ON USING NON-LINEAR CANONICAL CORRELATION ANALYSIS FOR VOICE CONVERSION BASED ON GAUSSIAN MIXTURE MODEL

Voice conversion algorithm aims to provide high level of similarity to the target voice with an acceptable level of quality.The main object of this paper was to build a nonlinear relationship between the parameters for the acoustical features of source and target speaker using Non-Linear Canonical Correlation Analysis(NLCCA) based on jointed Gaussian mixture model.Speaker indi-viduality transformation was achieved mainly by altering vocal tract characteristics represented by Line Spectral Frequencies(LSF).To obtain the transformed speech which sounded more like the target voices,prosody modification is involved through residual prediction.Both objective and subjective evaluations were conducted.The experimental results demonstrated that our proposed algorithm was effective and outperformed the conventional conversion method utilized by the Minimum Mean Square Error(MMSE) estimation.

Multimodal process monitoring based on transition-constrained Gaussian mixture model

Reliable process monitoring is important for ensuring process safety and product quality.A production process is generally characterized bymultiple operation modes,and monitoring thesemultimodal processes is challenging.Most multimodal monitoring methods rely on the assumption that the modes are independent of each other,which may not be appropriate for practical application.This study proposes a transition-constrained Gaussian mixture model method for efficient multimodal process monitoring.This technique can reduce falsely and frequently occurring mode transitions by considering the time series information in the mode identification of historical and online data.This process enables the identified modes to reflect the stability of actual working conditions,improve mode identification accuracy,and enhance monitoring reliability in cases of mode overlap.Case studies on a numerical simulation example and simulation of the penicillin fermentation process are provided to verify the effectiveness of the proposed approach inmultimodal process monitoring with mode overlap.

An Improved Moving Object Detection Algorithm Based on Gaussian Mixture Models

Aiming at the problems that the classical Gaussian mixture model is unable to detect the complete moving object, and is sensitive to the light mutation scenes and so on, an improved algorithm is proposed for moving object detection based on Gaussian mixture model and three-frame difference method. In the process of extracting the moving region, the improved three-frame difference method uses the dynamic segmentation threshold and edge detection technology, and it is first used to solve the problems such as the illumination mutation and the discontinuity of the target edge. Then, a new adaptive selection strategy of the number of Gaussian distributions is introduced to reduce the processing time and improve accuracy of detection. Finally, HSV color space is used to remove shadow regions, and the whole moving object is detected. Experimental results show that the proposed algorithm can detect moving objects in various situations effectively.

Retraction control of motorized seat belt system with linear state observer

A design and verification of linear state observers which estimate state information such as angular velocity and load torque for retraction control of the motorized seat belt (MSB) system were described. The motorized seat belt system provides functions to protect passengers and improve passenger's convenience. Each MSB function has its own required belt tension which is determined by the function's purpose. To realize the MSB functions, state information, such as seat belt winding velocity and seat belt tension are required. Using a linear state observer, the state information for MSB operations can be estimated without sensors. To design the linear state observer, the motorized seat belt system is analyzed and represented as a state space model which contains load torque as an augmented state. Based on the state space model, a linear state observer was designed and verified by experiments. Also, the retraction control of the MSB algorithm using linear state observer was designed and verified on the test bench. With the designed retraction control algorithm using the linear state observer, it is possible to realize various types of MSB functions.

Gaussian mixture model based adaptive control for uncertain nonlinear systems with complex state constraints

This paper addresses an uncertain nonlinear control system problem with complex state constraints and mismatched uncertainties.A novel Gaussian Mixture Model(GMM)based adaptive PID-Nonsingular Terminal Sliding Mode Control(NTSMC)(GMM-adaptive-PID-NTSMC)method is proposed.It is achieved by combining a GMM based adaptive potential function with a novel switching surface of PID-NTSMC.Next,the stability of the closed-loop system is proved.The main contribution of this paper is that the GMM method is applied to obtain the analytic description of the complex bounded state constraints,ensuring that the states'constraints are not violated with GMM-based adaptive potential function.The developed potential function can consider the influence of uncertainties.More importantly,the GMM-adaptive-PID-NTSMC can be generalized to control a more representative class of uncertain nonlinear systems with constrained states and mismatched uncertainties.In addition,the proposed controller enhances the robustness,and requires less control cost and reduces the steady state error with respect to the Artificial Potential Function based Nonsingular Terminal Sliding Mode Control(APF-NTSMC),GMM-NTSMC and GMM-adaptive-NTSMC.At last,numerical simulation is performed to validate the superior performance of the proposed controller.

