改进的“当前”统计模型变采样率目标跟踪算法

为满足实际雷达系统对高精度和高实时性的要求,提出了一种改进的"当前"统计模型变采样率机动目标跟踪算法。该算法针对"当前"统计模型必须预设加速度极值和机动频率的问题,提出一种加速度方差和机动频率在线同步自适应方法,建立改进的"当前"统计模型机动目标跟踪算法;针对在线自适应方法计算量大的问题,结合采样周期的大小与目标机动特性的关系,引入变采样率方法。仿真结果表明,与传统"当前"统计模型相比,改进的"当前"统计模型机动目标跟踪算法能显著提高对不同机动强度目标的跟踪精度;变采样率方法通过减少采样点数,节省了系统资源,提高了跟踪实时性;所提算法将两者结合,用传统的"当前"统计模型1.5~2倍的平均采样周期得到了更小的位置均方根误差,实现了用单模型方法同时改善跟踪精度和实时性的目的。     

一种改进的无线定位算法

Taylor级数展开法和Chan算法是两种性能优良的利用电波到达时间差(TDOA)的定位算法,前者简单实用,但是其缺点是对初值比较敏感；后者在视距(LOS)传播环境下有较高的定位精度,二者结合可以大大提高定位的精度。但是在非视距(NLOS)传播环境下Chan算法精度会受到较大的影响,从而影响到Taylor级数展开法的定位精度。本文根据NLOS传播环境下附加传播时延和均方根时延扩展的统计特性,对NLOS误差的均值和方差进行估计,对TDOA测量值进行修正,采用Chan算法计算初值,再利用Taylor级数展开法进行定位,并与其它两种基于Taylor级数展开法的定位方法结果进行了比较。仿真结果表明,该算法能够提高NLOS传播环境下的定位精度,性能优于另外两种基于Taylor级数展开法的定位方法。     

A Novel Method to Select Minimum Neighbors in Cooperative Localization Network

Cooperative localization is an emerging paradigm that circumvents the needs for high-power, high-density anchor deployment, and offers additional positioning accuracy by exchanging information between adjacent agents. Considering that positioning accuracy only need to meet the application requirements, too high positioning accuracy means that too much redundant information is exchanged. This paper presents a neighbors selection algorithm that tries to meet the positioning accuracy required by an application , while the system consumption is optimized. More specifically, this proposed algorithm tries to minimize the number of neighbors involved in cooperation on the basis of the equivalent Fisher information matrix. Before putting forward the algorithm, we first introduce the notion of equivalent Fisher information and characterize localization accuracy called the squared position error bound. Then we find that not only the Fisher information of neighbor nodes determines the positioning accuracy of the agent to be located, but also the direction of Fisher information affected the positioning accuracy. Based on this, an algorithm for selecting the minimum cooperative subset is proposed. Simulation results show that the proposed algorithm can obviously improving the utilization of energy compared to other commonly used methods by exchanging information with a few nodes, while achieving the specified positioning accuracy.

     

基于可见-近红外光谱特征波长选择的土壤有机质快速检测研究

选择光谱特征波长进行建模可以减少冗余波长的干扰,提高模型的预测精度。采用小波阈值消噪法对采集的104个土壤样本光谱数据进行了预处理,并通过间隔偏最小二乘法、无信息变量消除、连续投影算法和群智能算法等9种方法筛选了建模波长。结果表明,小波阈值消噪法能有效降低光谱中的噪声。利用波长选择方法筛选建模波长不仅能减少建模变量的个数,而且还能提高模型的预测精度,特别是离散粒子群优化算法利用26个波长进行建模,预测决定系数达到了0.81,预测的相对标准误差为2.31。实验结果证明,通过对光谱波长进行选择不但可以降低模型的复杂度,还能有效预测土壤有机质达的含量。     

