基于短时谱最小均方误差估计的语音增强和剩余噪声衰减

本文研究了三种基于语音短时谱最小均方误差估计的语音增强方法:短时谱幅度最小均方误差估计,短时对数谱最小均方误差估计和短时相对谱幅度最小均方误差估计,在理论分析基础上对它们进行了实验研究.计算机仿真结果表明,在加性白色高斯噪声污染下,当带噪语音信噪比为+5～-10dB时,处理后的语音信噪比提高了3.4～12dB。短时对致谱最小均方误差估计的效果最好,试听实验也证实了这一点。 文中还对增强后剩余噪声的衰减问题进行了研究。利用中心削波并对估值器中的增益函数进行修正,可明显地减弱剩余噪声。     

基于压缩感知的自适应卡尔曼滤波

针对稀疏流信号,提出了一种自适应卡尔曼滤波恢复方法,该算法基于压缩感知AIC结构,用有限长的窗口对信号进行观测,利用前后窗内信号之间的相关性,建立信号的状态转移方程,并与压缩感知获得的观测方程共同构成信号的状态空间模型,进而利用降阶的卡尔曼滤波算法近似得到信号的最小均方误差估计。信号重构阶段通过卡尔曼滤波迭代逐渐得到精确的支撑集,与以往仅用起始阶段的恢复结果获得支撑集的方法相比,本算法对起始阶段恢复支撑集的算法的精确程度要求不高,从而降低了整个算法的复杂度和要求的观测维度。仿真结果显示,这种自适应的卡尔曼滤波算法在宽带流信号的恢复中可以有效地降低所需观测维度,且最终结果可近似地收敛到信号的最小均方误差估计。     

基于成像激光雷达的地形辅助导航研究

成像激光雷达通过扫描可以测量高精度的地形数据,基于成像激光雷达的地形辅助导航系统是纠正惯导误差的有效手段.根据激光雷达测距模型提出了基于成像激光雷达的地形辅助导航系统,将激光雷达看作一个多维距离传感器阵列,利用卡尔曼滤波器迭代估计系统的状态误差,从而纠正惯导累积误差.根据激光雷达测距关系推导了系统的扩展卡尔曼滤波方程,并对成像激光雷达多维测量数据采用最小均方误差准则进行融合,融合滤波器组合了多个测量数据的信息,有效克服了测量噪声和数据丢失对单个滤波器的影响,从而提高了导航性能.然后利用局部可观测性对系统的性能进行了分析,并对提出的算法作了大量的仿真实验进行验证.     

基于正则化约束总体最小二乘的单站DOA-TDOA无源定位算法

针对利用单站外辐射源的目标无源定位问题,该文提出一种联合到达角度和时差信息的正则化约束总体最小二乘(RCTLS)定位算法。首先,将非线性的到达角度和时差的观测方程进行线性化处理,分析了方程系数矩阵可能出现的病态问题,将定位问题建立为RCTLS模型,并采用牛顿迭代方法对模型求解,从而得到目标位置估计。最后,推导了算法的理论误差,并按照均方误差最小的原则推导了正则化参数的最优值。仿真结果表明,算法的定位精度和鲁棒性均优于约束总体最小二乘(CTLS)算法。此外,对系统几何精度因子图的分析表明,目标及外辐射源的位置对定位精度也有影响。     

联合角度和时差的单站无源相干定位加权最小二乘算法

针对利用单个观测站接收多个外辐射源信号从而实现对目标定位的单站无源相干定位问题, 该文提出了一种联合角度和时差的加权最小二乘定位算法。首先, 将角度和时差的观测方程线性化处理, 考虑方程中的各项误差, 将定位问题建立为加权最小二乘模型。然后利用迭代方法对模型求解。最后, 对算法的定位性能进行了理论分析。仿真结果表明, 不同于仅时差定位方法至少需要3个外辐射源才能定位, 联合角度和时差定位方法仅需一个外辐射源即可定位, 且在同样数量外辐射源条件下估计精度高于仅时差定位;算法的均方误差低于最小二乘算法, 在时差测量噪声较大时定位精度仍然能逼近克拉美罗界。此外, 对系统几何精度因子图的分析表明, 目标及辐射源的位置对定位精度也有重要影响。      

基于DOA-TDOA-FDOA的单站无源相干定位代数解

针对利用单站接收多个外辐射源信号实现目标定位的问题,该文提出一种基于两步加权最小二乘的到达角度(DOA)、时差(TDOA)和频差(FDOA)联合定位代数解算法.首先,在第1步加权最小二乘估计中,通过引入辅助参数,将非线性的角度、时差和频差方程转化为伪线性形式,并利用加权最小二乘得到目标位置和速度的粗估计;而后在第2步加权最小二乘估计中利用辅助参数和目标位置参数之间的约束关系来构造另外一组线性方程,并再次利用加权最小二乘得到目标位置和速度的精确估计.推导了联合角度、时差和频差定位的克拉美罗界.理论分析和仿真结果表明,算法在观测误差较小时的定位误差可以达到克拉美罗界.     

