球谐域类正则化宽带超指向性波束形成算法

提出一种球谐域类正则化宽带超指向性波束形成算法,通过结合超指向性波束形成器与延迟求和波束形成器控制阵列白噪声增益和指向性因数,推导类正则化宽带超指向性波束形器的球谐域表达式。在此基础上设计新的可控指向性因数波束形成器,可在不显著放大低频白噪声的前提下实现设定的波束宽度。采用32元球阵的仿真结果表明,对于语音测试场景,类正则化波束形成器的最高PESQ得分相比于传统的超指向性波束形成器与延迟求和波束形成器分别提高了约0.5和0.4,平均词错率分别减少了约9.5%和8.1%。主观测试实验也表明类正则化波束形成器在方向性噪声和扩散场噪声环境下都可以获得更好的主观听觉感受。利用实验数据对算法性能进行测试,实验结果同样验证了该方法在实际声学环境中的有效性。     

同心圆环阵列模态波束形成器设计

针对贝塞尔函数零点造成的模态波束形成器性能损失的问题,提出了基于均匀同心圆环阵列的模态波束形成器设计方法。从阵元域出发,建立了阵元域和圆谐波域之间的转换关系,推导了圆谐波域阵列信号处理表达式,其中重点推导了延时求和、最小方差无失真响应和多约束波束形成器的设计方法。理论证明了在平面各向同性噪声场中的基于同心环阵的圆谐波域最小方差无失真响应波束形成器等价于相位模式波束形成器。在该结论的基础上,综合考虑相位模式波束形成器在一定频率范围内的稳健性和指向性,提出了一种同心圆环阵列的内环半径优化方法。使用16元双环阵列对本文提出的波束形成算法进行仿真和实验,结果表明:同心圆环阵列能够较好地解决贝塞尔函数零点问题;其中多约束波束形成器能够在多个关联的性能指标之间取得合理的折衷,实现优于相位模式和延时求和波束形成器的声源定位性能。对内环半径优化方法进行仿真和分析结果表明,该方法得到了相较于传统方法更优的阵列结构。      

一种加权稀疏约束稳健Capon波束形成方法

为了克服标准Capon波束形成器旁瓣级高以及存在角度失配时性能急剧下降等缺点, 在稀疏约束Capon波束形成器的基础上, 提出了一种加权稀疏约束Capon波束形成器. 该方法利用波束响应的稀疏分布特性, 在标准Capon波束形成优化模型中加入旁瓣区域波束响应稀疏约束(l₁ 范数约束), 使旁瓣区域波束响应向量中非零元素的个数最小化; 通过阵列采样数据协方差矩阵特征分解得到信号子空间及噪声子空间, 利用信号子空间与噪声子空间的正交特性, 构造加权矩阵对稀疏约束进行加权, 使得稀疏重构时波束响应向量中不同角度对应的元素得到不同程度的约束. 该方法有效地抑制了Capon波束形成器的高旁瓣级, 加深了干扰方位零陷, 提高了阵列输出信干噪比. 由于稀疏约束, 波束响应向主瓣集中, 期望信号方向附近的波束响应都较大, 从而也提高了阵列抗导向矢量角度失配的能力. 数值仿真和水池实验验证了所提方法的有效性.     

交替迭代多约束波束优化设计

声学阵列波束形成器的阵增益、旁瓣级、稳健性等多个指标相互关联,对于包含加权向量范数约束和旁瓣控制的高增益波束优化设计问题,针对传统的内点法波束优化算法硬件实现困难及计算复杂度大等问题,提出了多约束交替迭代求解方法。通过引入辅助变量将范数约束和旁瓣约束解耦合,进而将多约束优化转化为两组单约束优化问题。给出了优化子问题的解析或近似解。理论分析表明,与现有内点优化算法相比,所提方法计算量较小,单次迭代复杂度仅为变量及约束个数的二次多项式。采用仿真和实测的阵列流形验证了设计效果,结果表明所提方法可实现旁瓣级,阵增益和稳健性等多个波束设计指标的合理折中,计算精度与二阶锥规划算法相当,但计算量更小。      

时域实现结构球面阵的最差性能优化稳健宽带波束形成器设计

针对球面阵模态域波束形成器对传声器失配误差敏感的问题,研究了时域实现结构球面阵的最差性能优化稳健宽带波束形成器设计。揭示了波束响应误差上界的保守性是影响现有最差性能优化方法性能的主要原因,进而提出了一种波束响应误差上界更为紧凑的优化设计方法。理论分析表明,与现有最差性能优化方法相比,所提出方法的代价函数更小、且可行解的范围也更大。针对最差性能优化设计方法存在的波束指向上阵列响应随频率变化波动偏大而引起信号失真的问题,还给出了一种减小波束图在指向方向上受信号频率影响的约束方法。仿真分析结果表明:在相同传声器失配误差条件下,所提出的设计方法要优于现有方法,得到的波束图具有更低的旁瓣级。      