稳健收敛的时差定位迭代算法

为实现目标精确定位,减少初始值对目标定位精度产生的影响,该文提出了一种稳健收敛的时差定位迭代算法。该算法利用最大似然方法确定目标函数,运用牛顿法对目标位置进行迭代求解,对于计算过程中可能出现的病态 Hessian 矩阵,运用正则化理论构造一个对称正定的 Hessian 矩阵,使算法能够稳健高效的运行。实验结果表明：相对于传统牛顿法,本文算法在初始值的选取上具有稳健性,对误差选取较大的初始值,仍能够保证算法的收敛性,相对于现有闭合式定位方法,本文算法在噪声较大时具有较好的定位精度。     

Direction-of-Arrival Method Based on Randomize-Then-Optimize Approach

The direction-of-arrival(DOA) estimation problem can be solved by the methods based on sparse Bayesian learning(SBL). To assure the accuracy, SBL needs massive amounts of snapshots which may lead to a huge computational workload. In order to reduce the snapshot number and computational complexity, a randomizethen-optimize(RTO) algorithm based DOA estimation method is proposed. The “learning” process for updating hyperparameters in SBL can be avoided by using the optimization and Metropolis-Hasti...     

基于二次分割的银行票据彩色印章的滤除

由于彩色印章的存在,传统的印章滤除算法常常会弱化票据中印章的遮挡字符,降低票据识别正确率,这很难满足实际应用的需要。针对这种情况,提出一种基于彩色图像二次分割的算法,该算法能够实现印章的有效滤除。算法首先对票据背景和印章部分进行滤波。然后,基于印章颜色分析进行图像的第二次分割,判断并提取出印章遮挡的字符。根据票据背景平均灰度值自适应的增强遮挡字符,并滤除干扰印章部分,实现遮挡字符细节得以保留,从而提高了票据识别的正确率。实验结果表明,相比于现有的方法,该文采用的方法将识别准确率提高了约10%,实验证明了该方法的有效性,并具有较好的应用前景和商业价值。     

A general approach to network analyzer calibration

A general-purpose algorithm, the TCX algorithm, for network analyzer and test fixture calibration is presented. The algorithm can handle most of the existing calibration methods, including symmetrical test fixtures. Any combination of one-port or two-port standards can be used. There is a possibility of partial self-calibration, if one of the standards is a two-port network or a through connection. The algorithm is applied to get simple equations covering the TSD, LDX (LRL), LAX, and LMX methods (X being an unknown one-port or symmetrical two-port network). A transmission path is allowed between the ports of standard X. In the TSD method, the delay line can be replaced with an attenuation network or with a matched load; also, the through line can have an unknown delay line and attenuation. A method of root choice for LRL and similar methods in conjunction with test fixtures is described. The method of least-squares-fit can be applied, when redundant data are available. It gives an essential improvement of accuracy in the simulation of a symmetrical test fixture     

Optimizing the channel selection and classification accuracy in EEG-based BCI

Multichannel EEG is generally used in brain-computer interfaces (BCIs), whereby performing EEG channel selection 1) improves BCI performance by removing irrelevant or noisy channels and 2) enhances user convenience from the use of lesser channels. This paper proposes a novel sparse common spatial pattern (SCSP) algorithm for EEG channel selection. The proposed SCSP algorithm is formulated as an optimization problem to select the least number of channels within a constraint of classification accuracy. As such, the proposed approach can be customized to yield the best classification accuracy by removing the noisy and irrelevant channels, or retain the least number of channels without compromising the classification accuracy obtained by using all the channels. The proposed SCSP algorithm is evaluated using two motor imagery datasets, one with a moderate number of channels and another with a large number of channels. In both datasets, the proposed SCSP channel selection significantly reduced the number of channels, and outperformed existing channel selection methods based on Fisher criterion, mutual information, support vector machine, common spatial pattern, and regularized common spatial pattern in classification accuracy. The proposed SCSP algorithm also yielded an average improvement of 10% in classification accuracy compared to the use of three channels (C3, C4, and Cz).     