宽带平坦增益光纤喇曼放大器泵浦配置的快速搜索算法

本文提出了一种用于宽带平坦增益喇曼光纤放大器的泵浦配置的快速搜索算法,该算法只需要指定放大器采用的介质光纤参数、信号功率水平、放大器的目标增益和允许采用的泵浦数,就能够搜索到最小均方误差意义下泵浦的配置结果.算法利用最小均方差意义下的最佳逼近,把搜索波长的遗传算法和确定功率积分的具有线性约束条件的二次型凸规划问题结合起来,从而加快了搜索速度.     

语音增强:使用burg谱先验信噪比估计消除"音乐噪声"

针对谱衰减增强算法中存在的"音乐噪声"问题,提出一种新的估计先验信噪比的方法.文章深入分析了谱衰减算法中"音乐噪声"的产生原因和抑制机理,并基于对先验信噪比的平滑,准确性和分辨率三者之间的关系以及它们对增强算法的影响等方面的综合考虑,选择burg方法结合直接判决准则进行估计,从而得到更加接近于真实情况下的先验信噪比.实验证明,在最小均方误差(MMSE)和对数谱最小均方误差(LSA-MMSE)增强算法中使用文章提出的先验信噪比估计,可以有效抑制"音乐噪声",同时,语音失真也进一步降低.     

宽视场湍流轮廓线测量方法研究

多层共轭自适应光学(MCAO)等技术可以有效改善湍流大气的干扰。为了MCAO的设计和性能优化,需要对址点的大气湍流轮廓线进行测量。基于面源目标像相对运动方差或协方差统计反演的湍流轮廓线测量方法得到的反演线性方程组需要通过离散化获得,离散化误差导致观测视场不能太大。较小的视场使方程数目受限,导致湍流轮廓线反演结果受运动方差或协方差测量误差的影响。以PML方法为研究实例,提出了分层积分系数矩阵方法,该方法可以显著改善离散化误差,大幅增加观测视场范围,降低协方差误差对测量结果的影响,提高湍流轮廓线的反演精度。数值模拟结果表明,基于分层积分系数矩阵的PML方法的视场可达到400″,测量误差也可大幅度降低。     

基于宽巷组合模型求解整周模糊度的研究

传统的最小二乘降相关法(LAMBDA)在求解模糊度时,通常会因观测历元较少而出现法方程秩亏的现象,导致无法成功求解模糊度。针对此问题,提出利用双频宽巷的线性组合观测模型解算模糊度。同时,研究了观测量误差均方差值和OVT检验历元个数对模糊度解算的影响。结果表明,整周模糊度解算成功率受伪距双差观测误差均方差值的变化影响较大,模糊度解算成功率随着检验历元个数增加而增大。     

Miniaturized microstrip lowpass filter using cylindrical-shaped resonators for integrated applications

A novel compact microstrip lowpass filter is presented in this paper. The structure uses stepped impedance and radial patches and it has the ??3 dB cut-off frequency at 4.39 GHz. The proposed lowpass filter has a wide rejection band from 6 to 30 GHz with rejection level more than ??20 dB. It has also other advantages such as: low insertion-loss <?0.1 dB and high return-loss more than 22 dB in 80% of the passband. The circuit size of the presented filter is about 90 mm². Such important advantages make the proposed lowpass filter applicable to be integrated in many high frequency circuits such as power amplifiers, oscillators, couplers and power dividers. It also can block the unwanted video transmitter interference nearby 4 GHz such as implementing between transmitter and antenna. The lowpass filter has simulated and fabricated and a good agreement was seen between the simulation and fabrication results.     

基于LTCC技术的低通滤波器快速设计与测试方法

介绍了一种通带为0-1.2 GHz的LTCC多层低通滤波器快速设计方法。利用滤波器设计软件,通过选择相应的参数,可以快速地设计出低通滤波器电路图,再将原型电路在三维电磁场仿真软件HFSS中建立滤波器模型。根据厂商提供的电容、电感等元器件模型库,根据模型库中的电容、电感值估算本次设计所需的元件大小,在HFSS中可以快速的建立模型,仿真结果可以很快的满足指标要求。最后采用标准LTCC工艺实现出尺寸为3.2 mm×1.6 mm×1.0 mm的低通滤波器。运用该方法可以帮助工程师快速地设计LTCC滤波器,有很强的实用性和便利性。     

Power deposition of a microstrip applicator radiating into alayered biological structure

The power deposited by a microstrip antenna into a layered biological structure is presented. The solution is based on an integral equation for the surface current density on the antenna and on an electric Green's dyadic for the fields inside a planar stratified medium. The integral equation is solved using the method of moments in conjunction with the point-matching technique. The modeling of the surface current takes the edge conditions into account. Special attention is devoted to a correct modeling of the excitation of the antenna by a coaxial feed. The numerical results focus on the power deposition as a function of depth     

