AN IMPROVED ALGORITHM OF GMM VOICE CONVERSION SYSTEM BASED ON CHANGING THE TIME-SCALE

This paper improves and presents an advanced method of the voice conversion system based on Gaussian Mixture Models(GMM) models by changing the time-scale of speech.The Speech Transformation and Representation using Adaptive Interpolation of weiGHTed spectrum(STRAIGHT) model is adopted to extract the spectrum features,and the GMM models are trained to generate the conversion function.The spectrum features of a source speech will be converted by the conversion function.The time-scale of speech is changed by extracting the converted features and adding to the spectrum.The conversion voice was evaluated by subjective and objective measurements.The results confirm that the transformed speech not only approximates the characteristics of the target speaker,but also more natural and more intelligible.

Data point selection for weighted least square fitting of cavity decay time constant

For the accurate extraction of cavity decay time, a selection of data points is supplemented to the weighted least square method. We derive the expected precision, accuracy and computation cost of this improved method, and examine these performances by simulation. By comparing this method with the nonlinear least square fitting （NLSF） method and the linear regression of the sum （LRS） method in derivations and simulations, we find that this method can achieve the same or even better precision, comparable accuracy, and lower computation cost. We test this method by experimental decay signals. The results are in agreement with the ones obtained from the nonlinear least square fitting method.

A redundant subspace weighting procedure for clock ensemble

A redundant-subspace-weighting(RSW)-based approach is proposed to enhance the frequency stability on a time scale of a clock ensemble.In this method,multiple overlapping subspaces are constructed in the clock ensemble,and the weight of each clock in this ensemble is defined by using the spatial covariance matrix.The superimposition average of covariances in different subspaces reduces the correlations between clocks in the same laboratory to some extent.After optimizing the parameters of this weighting procedure,the frequency stabilities of virtual clock ensembles are significantly improved in most cases.

Time-Spectral Solution of Initial-Value Problems—Subdomain Approach

Temporal and spatial subdomain techniques are proposed for a time-spectral method for solution of initial-value problems. The spectral method, called the generalised weighted residual method (GWRM), is a generalisation of weighted residual methods to the time and parameter domains [1]. A semi-analytical Chebyshev polynomial ansatz is employed, and the problem reduces to determine the coefficients of the ansatz from linear or nonlinear algebraic systems of equations. In order to avoid large memory storage and computational cost, it is preferable to subdivide the temporal and spatial domains into subdomains. Methods and examples of this article demonstrate how this can be achieved.

A Simple Additive Weighting Method for Time-SeriesMultiindices Decision Making and Its Applications

This paper takes the evaluation of overall economic benefit by an example and proposes a simple additive weighting method for time-series multiindices decision making. The method can automatically determine the weight coefficients among the multiindices and the years respectively and it also can obtain the objective evaluation results and conclusions.

A New Time-Aware Collaborative Filtering Intelligent Recommendation System

Aiming at the problem that the traditional collaborative filtering recommendation algorithm does not fully consider the influence of correlation between projects on recommendation accuracy,this paper introduces project attribute fuzzy matrix,measures the project relevance through fuzzy clustering method,and classifies all project attributes.Then,the weight of the project relevance is introduced in the user similarity calculation,so that the nearest neighbor search is more accurate.In the prediction scoring section,considering the change of user interest with time,it is proposed to use the time weighting function to improve the influence of the time effect of the evaluation,so that the newer evaluation information in the system has a relatively large weight.The experimental results show that the improved algorithm improves the recommendation accuracy and improves the recommendation quality.

DECAY ESTIMATES OF PLANAR STATIONARY WAVES FOR DAMED WAVE EQUATIONS WITH NONLINEAR CONVECTION IN MULTI-DIMENSIONAL HALF SPACE

