Sound Source Localization Based on SRP-PHAT Spatial Spectrum and Deep Neural Network

Microphone array-based sound source localization(SSL)is a challenging task in adverse acoustic scenarios.To address this,a novel SSL algorithm based on deep neural network(DNN)using steered response power-phase transform(SRP-PHAT)spatial spectrum as input feature is presented in this paper.Since the SRP-PHAT spatial power spectrum contains spatial location information,it is adopted as the input feature for sound source localization.DNN is exploited to extract the efficient location information from SRP-PHAT spatial power spectrum due to its advantage on extracting high-level features.SRP-PHAT at each steering position within a frame is arranged into a vector,which is treated as DNN input.A DNN model which can map the SRP-PHAT spatial spectrum to the azimuth of sound source is learned from the training signals.The azimuth of sound source is estimated through trained DNN model from the testing signals.Experiment results demonstrate that the proposed algorithm significantly improves localization performance whether the training and testing condition setup are the same or not,and is more robust to noise and reverberation.

NOVEL SPECTRUM SENSING METHOD BASED ON THE SPATIAL SPECTRUM FOR COGNITIVE RADIO SYSTEMS

Spectrum sensing is the key problem for Cognitive Radio(CR) systems.A method based on the Peak-to-Average Amplitude-Ratio(PAAR) of the Spatial Spectrum(SS) of the received signals is proposed to sense the available spectrum for the cognitive users with the help of the multiple antennas at the receiver of the cognitive users.The greatest advantage of the new method is that it requires no information of the noise power and is free of the noise power uncertainty.Both the simulation and the analytical results show that the proposed method is robust to noise uncertainty,and greatly outperform the classical Energy Detector(ED) method.

Effect of spatial spectrum overlap on Fourier ptychographic microscopy

Fourier ptychographic microscopy(FPM)is a newly developed imaging technique which stands out by virtue of its high-resolution and wide FOV.It improves a microscope's imaging perfor-mance beyond the diffraction limit of the employed optical components by illuminating the sample with oblique waves of different incident angles,similar to the concept of synthetic aperture.We propose to use an objective lens with high-NA to generate oblique illuminating waves in FPM.We demonstrate utilizing an objective lens with higher NA to iluminate the sample leads to better resolution by simulations,in which a resolution of 0.28 pum is achieved by using a high-NA illuminating objective lens(NA=1.49)and a low-NA collecting objective lens(NA=0.2)in coherent imaging(λ=488 nm).We then deeply study FPM's exact relevance of convergence speed to spatial spectrum overlap in frequency domain.The simulation results show that an overlap of about 60%is the optimal choice to acquire a high-quality recovery(520*520 pixels)with about 2 min's computing time.In addition,we testify the robustness of the algorithm of FPM to additive noises and its suitability for phase objects,which further proves FPM's potential application in biomedical imaging.

SPATIAL SPECTRUM ESTIMATION OF COHERENT SOURCES IMPINGING ON NONUNIFORM ARRAYS

The problem of bearing estimation of coherent signal impingiog on an array ofarbitrary geometry is studied.Both approaches are developed,the first one synthesizes theobserved array data into the outputs of a linear uniform array and then processes them viaconventional techniques such as spatial smoothing etc.The second approach is based on theinvariant subspace rotation operation which is equivalent to the translational displacement ofthe artay,linearly independent signal vectors are obtained thereby to span perfectly the signalsubspace.As compared with the conventional processing techniques,the approach based oninvariant subspace rotation does not lead to an effective decrease in aperture size and thereforea decrease in resolution capability.The computer simulations are conducted to demonstrate theeffectiveness of the presented approaches.

A detecting method for line-spectrum target by fusing output DC jump to fluctuations ratio of sub-band spatial spectrum and beam space

The low frequency line components of the radiated noise from an underwater target usually have both high spectrum level and sustained stability. This feature could be used to increase the detection performance of conventional broadband energy integration method. The required spectrum level is theoretically discussed when the detection performance of the known line detection is better than that of broadband energy integration method. Under the condition of the target can be detected in line spectrum band, the relationship between the line spectrum level and signal to noise ratio （SNR） is also discussed. This paper proposes a line spectrum target detection method that a matrix using DC jump to fluctuations ratios of sub-band spatial spectrum and beam space output is constructed. This matrix acts as a filter that the line spectrum target with certain frequency and azimuth is passed at most. By fusing with the other sub band results, the conventional detection performance can be improved. At the same time, the applicable condition and detection performance are analyzed in the paper. The simulation and the sea trial data processing results show that the algorithm can effectively extract weak goal line spectrum target under the condition of multi-interference. The algorithm doesn＇t need multi-frame statistics and the detection performance is closer to the optimal line spectrum method.