Antenna Power Cutback Manipulation to Optimize Relay Power Allocation over MAMR Networks Employing Zero Forcing Scheme

Power allocation between relays and data streams is an important aspect of MIMO relay systems. Different methods are designed to consider weight matrices within relay channels. Commonly employed methods are based on linearly distributed multi antenna relay beamforming with match filter, zero forcing and MMSE techniques. This article proposes a cutback relay power allocation technique to design a suboptimal beamforming weight matrix by ZF–ZF method to decrease BER considerably at high SNRs. This method allocates the unused relay antenna power to others through the associated relay. This allows maximum workable power to increase; consequently the system performance is improved.     

Planar near-field to far-field transformation using an array ofdipole probes

A method is presented for computing far-field antenna patterns from measured near-field data measured by an array of planar dipole probes. The method utilizes the near-field data to determine some equivalent magnetic current sources over a fictitious planar surface which encompasses the antenna. These currents are then used to find the far fields. The near-field measurement is carried out by terminating each dipole with 50 Ω load impedances and measuring the complex voltages across the loads. An electric field integral equation (EFIE) is developed to relate the measured complex voltages to the equivalent magnetic currents. The mutual coupling between the array of probes and the test antenna modeled by magnetic dipoles is taken into account. The method of moments with Galerkin's type solution procedure is used to transform the integral equation into a matrix one. The matrix equation is solved with the conjugate gradient-fast Fourier transformation (CG-FFT) method exploiting the block Toeplitz structure of the matrix. Numerical results are presented for several antenna configurations to show the validity of the method     

多层电厚微带天线的阻抗特性

本文提出了一种多层介质电厚微带天线的一种分析方法。从无界空间中的电并矢格林函数出发,导出了分层介质中任意一点具有任意取向的水平电偶极子场的表达式。以此为基础,利用Richmond反作用积分方程,建立了关于贴片上电流分布的积分方程。通过适当选取电流基函数,使本方法适用于电厚介质微带天线。应用Galerkin方法建立矩阵方程。借助于计算机求解该矩阵方程,可得到天线的各项特性。文中用实例进行了验证。     

Planar near-field to far-field transformation using an equivalentmagnetic current approach

An alternative method is presented for computing far-field antenna patterns from near-field measurements. The method utilizes the near-field data to determine equivalent magnetic current sources over a fictitious planar surface that encompasses the antenna, and these currents are used to ascertain the far fields. Under certain approximations, the currents should produce the correct far fields in all regions in front of the antenna regardless of the geometry over which the near-field measurements are made. An electric field integral equation (EFIE) is developed to relate the near fields to the equivalent magnetic currents. The method of moments is used to transform the integral equation into a matrix one. The matrix equation is solved with the conjugate gradient method, and in the case of a rectangular matrix, a least-squares solution for the currents is found without explicitly computing the normal form of the equation. Near-field to far-field transformation for planar scanning may be efficiently performed under certain conditions. Numerical results are presented for several antenna configurations     

An integral equation and its application to spiral antennas onsemi-infinite dielectric materials

This paper presents an integral equation that can handle wire antennas on a semi-infinite dielectric material. The integral equation is reduced to a set of linear equations by the method of moments. For efficiency, the impedance matrix element Z_m,n is divided into two parts on the basis of weighted Green's function extractions. The far-zone radiation field, which is formulated using the stationary phase method, is also described. After the validity of the presented numerical techniques is checked using a bow-tie antenna, a spiral antenna is analyzed. The current distribution, radiation pattern, axial ratio, power gain, and input impedance are discussed. It is found that the radiation field inside a dielectric material is circularly polarized. As the relative permittivity of the dielectric material increases, the angle coverage over which the axial ratio is less than 3 dB becomes narrower     

FIR滤波器设计:基于进化规划的频率采样技术

本文介绍了进化规划在要样技术中的应用,结合ＦＩＲ数字低通、带通滤波器设计的两个例子,给出了算法实现的具体操作步骤和实验结果。实验数据表明采用进化规划确定的频率过渡带样本值是最优的,设计的ＦＩＲ滤波器的性能优于查表法。     

基于Inagaki模式方法分析导体内谐振特性

当应用电场积分方程或磁场积分方程对导体散射特性进行矩量法分析时,在某些离散的频率点即内谐振点上,常常出现解的不稳定或不唯一情况。为了解决这一问题,该文提出了一种新型的消除内谐振的方法。这种方法基于电场积分方程,利用Inagaki模性质有效地去除了谐振模式,获得内谐振条件下正确的导体散射特性。该方法具有概念清晰和计算简便等优点。计算结果与公开发表的文献结果以及解析解相比,一致性良好。