DOA estimation algorithm based on covariance matrix reconstruction method.Firstly,the noise field can be decomposed into symmetrical noise field and asymmetrical noise field.We utilize symmetry property of colored noise matrix and the feature that the imaginary part of covariance matrix has no relation with the symmetry noise to remove the real part of covariance matrix.This operation helps to suppress the influence of colored noise on DOA estimation accuracy.Based on the principle of the imaginary matrix part displacement and the dimension reduction transformation method,the real part of covariance matrix is reconstructed,which helps to suppress the bilateral spectrum interference.Thereafter,Toeplitz method is applied for the covariance matrix decorrelation amendment,and a noise subspace is formed by singular value decomposition(SVD).Finally,we can estimate the DOA of target signals.Both theoretical analysis results and numerical simulation results verify the symmetrical noise suppression performance of this algorithm,and the estimation performance of target azimuth is improved obviously.This method has the characteristics of lower operational complexity,higher degrees of freedom and stronger target resolution.     

面向自定义语音唤醒的关键词相关的单通道语音增强

提出一种面向自定义语音唤醒的单通道语音增强方法。该方法预先将关键词音素信息存入文本编码矩阵,并在常规语音增强模型基础上添加一个基于注意力机制的音素偏置模块。该模块利用语音增强模型中间特征从文本编码矩阵中获取当前帧的音素信息,并将其融入语音增强模型的后续计算中,从而提升语音增强模型对关键词相关音素的增强效果。在不同噪声环境下的实验结果表明,该方法可以更有效地抑制关键词部分噪声。同时所提出方法对比常规语音增强方法与其他文本相关语音增强方法,在自定义语音唤醒性能上可以分别获得14.3%和7.6%的相对提升。     

Clustering gene expression data analysis using an improved EM algorithm based on multivariate elliptical contoured mixture models

Clustering gene expression data is an important research topic in bioinformatics because knowing which genes act similarly can lead to the discovery of important biological information. Many clustering algorithms have been used in the field of gene clustering. The multivariate Gaussian mixture distribution function was frequently used as the component of the finite mixture model for clustering, however the clustering cannot be restricted to the normal distribution in the real dataset. In order to make the cluster algorithm strong adaptability, this paper proposes a new scheme for clustering gene expression data based on the multivariate elliptical contoured mixture models (MECMMs). To solve the problem of over-reliance on the initialization, we propose an improved expectation maximization (EM) algorithm by adding and deleting initial value for the classical EM algorithm, and the number of clusters can be treated as a known parameter and inferred with the QAIC criterion. The improved EM algorithm based on the MECMMs is tested and compared with some other clustering algorithms, the performance of our clustering algorithm has been extensively compared over several simulated and real gene expression datasets. Our results indicated that improved EM clustering algorithm is superior to the classical EM algorithm and the support vector machines (SVMs) algorithm, and can be widely used for gene clustering.     

基于矢量水听器阵列空域预滤波处理的多重信号分类算法研究

矢量水听器能同时拾取声压和振速信息,在相同的信噪比、阵元数及阵列孔径下,矢量阵定向性能优于声压阵列。目前,以多重信号分类算法(Multiple signal classification,MUSIC)为代表的高分辨定向算法已经广泛应用于矢量水听器阵列中。但是随着信噪比降低、信号源方位间隔减小,传统MUSIC算法定向精度及分辨概率显著下降。本文采用最小二乘法设计适用于矢量水听器水平阵列的矩阵空域滤波器,用于阵列数据的空间滤波预处理,可以对阻带扇面噪声进行有效抑制。由滤波后的数据协方差矩阵可以得到新的噪声子空间,在传统MUSIC算法基础上修正通带扇面内阵列流型的畸变后即可得到滤波后MUSIC算法的方位谱。仿真结果表明,当信噪比较低时,改进算法有效提高了通带扇面内目标方位分辨性能。最后本文对四基元矢量水平阵列海试数据进行了处理,改进算法对窄带信号定向较常规算法-3 dB束宽减小了13°,旁瓣级降低约8 dB。对有一定带宽的行船辐射噪声定向处理得到了更加精确的航迹图,海试数据处理结果证明了该算法的可行性和有效性。     

Single-channel speech enhancement method using reconstructive NMF with spectrotemporal speech presence probabilities

In this paper, a novel single microphone channel-based speech enhancement technique is presented. While most of the conventional nonnegative matrix factorization-based approaches focus on generating a basis matrix of speech and noise for enhancement, the proposed algorithm performs an additional process to reconstruct speech from noisy speech when these two elements are highly overlapped in selected spectral bands. This process involves a log-spectral amplitude based estimator, which provides the spectrotemporal speech presence probability to obtain a more accurate reconstruction. Moreover, the proposed algorithm applies an unsupervised learning method to the input noise, so it is adaptable to any type of environmental noise without a pre-trained dictionary. The experimental results demonstrate that the proposed algorithm obtains improved speech enhancement performance compared with conventional single channel-based approaches.     

A wavelet-based forward BSS algorithm for acoustic noise reduction and speech enhancement

In this paper, we address the problem of noise reduction and speech enhancement by adaptive filtering algorithm. Recently, the well known forward blind source separation (FBSS) structure has been largely studied and intensively used to reduce acoustic noise components and to enhance speech signal. The FBSS structure is often combined with adaptive algorithms to accelerate the adaptation of the cross-filters, and to improve noise suppression at the output. In this paper, we propose to use a wavelet transform decomposition in the FBSS structure by using a two-channel forward wavelet symmetric adaptive decorrelating (WFSAD) algorithm. The proposed WFSAD algorithm provides a better compromise between time and frequency resolution and improves robustness of the noise reduction process when compared with the classical two-channel forward symmetric adaptive decorrelating (FSAD) algorithm. Simulation results prove the efficiency of the proposed WFBSS algorithm in comparison with conventional ones in terms of several objective and subjective criteria.