逆波束形成的旁瓣抑制

为了降低普通逆波束形成旁瓣,提出了一种新的旁瓣抵消技术。该技术利用指向性函数中的旁瓣部分去估计声场的响应,然后从基阵估计的声场减去这部分指向性函数旁瓣估计的声场,从而得到最终的声场估计。由于较多的旁瓣响应被估计的声场扣除,最终估计的声场中的旁瓣大大降低。利用该技术可有效地抑制干扰,同时又可保持普通逆波束形成的方位高分辨力。湖试结果显示了该技术的有效性和可行性。     

结合精确阵列响应控制的时域宽带波束图综合

提出了采用时域结构的低计算复杂度宽带波束图综合方法。首先,引入自适应零陷波束设计思想,通过在特定频率与角度设置虚拟干扰源,实现对旁瓣区域阵列响应的精确控制,多次迭代控制后满足波束设计要求。其次,所提方法能够在现有波束设计基础上实现灵活调整,根据旁瓣级等设计参数的变化快速更新波束图。理论研究表明,相较现有采用内点法的波束图综合方法,所提方法降低了单次迭代计算复杂度,有效提升了波束设计效率。仿真实验与半消声室实验结果表明,该方法所设计波束形成器相较传统固定波束形成器具有更低的旁瓣级,并且对方向性干扰有更好的抑制效果。     

基于分组时延预处理的时域波束形成方法*

对于不等强度多目标情况下的弱目标检测问题,依据波束形成归一化指向性函数在目标方向上输出值为1、在非目标方向上输出值为小于1的特性,提出一种基于分组时延预处理的时域波束形成方法。该方法首先对线列阵接收数据进行分组处理;然后,按搜索角度对各组数据进行时延预处理,并对各组预处理结果进行相加,得到一组新数据;最后,对该组新数据按时域波束形成进行处理,得到该搜索角度对应波束值。理论推导、数值仿真和试验数据处理结果均表明：相比常规时域波束形成,该方法所得波束形成指向性函数发生了变化,旁瓣级得到了13dB以上的改善,降低了强目标旁瓣级对弱目标检测的影响。     

采用空间谱加权稀疏约束的稳健Capon波束形成方法

针对标准Capon波束形成器旁瓣级高以及期望信号方向存在误差时,性能会严重下降的问题,提出一种基于噪声及干扰空间谱加权的稀疏约束Capon波束形成方法。该方法利用波束响应本身具有的稀疏特性,以及在稀疏重构算法中l₁范数具有抑制较大值保护较小值的特性,使用噪声及干扰的空间谱对不同角度的波束响应进行加权,然后施以稀疏约束。仿真、水池试验与湖上试验中,与其它几种方法比较表明该方法可以降低波束形成器的旁瓣级,获得较深的零陷,提高了阵列输出信干噪比与抗导向向量误差的能力。      

时域宽带恒定束宽波束形成的实验研究

由于实际水声系统中的阵列模型往往与理想阵列模型失配,导致基于理想阵列模型设计的时域恒定束宽波束形成器所形成的波束图在实际应用中严重畸变,从而不能有效地抑制旁瓣干扰和无失真地接收宽带信号。本文利用水池实验说明阵列模型失配对基于理想阵列模型设计的时域恒定束宽波束形成器的影响,并提出一种有效的设计方法以减小这种影响,即通过实际测量,获得包含失配信息在内的实际阵列模型,并将该模型应用到时域恒定束宽波束器的设计当中。实验结果表明,基于实际阵列模型的设计可以有效地减小由于阵列模型失配带来的影响,这对设计实际系统可以使用的时域恒定束宽波束形成器具有重要的意义。     

Comparison of the measured and theoretical performance of a broadband circular microphone array

The design and construction of a circular microphone array (CMA) that has a wide frequency range suitable for human hearing is presented. The design of the CMA was achieved using a technique based on simulated directivity index (DI) curves. The simulated DI curves encapsulate the critical microphone array performance limitations: spatial aliasing, measurement noise, and microphone placement errors. This paper demonstrates how the non-regularized DI curves for a given beamforming order clearly define the bandwidth of operation, in other words, the frequency band for which the beamformer has relatively constant and maximum directivity. Detailed and comprehensive experimental data that characterizes the CMA beamformer are also presented.     