Spatio-Temporal Population Density and Spatial Dynamic Spectrum Allocation

A realistic population density distribution scenario in conjunction with the spatial dynamic spectrum allocation （DSA） is taken into account to mitigate the spectrum wastage in terms of extra guard bands. For the insertion of the extra guard bands, an efficient strategy based on self-assessment is applied to each victim cell individually and independently. Consequently, it is no more required to spread the extra guard band over the whole DSA region. Simulation results StlOW an improvement of 3% -4% in percentage of satisfied users for Universal Mobile Telecommunications System （UMTS） network and 4%-5% for Digital Video Broadcasting Terrestrial （DVB-T） network.

Spectrum spatial structure characteristic analysis of remote sensing alteration information and interference factors

Based on the statistical characteristics of remote sensing data, the spatial geometric structure characteristics of spectral data and distribution of background, interference and alteration information in characteristic space were researched through the analysis of two-dimensional and three-dimensional scatter diagrams. The results indicate that the hyper-space of remote sensing multi-data aggregation belongs to low-dimensional geometric structure, i.e. hyperplane form, and anomalous point groups including alteration information usually dissociate out of hyperplane. Scatter diagrams of remote sensing data band are mainly presented as two distribution forms of single-ellipse and dual-ellipse. Clarifying the relations of three objects of background, disturbance and alteration information in remote sensing images provides an important technical thought and guidance for accurately detecting and extracting remote sensing alteration information.

2-D SPATIAL-SPECTRUM ESTIMATION FORWIDE-BAND SOURCES BASED ON INTERPOLATED CIRCULAR ARRAYS

In array signal processing, 2-D spatial-spectrum estimation is required to determine DOA of multiple signals. The circular array of sensors is found to possess several nice properties for DOA estimation of wide-band sources. C. U. Padmini, et al.(1994) had suggested that the frequency-direction ambiguity in azimuth estimation of wide-baud signals received by a uniform linear array (ULA) can be avoided by using a circular array, even without the use of any delay elements. In 2-D spatial-spectrum estimation for wide-band signals, the authors find that it is impossible to avoid the ambiguity in source frequency-elevation angle pairs using a circular array. In this paper, interpolated circular arrays are used to perform 2-D spatial-spectrum estimation for wide-band sources. In the estimation, a large aperture circular array (Υ】λmin/2) is found to possess superior resolution capability and robustness.

A new spatial coherence model and analytical coeffi cients for multi-support response spectrum combination

In this paper, a new spatial coherence model of seismic ground motions is proposed by a fitting procedure. The analytical expressions of modal combination （correlation） coefficients of structural response are developed for multi-support seismic excitations. The coefficients from both the numerical integration and analytical solutions are compared to verify the accuracy of the solutions. It is shown that the analytical expressions of numerical modal combination coefficients are of high accuracy. The results of random responses of an example bridge show that the analytical modal combination coefficients developed in this paper are accurate enough to meet the requirements needed in practice. In addition, the computational efficiency of the analytical solutions of the modal combination coefficients is demonstrated by the response computation of the example bridge. It is found that the time required for the structural response analysis by using the analytical modal combination coefficients is less than 1/20 of that using numerical integral methods.

半监督空谱局部判别分析的高光谱影像特征提取

为充分利用高光谱影像中蕴含的空谱特征,提出了一种半监督空谱局部判别分析的高光谱影像特征提取算法(S4LFDA)。鉴于高光谱数据集具有空间一致性,首先将像元进行空间重构,保存高光谱数据的近邻关系;其次引入光谱信息散度重构像元间的相似度;为了充分利用大量无标签样本提高算法性能,采用模糊C均值聚类算法对样本进行聚类分析得到伪标签;然后通过增加规范化项到局部力导引算法(FDA)的类内散度矩阵和类间散度矩阵中,以此保持无标签样本的聚类结构一致性;最后通过局部FDA算法来保持有标签样本类间散度最大化和类内散度最小化并求解最佳投影向量。S4LFDA算法既保持了数据集在光谱域的可分性,又保持了像元在空间区域内的近邻关系,合理利用有标签样本及无标签样本,提高了算法的分类性能。在Pavia University和Indian Pines数据集上进行实验,总体分类精度达到95.60%和94.38%。与其他维数约简算法相比,该算法有效提高了地物分类性能。