采用动态窗口宽度和二次平均算法的SSOA定位方法

信号的中值估计是到达信号强度(SSOA)定位的核心,其误差主要来源于信道的快衰落和阴影(慢)衰落。信号平均是获得信号中值估计的重要手段,该文提出了二次平均和动态窗口宽度算法以提高中值估计精度,从而达到进一步提高定位精度的目的。计算机仿真结果表明,在衰落信道中,二次平均算法的定位精度比传统方法约高30％。采用动态窗口宽度后,可保证在移动台低速运动时仍具有较高的定位精度。     

基于数据融合理论的多传感器几何定位算法

为解决多部(3部及以上)2D传感器网络对三维空间目标的定位估计和定位精度问题,克服地球曲率对观测模型的影响,建立了考虑实际地球曲率的等效地球模型和传感器观测模型,提出了此模型中基于二次数据融合的多传感器组网几何定位算法,该方法将几何定位与数据融合理论相结合,并对融合数据进行二次融合,充分利用了各传感器的量测数据。仿真实验证明了方法的有效性和实用性,在多部2D传感器组网的情况下可对三维空间内目标实现精确定位,定位误差趋近于克拉美-罗下界(CRLB),具有工程实用价值。     

Fusing different measurements and algorithms to improve RFID localization accuracy

This paper considers the topic of two-dimensional object localization using radio frequency identification (RFID) technology. One of the important problems that arise during the development of RFID positioning systems is choosing a localization algorithm and a type of measurement data. Usually developers try to find such a combination of an algorithm and measurement type that allows to achieve maximal localization accuracy for a particular scenario. However, in some situations there can be several combinations of algorithms and measurements with equally high accuracy. In order to simplify the choosing problem and to additionally increase accuracy it is developed a combinational localization method. The method is based on averaging position estimates formed by several point-based and zone-based algorithms that process different measurements. In our work there are used three point-based and three zone-based algorithms: a k nearest neighbors algorithm, trilateration, intersectional algorithm, the methods of support vector machine, artificial neural networks, and a naive Bayes classifier. As an input for the algorithms we utilized received signal strength, read rate, and proximity measurements. During the experiments we found that our method decreases the mean error by 15% and the maximum error by 14% compared to the best single algorithm.     

基于梳状导频分布的OFDM信道估计改进算法

提出了基于梳状导频分布的OFDM信道估计的改进算法,分别在基于频域(DFT)和时域(IDFT)的迫零内插算法的基础上运用了数据翻转的思想.该算法可以大大地减少DFT/IDFT变换时所产生的边缘效应,并且通过优化边缘位置的估计性能来提升整个OFDM信道估计的精确度.仿真结果表明,该算法相比经典算法在性能上有较大提升.     

Subpixel Target Location Techniques for 3—D Coordinate Measuring System

The close photogrammetric 3-D coordinate meaurement is a new measuring technology in the fields of the coordinate measurement machine (CMM) in recent yearsl.In this method,we usually place some targets on the measured object and take image of targets to determine the object coordinate.The subpixel location of target image plays an important role in high accuracy 3-D coordinate measuring procedure,In this paper,some subpixel location methods are reviewed and some factors which affect location precision are analyzed. Then we propose bilinear interpolation centroid algorithm.The experiments have shown this algorithm can improve accuracy of target centroid by increasing available pixels.     

一种改进的高维度离群点检测方法

离群点检测作为数据挖掘的一个重要研究方向,可以从大量数据中发现少量与多数数据有明显区别的数据对象。高维度环境是离群点检测技术的一个重要场景,现实生活中的高维数据中大量无关或嘈杂的特征给基于子空间/特征选择的高维离群点检测方法提出了重大挑战。Pang等人提出了一种将离群点打分和特征选择结合起来的方案CINFO,准确度相比传统的高维度离群点检测算法有所提升。CINFO方法在效率上有改进空间,本文通过引入扩展的孤立森林算法(Extended Isolation Forest,EIF)对CINFO方法进行改进,在几乎不损失精度的情况下,明显提升了算法效率。     