波束空间能量约束的稳健自适应波束形成

为改善波束形成器在一定失配条件下的工作性能,提出了一种稳健自适应波束形成方法。该方法通过设计信号在波束空间能量分布的约束条件,利用阵列数据协方差矩阵,结合迭代二阶锥规化方法,实现信号驾驶向量的估计,最后利用该估计得到最小方差无失真响应波束形成器。该方法在较恶劣的失配条件下也可以自适应地准确估计信号驾驶向量并保证其成分无失真地通过,使系统的稳健性得到改善。数据仿真结果表明,该方法在理想条件下与成熟的对角加载法性能相当,在失配条件下则优于后者。      

Acoustic analysis by spherical microphone array processing of room impulse responses

Spherical microphone arrays have been recently used for room acoustics analysis, to detect the direction-of-arrival of early room reflections, and compute directional room impulse responses and other spatial room acoustics parameters. Previous works presented methods for room acoustics analysis using spherical arrays that are based on beamforming, e.g., delay-and-sum, regular beamforming, and Dolph-Chebyshev beamforming. Although beamforming methods provide useful directional selectivity, optimal array processing methods can provide enhanced performance. However, these algorithms require an array cross-spectrum matrix with a full rank, while array data based on room impulse responses may not satisfy this condition due to the single frame data. This paper presents a smoothing technique for the cross-spectrum matrix in the frequency domain, designed for spherical microphone arrays, that can solve the problem of low rank when using room impulse response data, therefore facilitating the use of optimal array processing methods. Frequency smoothing is shown to be performed effectively using spherical arrays, due to the decoupling of frequency and angular components in the spherical harmonics domain. Experimental study with data measured in a real auditorium illustrates the performance of optimal array processing methods such as MUSIC and MVDR compared to beamforming.     

Adaptive field-of-view imaging for efficient receive beamforming in medical ultrasound imaging systems

Quadrature demodulation-based phase rotation beamforming (QD-PRBF) is commonly used to support dynamic receive focusing in medical ultrasound systems. However, it is computationally demanding since it requires two demodulation filters for each receive channel. To reduce the computational requirements of QD-PRBF, we have previously developed two-stage demodulation (TSD), which reduces the number of lowpass filters by performing demodulation filtering on summation signals. However, it suffers from image quality degradation due to aliasing at lower beamforming frequencies. To improve the performance of TSD-PRBF with reduced number of beamforming points, we propose a new adaptive field-of-view (AFOV) imaging method. In AFOV imaging, the beamforming frequency is adjusted depending on displayed FOV size and the center frequency of received signals. To study its impact on image quality, simulation was conducted using Field II, phantom data were acquired from a commercial ultrasound machine, and the image quality was quantified using spatial (i.e., axial and lateral) and contrast resolution. The developed beamformer (i.e., TSD-AFOV-PRBF) with 1024 beamforming points provided comparable image resolution to QD-PRBF for typical FOV sizes (e.g., 4.6% and 1.3% degradation in contrast resolution for 160 mm and 112 mm, respectively for a 3.5 MHz transducer). Furthermore, it reduced the number of operations by 86.8% compared to QD-PRBF. These results indicate that the developed TSD-AFOV-PRBF can lower the computational requirement for receive beamforming without significant image quality degradation.     

An acousto-optic beamformer

There is a great variety of beamforming techniques that can be used for localization of sound sources. The differences among them usually lie in the array layout or in the specific signal processing algorithm used to compute the beamforming output. Any beamforming system consists of a finite number of transducers, which makes beamforming methods vulnerable to spatial aliasing above a certain frequency. The present work uses the acousto-optic effect, i.e., the interaction between sound and light, to localize sound sources in a plane. The use of a beam of light as the sensing element is equivalent to a continuous line aperture with an infinite number of microphones. This makes the proposed acousto-optic beamformer immune to spatial aliasing. This unique feature is illustrated by means of simulations and experimental results within the entire audible frequency range. For ease of comparison, the study is supplemented with measurements carried out with a line array of microphones.     

球谐域自适应混响抵消与声源定位算法

提出了一种基于球谐域的自适应混响抵消与声源定位算法,该方法通过去混响处理改善语音质量,并提高球谐域定位算法在混响环境下的定位性能。推导了基于多通道线性预测的自适应混响抵消算法在球谐域的表达式,针对刚球模型提出分阶处理的去混响方法,并对去混响后的信号进行波达方向估计。采用32元球阵的仿真结果表明,相比于球谐域不分阶去混响方法,该方法最大可减少约2/3的运算量,同时语音PESQ得分及SRMR均显著提高。利用实验数据对算法性能进行测试,实验结果验证了该方法在实际声学环境中去混响和声源定位的有效性。      

Robust supergain beamforming for circular array via second-order cone programming

The phenomenon of supergain for a circular array and its robust beamforming are presented. The coplanar superdirective array gain of the circular array, although it is not so extreme as an endfire line array, outperforms a lot over that of a conventional delay-and-sum beamformer in isotropic noise fields when the inter-element spacings are much smaller than one-half wavelength. However, optimum beamforming algorithms can be extremely sensitive to slight errors in array characteristics. The performance are known to degrade significantly if some of underlying assumptions on the sensor array is violated. Therefore, white noise gain constraint is used to improve the robustness of the supergain beamformer against random errors. We show that the design of the weight vector of robust supergain beamformer can be reformulated as a form of second-order cone programming and resolved efficiently via the well-established interior point method. Results of computer simulation for a 24-element circular array confirm satisfactory performance of the approach proposed in this paper.