一种波段聚类和多尺度结构特征融合的高光谱图像分类模型

高光谱图像包含丰富的地物信息,在农业、工业和军事等领域应用广泛。因此,高光谱图像的识别与分类是一项重要的研究课题。然而,高光谱图像存在光谱维度高、噪声大、标记样本有限等问题,并未取得很好的分类效果。针对以上问题,提出一种波段聚类和多尺度结构特征融合的高光谱图像分类模型(ASPS-MRTV)。该方法主要包括以下几个步骤,首先,对高光谱数据进行归一化处理,将归一化后的三维图像按光谱维等分为n个子空间;其次,采用粗细划分策略构造自适应子空间光谱特征提取框架,将每个空间波段拉伸为一维向量后用信息散度构造波段的相似性矩阵,按照聚类的思想对n个子空间进行自适应;然后,将每个自适应子空间的光谱波段平均值进行叠加,形成光谱特征;最后,对所得到的光谱特征数据利用多尺度相对全变分技术提取结构特征。为了增强样本的线性可分性,在数据堆叠之后进行核主成分分析,最终形成空谱特征。对比实验中统一使用惩罚参数C和核参数σ都为24.5的SVM进行分类。经测试,ASPS-MRTV网络模型在Indian Pines、 University of Pavia两个数据集上分别以5%, 1%的训练样本达到了97.06%、 98.98%的总体分类精度。实验结果表明,该模型与SVM、 ASPS(ED)、 ASPS(ID)、 ASPS-LBP、 ASPS-GlCM、 ASPS-BF模型相比,在分类性能和计算效率方面都取得了更优的效果,有效提高小样本下高光谱图像的分类精度。

Media Access with Spatial Reuse for Cooperative Spectrum Sensing

The increasing interest for wireless communication services and scarcity of radio spectrum resources have created the need for a more flexible and efficient usage of the radio frequency bands. Cognitive Radio (CR) emerges as an important trend for a solution to this problem. Spectrum sensing is a crucial function in a CR system. Cooperative spectrum sensing can overcome fading and shadowing effects, and hence increase the reliability of primary user detection. In this paper we consider a system model of a dedicated detect-andforward wireless sensor network (DetF WSN) for cooperative spectrum sensing with k-out-of-n decision fusion in the presence of reporting channels errors. Using this model we consider the design of a spatial reuse media access control (MAC) protocol based on TDMA/OFDMA to resolve conflicts and conserve resources for intra-WSN communication. The influence of the MAC protocol on spectrum sensing performance of the WSN is a key consideration. Two design approaches, using greedy and adaptive simulated annealing (ASA) algorithms, are considered in detail. Performance results assuming a grid network in a Rician fading environment are presented for the two design approaches.

基于空间光调制器的艾里光束能量调制方法

艾里光束由于具有无衍射、自加速和自愈合等特性,具有重要的应用前景。本文基于空间光调制器可编程特性,提出了一种可调谐的艾里光束能量调制新方法。从理论上深入分析通过在相位图中引入一个可控的调制相位而实现了艾里光束能量调控的方法,并且量化了调制参数与能量调制范围的关系。仿真和实验结果表明通过改变调制相位可以灵活地调节光束能量分布,本文的研究有助于进一步推动艾里光束的应用。

基于波束形成的变压器噪声成像技术研究与应用

作为变电站中最为昂贵和重要的设备,变压器一直是科研人员的研究重点对象。目前对变压器噪声源识别的相关研究较少,因此结合波束形成技术针对变压器噪声源识别展开研究。首先根据所调研的变压器噪声特性,确定了所关注的噪声频段和所用的声源识别方法,阐述了宽带信号的波达方向估计与波束形成的相关理论,进而整理出了一种基于频域数据的宽带变压器信号定位思路,利用仿真结果证实了所提定位思路的合理性与准确性。最后基于LabVIEW平台搭建了一套变压器噪声源识别系统,并通过对不同情况下实际运行变压器的现场试验,对系统的定位效果进行了测试,其定位误差在2°左右,结果证明噪声源识别思路和搭建的系统平台具有良好的可靠性与可用性。

移动强干扰条件下大孔径阵列稀疏空间谱估计

无源声呐阵列的目标方位估计性能受到强干扰的显著影响,在干扰源快速移动时大孔径阵列还可能存在快拍不足的问题,为此提出一种结合波束域预处理以实现干扰抑制和数据降维的稀疏空间谱估计方法。首先在预处理阶段对干扰方向形成自适应零点,并通过协方差矩阵锥化展宽零点,然后对预处理后的波束域数据应用稀疏贝叶斯学习进行空间谱估计。与现有方法相比,波束域预处理可以有效减少阻带范围内干扰声源的影响,提高通带范围内空间谱估计的精度和计算效率,零点展宽可以进一步加强稳健性,避免出现快拍不足问题。仿真分析和海试结果表明,所提方法在移动强干扰条件下仍能准确估计目标声源的多途到达角,且计算时间远低于阵元域稀疏方法。