This paper is concerned with the initial-boundary value problem for damped wave equations with a nonlinear convection term in the multi-dimensional half space R n ＋ ： u tt u ＋ u t ＋ divf （u） = 0, t 〉 0, x = （x 1 , x ′ ） ∈ R n ＋ ：= R ＋ × R n 1 , u（0, x） = u 0 （x） → u ＋ , as x 1 → ＋ ∞ , u t （0, x） = u 1 （x）, u（t, 0, x ′ ） = u b , x ′ = （x 2 , x 3 , ··· , x n ） ∈ R n 1 . （I） For the non-degenerate case f ′ 1 （u ＋ ） 〈 0, it was shown in [10] that the above initialboundary value problem （I） admits a unique global solution u（t, x） which converges to the corresponding planar stationary wave φ（x 1 ） uniformly in x 1 ∈ R ＋ as time tends to infinity provided that the initial perturbation and/or the strength of the stationary wave are sufficiently small. And in [10] Ueda, Nakamura, and Kawashima proved the algebraic decay estimates of the tangential derivatives of the solution u（t, x） for t → ＋ ∞ by using the space-time weighted energy method initiated by Kawashima and Matsumura [5] and improved by Nishihkawa [7]. Moreover, by using the same weighted energy method, an additional algebraic convergence rate in the normal direction was obtained by assuming that the initial perturbation decays algebraically. We note, however, that the analysis in [10] relies heavily on the assumption that f ′ （u） 〈 0. The main purpose of this paper isdevoted to discussing the case of f ′ 1 （u b ） ≥ 0 and we show that similar results still hold for such a case. Our analysis is based on some delicate energy estimates.

Real-Time Update of Sequence Placement Logic for High Arch Dams Based on Evidence Weight Discount

Sequence placement logic plays a significant role in construction simulation of high arch dams and directly affects the simulation process and results. To establish a sequence logic for dam block placement, the construction scheme, real-time construction process, and random factors of the site all need to be considered in detail. There are few studies available currently that take all these factors into consideration. To address this problem, a real-time update of sequence placement logic for high arch dams based on evidence weight discount is proposed in this study. First, the subjective weight of the dam block sequence priority criteria is built using a consistent matrix method based on the construction scheme. Second, using evidence theory, dynamic objective weight of the priority criteria and basic probability assignment is built. Finally, using a weight self-adaptive adjustment method and comprehensive evidence discounting, the placing probabilities of different dam blocks are obtained. A case study indicates that this method can realize real-time update of sequence placement logic. © 2017, Tianjin University and Springer-Verlag Berlin Heidelberg.

THE VLASOV-POISSON-BOLTZMANN SYSTEM NEAR MAXWELLIANS FOR LONG-RANGE INTERACTIONS

In this article, we are concerned with the construction of global smooth small-amplitude solutions to the Cauchy problem of the one species Vlasov-Poisson-Boltzmann system near Maxwellians for long-range interactions. Compared with the former result obtained by Duan and Liu in [12] for the two species model, we do not ask the initial perturbation to satisfy the neutral condition and our result covers all physical collision kernels for the full range of intermolecular repulsive potentials.

Application of Data from a Lightning Location System in Huai'an Region

Based on data of cloud-to-ground lightning detected by a lightning location system in Jiangsu Province from 2007 to 2012and data of thunderstorm days from seven meteorological observation stations in Huai'an during 1961-2012,the spatial and temporal distribution characteristics of cloud-to-ground lightning in Huai'an were analyzed.The results show that the total frequency of cloud-to-ground lightning in one day was closely related to the intensity and frequency of thunder and lightning.The seasonal variations in the total frequency of cloud-to-ground lightning in Huai'an.The total frequency was shown in unimodal distribution,and the extreme values always appeared in July or August.Most cloud-to-ground lightning was negative,and its frequency accounted for about95.3% of the total frequency.Cloud-to-ground lightning density was high in the south of Huai'an and low in the north,and the serious and very serious ground flash zones concentrated in the south of Hongze County,and most parts of Xuyi and Jinhu County,so lightning protection should be an important task particularly in these regions.

DEGENERATE BOUNDARY LAYER SOLUTIONS TO THE GENERALIZED BENJAMIN-BONAMAHONY-BURGERS EQUATION

This paper is concerned with the convergence rates of the global solutions of the generalized Benjamin-Bona-Mahony-Burgers（BBM-Burgers） equation to the corresponding degenerate boundary layer solutions in the half-space.It is shown that the convergence rate is t-（α/4） as t →∞ provided that the initial perturbation lies in H α 1 for α 〈 α（q）：= 3 ＋（2/q）,where q is the degeneracy exponent of the flux function.Our analysis is based on the space-time weighted energy method combined with a Hardy type inequality with the best possible constant introduced in [1]