Filter design and comparison for two fast CWT algorithms

The continuous wavelet transform (CWT) is a powerful technique for signal analysis. Direct CWT computation by FFT requires O(N log₂ N) operations per scale, where N is the data length. The a trous algorithm and the Shensa (1992) algorithm are two fast methods to compute CWT recursively that require only O(N) operations per scale. Both of them can be described by the multiresolution analysis (MRA) structure but with different MRA filters. This paper proposes methods to design the MRA filters of the two algorithms to improve their accuracy on CWT computation. We begin with the formulation of the CWT computation error using the MRA structure. The MRA filters of the two algorithms are then designed to minimize the error. In either algorithm, both the lowpass and bandpass MRA filters can be optimized. The a trous algorithm has closed-form solutions for the two filters. The Shensa algorithm, on the other hand, has an analytic solution for the bandpass filter only. Finding the optimum lowpass filter requires a multidimensional numerical search. Simulation studies show that by using the proposed optimum filters, the Shensa algorithm, in general, outperforms the a trous algorithm     

基于多参数配准模型的脑核磁影像分割算法

配准技术在基于多图谱的分割方法中能有效地将医学图谱的先验知识融入分割过程,再结合以高效的标记融合算法,最终实现精确地自动分割.针对图谱配准的较大误差及其对标记融合的重要影响,本文建立了一种新的概率图模型框架并以此提出了基于多参数配准模型的分割算法,将此方法与高效的标记融合算法相结合,可以提高目标图像中特定组织区域的分割精度,更使其在少量图谱分割的情形下具有重要应用.首先,使用多种配准参数对所有目标图像进行配准;然后,分别采用不同的算法对配准图像进行灰度融合和标记融合,实现训练图像的重构过程;最后,利用高效的标记融合算法对重构后的图像进行融合得到最终精确的分割结果.实验结果表明该方法均优于本文其他分割算法,能够有效提升脑部组织分割精度.     

基于改进贝叶斯分类的Android恶意软件检测

针对Android手机安全受恶意软件威胁越来越严重这一问题,提出一种改进的Android恶意软件检测算法。监控从Android移动设备应用程序获取的多种行为特征值,应用机器学习技术,通过与卡方检验滤波测试结合的方式改进传统的朴素贝叶斯算法,检测Android系统中的恶意软件。通过实验仿真,结果表明在采取朴素贝叶斯分类模型之前,使用卡方检验过滤应用程序的行为特征,可以使基于Android的恶意软件检测技术拥有较低的误报率和较高的精度。     

基于亚像素精度分割的晶粒图像识别方法

机器视觉在工业产品质量控制中的应用越来越广,针对LED晶粒的质量检测问题,提出了一种基于亚像素精度阈值分割的方法进行LED晶粒图像的识别。首先通过图像增强及预处理,然后通过基于亚像素精度的阈值分割方法进行晶粒图像的预处理,再通过K值聚类算法进行感兴趣的区域提取,最后通过NCC归一化的方法进行晶粒位置的识别及定位。实验结果表明,提出的亚像素精度阈值分割和NCC归一化定位算法相结合的方法应用在LED分拣机中,能准确地识别出LED晶粒图像。     

Numerical evaluation of the Lambert W function and application to generation of generalized Gaussian noise with exponent 1/2

We address the problem of synthesizing a generalized Gaussian noise with exponent 1/2 by means of a nonlinear memoryless transformation applied to a uniform noise. We show that this transformation is expressable in terms of a special function known under the name of the Lambert W function. We review the main methods for numerical evaluation of the relevant branch of the (multivalued) Lambert W function with controlled accuracy and complement them with an original rational function approximation. Based on these methods, synthesis of the generalized Gaussian noise can be performed with arbitrary accuracy. We construct a simple and fast evaluation algorithm with prescribed accuracy, which is especially suited for Monte Carlo simulation requiring large numbers of realizations of the generalized Gaussian noise.