空间特性光谱反演及对户外光谱修正的可行性研究

户外高光谱探测可以快速获取样品的光谱信息,但受环境光线和样品二向反射特性的影响,采集到的光谱并不能准确反映样品的真实信息,对户外探测精度有一定影响。为了提高户外高光谱探测精度,提出了一种使用空间特性光谱对户外光谱进行修正的方法,以冬枣、红提、小白杏为研究对象,使用Walthall、Shibayama、Ross-li、Roujean与Rahman这5种BRDF模型反演3种果品的空间特性光谱,利用反演的空间特性光谱对户外光谱进行修正,之后分别建立暗室光谱、户外光谱与修正光谱的品质预测模型。反演结果表明:3种果品的空间特性光谱均有较好的反演效果,反演误差从低到高依次为冬枣、小白杏、红提,平均决定系数R2分别为0.957、0.947、0.927,平均误差分别为3.56%、4.90%、8.23%;5种BRDF模型中,Walthall模型的反演效果最佳,平均决定系数R2与误差分别为0.949、5.33%,Ross-li模型的反演效果最差,平均决定系数与误差分别为0.934、6.05%。户外光谱修正结果表明:户外光谱经过修正后噪声减少,光谱更为平滑,且光谱趋势与暗室光谱一致,受反演效果影响,冬枣光谱的修正效果最佳,而红提与小白杏的修正光谱中噪声较多。品质预测模型结果表明:3种果品的品质预测模型预测效果差异较大,从高到低依次为冬枣、小白杏、红提,可能与3种果品的品质不同有关;5种BRDF模型得到的修正光谱所建立的模型预测效果不同,但无显著差异;预测模型中,使用暗室光谱建立的预测模型最优,修正光谱建立的模型预测能力优于户外光谱建立的模型,表明户外光谱经过修正后模型预测能力得到提升。BRDF模型能够较好地反演果品的空间特性光谱,修正后的光谱与暗室光谱较为接近,修正光谱建立的模型优于户外光谱建立的模型,表明使用空间特性光谱对户外光谱进行修正的方法是可行的,可为提高户外无损检测的精度提供一种新的思路。

圆柱形空间约束对LIBS检测煤粉颗粒流的优化研究

为了提高激光诱导击穿光谱技术(LIBS)检测煤粉颗粒流时光谱信号的质量和检测结果的可靠性,根据颗粒流的检测特点,设计了一种圆柱形空间约束,并分析了其对LIBS直接检测煤粉颗粒流的实验系统的优化效果。先后对比分析了添加约束前后煤中一些重要元素的特征谱线的绝对强度、信噪比(SNR)和脉冲到脉冲间的相对标准偏差(RSD)等重要信息。结果表明:圆柱形空间约束能够有效地提升LIBS检测煤粉颗粒流的光谱强度和信号的稳定光谱信号的重复性,达到提高检测结果的可靠性的目的。

基于DSP的快速MUSIC测角算法

基于阵列信号处理的MUSIC算法,由于其超高的角分辨率和角精度,受到广泛的关注。但是其高运算复杂度,较大地限制了算法的硬件实现。为此,基于MUSIC算法的基本原理,文中分析了该算法在DSP(TI1843)硬件系统实现中关键子步骤的运行时长。针对最为耗时和占用内存最大的空间谱构建和谱峰搜索子步骤,分别提出了便于工程实现的简化方法,并通过硬件平台移植验证了算法的可行性,对比发现该方法大大减少了运算时间和内存空间,具有较高实际应用价值。

强干扰下基于三阶互累积量的FMCW雷达信号DOA估计算法

针对车载调频连续波雷达存在相互干扰的实际情况,本文提出了一种基于三阶互累积量的多目标波达方向超分辨估计方法.该方法利用同步采集的阵列信号并根据回波信号和干扰信号的非相关特性,构造三阶互累积量矩阵,随后根据子空间方法获得二维空间谱,实现距离和方位的联合估计.计算机仿真结果表明,本文所提方法能够在抑制强干扰影响的同时实现3°以内多目标识别,与现有超分辨法和重构法相比,具有更高的角度分辨率和估计精度.