单相并网储能双向变流器线性时间周期模型的稳定性分析

针对传统线性时不变模型无法描述储能双向变流器多频次谐波间交互耦合现象的问题,提出了一种单相并网储能双向变流器的线性时间周期模型,并进行稳定性分析。基于线性时间周期理论对单相并网储能双向变流器进行建模,得到非线性时间周期模型;围绕周期轨道将该模型线性化,得到储能双向变流器的线性时间周期时域模型;通过傅里叶展开推出了储能双向变流器的线性时间周期频域模型;利用李雅普诺夫第一方法分析比例积分控制器在不同参数下的储能双向变流器小信号模型的稳定性及稳定边界。MATLAB/Simulink仿真实验结果表明:建立的单相并网储能双向变流器的线性时间周期模型具有准确性和稳定性,稳定性边界正确。

中国发展方式绿色转型的统计测度及时空演进分析

加快发展方式绿色转型,是贯彻落实新发展理念的战略要求,是践行绿色发展理念的时代命题。评估中国发展方式绿色转型进程,对于实现“双碳”目标、推动经济社会高质量发展具有重要现实意义。本文采用熵权法,选取2011~2022年我国30个省(区、市)的面板数据,对各省(区、市)发展方式绿色转型水平进行测度并分析其时空演变特征和地区差异。研究发现:(1)中国发展方式绿色转型进程取得了阶段性成果,东部地区发展方式绿色转型程度最高;(2)东部地区与其他三大地理区域的发展方式绿色转型水平差异显著,其中东部地区与东北地区的差异最大;(3)发展方式绿色转型是一个渐进式过程,存在稳定的β收敛特征。研究结论对于缩小地区间发展方式绿色转型差异、加快发展方式绿色转型进程具有参考意义。

一种改进组合加权的TDOA室内二维定位算法

针对现有组合加权算法对定位区域边缘的目标定位时精度较低的问题,在现有组合加权算法的基础上提出了一种改进算法。首先,将基站分组,以到达时差算法得到目标位置的多个估计结果;其次,计算各估计结果之间距离值并排序,以滑动窗口法判断是否存在基站组出现异常定位估计;最后,当任意基站组的定位结果发生异常时,使用目标位置估计结果及其估计克拉美罗下界值设计两个加权步骤的权值,通过二步组合加权算法得到最终定位结果。仿真结果表明,所提算法有效减少了原组合加权算法对定位区域边缘的目标定位时的误差,当测量噪声标准差为0.8 m时,所提算法相较于原算法在正方形边缘区域定位均方根误差减小了0.35 m;在定位狭窄矩形区域时,所提算法平均定位均方根误差减小了0.11 m。

A simple data assimilation method for improving the MODIS LAI time-series data products based on the object analysis and gradient inverse weighted filter

A simple data assimilation method for improving estimation of moderate resolution imaging spectroradiometer （MODIS） leaf area index （LAI） time-series data products based on the gradient inverse weighted filter and object analysis is proposed. The properties and quality control （QC） of MODIS LAI data products are introduced. Also, the gradient inverse weighted filter and object analysis are analyzed. An experiment based on the simple data assimilation method is performed using MODIS LAI data sets from 2000 to 2005 of Guizhou Province in China.

电性源瞬变电磁法频-时转换的效果分析

电性源短偏移距瞬变电磁法是近年来新提出的固体矿产深部电磁探测方法之一,较传统的回线源瞬变电磁具有探测深度深、仪器轻便和分辨率高等优点。但该方法由于偏移距较短,距发射源较近,正演中的频-时转换会受到转换方法及其他工作参数的影响,进而直接影响探测结果,因此有必要开展频-时转换效果分析评价。通过对影响频-时转换的因素以及常用的四种频-时转换方法(Guptasarma线性滤波法、折线逼近法、余弦变换的数值滤波法和正弦变换的数值滤波法)在电性源短偏移距瞬变电磁正演响应计算中的对比,从计算时间和转换精度两方面分析了在不同条件下的转换效果,为一维及高维度的瞬变电磁正演计算提供频-时转换方法参考。对比结果表明:1)影响时间域瞬变电磁场正演计算(即频-时转换效果)的精度不仅和转换方法本身有关,而且也和偏移距及地下介质电阻率等参数有关;2)通过正演模拟,在低阻探测区采用折线逼近法与正弦变换的数值滤波法相对Guptasarma线性滤波法和余弦变换的数值滤波法转换精度更高,整体误差更小,而Guptasarma线性滤波法的计算速度最快,早期计算精度较高,晚期计算精度较差;在高阻探测环境中,采用折线逼近法计算精度和速度均相对较好,Guptasarma线性滤波法、正弦变换的数值滤波法及余弦变换的数值滤波转换效果均较差,特别是在晚期。因此,对于未知的探测区域,若实现高精度的电性源瞬变电磁正演计算,采用折线逼近法进行频-时转换相对效果最佳。

未知信号传播速度的TDOA定位方法研究

研究了未知信号传播速度时的到达时间差(Time-Difference-Of-Arrival,TDOA)定位问题,提出两种联合估计信号传播速度和目标位置的定位方法。第一种方法为两步加权最小二乘方法。在该方法中,首先不考虑变量之间关系,得到一个关于未知变量的初始加权最小二乘估计。为改进第一步估计的性能,第二步考虑第一步估计中变量之间的关系,将其转换为一个标准的广义信赖域子问题,最终获得更高精度的估计性能。第二种方法为半正定松弛方法,通过构建非线性非凸加权最小二乘问题,然后利用半正定松弛技术将其松弛为凸的半正定规划问题,容易求解。仿真结果表明,所提两种方法均能够在高斯噪声下,且噪声不太大时达到克拉美-罗界。

基于累积位移特征与时间序列组合模型的滑坡位移预测

分析滑坡累积位移演化特征对位移预测具有重要的意义,根据不同增长趋势的位移曲线构造合适的模型来预测位移,将有效提高预测结果的准确度。本文分析三峡库区阶跃型滑坡累积位移变化趋势,将其位移曲线划分为4类特征:等幅-阶跃型、减幅-阶跃型、增幅-阶跃型、复合型,并建立时间序列组合预测模型。以八字门滑坡监测点ZG111及白家包滑坡监测点ZG326为例,依据时间序列原理,采用变分模态分解法(VMD)将累积位移分解为趋势性位移、周期性位移、随机性位移;利用一元线性回归、幂函数非线性回归方法对趋势性位移进行建模分析,预测结果采用加权改进后的最小二乘法(WLS);用麻雀搜索算法(SSA)优化BP神经网络模型,并结合滚动预测的思想,预测周期性位移、随机性位移。最终得到的各位移预测值之和即累积位移预测结果,结果表明:趋势性位移预测MAPE分别为1.2%、0.77%;周期性位移、随机性位移拟合效果较好,预测结果能较好的符合位移整体变化趋势;累积位移预测MAPE在2%以内,预测结果与实际值具有良好的一致性。本文提出的预测模型满足预测精度的要求,能完成滑坡将来位移量的预测,具有较强的工程实用价值,为滑坡灾害预测和防治方面的研究工作提供指导。

基于PANGU程序的球床式高温气冷堆中子动力学参数计算

有效缓发中子份额(β_(eff))等中子动力学参数是核反应堆动态分析的基础。获取中子动力学参数的方法包括实验测量与数值计算,而对于球床式高温气冷堆,现有的条件还不能够直接进行实验测量,因此,数值计算是唯一现实可取的方法。本文研制了包含缓发中子数据的多群截面数据库XPZLIB,基于高温气冷堆栅格物理程序XPZ,对缓发中子数据进行了均匀化和能群归并处理,基于堆芯物理分析程序PANGU,使用重要性函数加权方法计算了高温气冷堆中子动力学参数,并利用蒙特卡罗程序检验了计算结果的准确性。结果证明,数值计算方法可以用于计算高温气冷堆动力学参数,并能为其动力学分析提供基础。

一种智能手机GNSS精密定位随机模型构建方法及定位性能分析

全球导航卫星系统(GNSS)定位芯片已经成为智能手机中实现导航定位的重要传感器,随着信息化、智能化社会的不断发展和深入,人们对手机高精度定位服务的需求不断增加。在分析伪距残差与高度角、载噪比(C/N_(0))及手机GNSS原始观测信息的相关性时发现:相比于C/N_(0),手机GNSS原始观测信息中的ReceivedSvTimeUncertaintyNanos与伪距残差的相关性更高。因此,提出了一种基于岭回归和智能手机原始观测信息的GNSS精密定位随机模型构建方法。通过构建以手机原始观测信息为特征变量的岭回归模型,利用谷歌公开数据集训练该模型,学习伪距残差与各种特征变量之间的映射关系,进而在线预测伪距噪声,以此作为GNSS定位中观测值定权的依据。最后,进行实时动态定位(RTK)和精密单点定位(PPP)实验以验证新定权方法的定位性能,实验结果表明,相比传统C/N_(0)定权方法,新定权方法的定位精度在3组实验中平均提升幅度分别为24.9%,44.8%及39.3%。