MOTION ARTIFACT REDUCTION IN FUNCTIONAL NEAR INFRARED SPECTROSCOPY SIGNALS BY AUTOREGRESSIVE MOVING AVERAGE MODELING BASED KALMAN FILTERING

Functional near infrared spectrosecopy(NIRS)is a technique that is used for noninvasive measurement of the oxyhemoglobin(HbO_(2))and deoxyhemoglobin(HHb)concentrations in the brain tissue.Since the ratio of the concentration of these two agents is correlated with the neuronal activity,ONIRS can be usod for the monitoring and quantifying the cortical activity.The portability of NIRS makes it a good candidate for studies involving subject's movement.The NIRS measurements,however,are sensitive to artifacts generated by subject's head motion.This makes fNIRS signals less effective in such applications.In this paper,the autoregressive moving average(ARMA)modeling of the NIRS signal is proposed for state-space representation of the signal which is then fed to the Kalman filter for estimating the motionless signal from motion corrupted signal.Results are compared to the autoregressive model(AR)based approach,which has been done previously,and show that the ARMA models outperform AR models.We attribute it to the richer structure,containing more terms indeed,of ARMA than AR.We show that the signal to noise ratio(SNR)is about 2 dB higher for ARMA based method.

Comparison of Khasi Speech Representations with Different Spectral Features and Hidden Markov States

In this paper,we present a comparison of Khasi speech representations with four different spectral features and novel extension towards the development of Khasi speech corpora.These four features include linear predictive coding(LPC),linear prediction cepstrum coefficient(LPCC),perceptual linear prediction(PLP),and Mel frequency cepstral coefficient(MFCC).The 10-hour speech data were used for training and 3-hour data for testing.For each spectral feature,different hidden Markov model(HMM)based recognizers with variations in HMM states and different Gaussian mixture models(GMMs)were built.The performance was evaluated by using the word error rate(WER).The experimental results show that MFCC provides a better representation for Khasi speech compared with the other three spectral features.

Model Reduction of 2-D IIR Filters

The work presented in this paper concerns with analysis and synthesis of the two-dimensional Infinite Impulse Response (IIR) filters based on model order reduction. The synthesis is performed with two methods, the Prony's method (Prony modified) and Iterative method, in the spatial domain, and with the method of Semi-Definite iterative Programming (SDP), in the frequency domain. After synthesis, we make an order reduction of the filter model by the Quasi-Gramians method.

基于最优控制理论的国产光抽运小铯钟频率控制算法

原子钟频率控制是时间保持工作中的关键技术.当前守时工作中的频率控制主要针对国外微波钟采用开环控制算法,但由于国产光抽运小铯钟(下称国产钟)的工作原理和性能不同于国外同类型原子钟,因此该算法不能很好适应国产钟.为了提升我国标准时间的自主性和安全性,本文基于国产钟的噪声特性,在最优控制理论的框架下研究了线性二次高斯控制算法,该算法属于闭环控制算法,从同步时间、频率控制准确度和频率控制稳定度方面研究国产钟性能,最后分析了不同控制间隔对国产钟性能的影响.结果表明随着二次损失函数中约束矩阵W_(R)的增大,同步时间延长,控制准确度降低,控制短期稳定度提高.W_(R)相同情况下,随着控制间隔的增大,同步时间延长,控制准确度降低,控制短期稳定度提高,对于W_(R)=1时,控制间隔为1 h的同步时间为5小时,控制准确度为1.83 ns,1 h的Allan偏差为1.81×10^(-13);控制间隔为8 h的同步时间为28 h,控制准确度为4.48 ns,1 h的Allan偏差为1.48×10^(-13).控制国产光抽运小铯钟的中长期稳定度都得到提高.

电力电子化电力系统多时间尺度时变动态小干扰稳定问题

面向国家能源安全和双碳战略的重大需求,大规模多样化的电力电子装备以不同的形式和功能参与到现代电力系统源-网-荷各环节中,受多样化电力电子装备多时间尺度时变动态影响的电力电子化电力系统稳定机制发生根本性改变。近年来,国内外电力系统中持续出现与电力电子装备相关的各种机理不明的振荡事故,表现为振荡频率范围较宽的多时间尺度时变动态问题,严重威胁电力电子化电力系统的安全稳定运行。该文从多样化电力电子装备原始特征出发,归纳电力电子化电力系统多时间尺度时变动态基本特性,并分别从小干扰建模和动态稳定分析归纳对不同理论方法的基本认识,随后从适用于电力电子化电力系统多时间尺度时变动态小干扰稳定快速分析计算为根本目标,凝练基于线性周期时变理论小干扰建模和动态稳定直接分析的研究思路。

多种残差补偿的贝叶斯网络下的短期交通预测

为了解决道路车流量的数据生成条件时变场景下的交通预测问题,本文建立道路交通控制与交通流预测数据之间的联系,提出一种基于多种残差补偿的贝叶斯网络的短期交通预测方法。提取城市中大规模多路口主干道车道及车辆信息构造多个平行的贝叶斯网络,使用贝叶斯关系及期望最大化算法进行短期交通预测。再通过数据自相关残差补偿、车辆换道和多路口连通性的线性残差补偿提高了预测的精度,解决了传统研究对相邻路口和换道导致的误差等因素处理能力不足的问题。仿真结果表明:使用贝叶斯网络预测交通流,并基于车辆行为的残差进行精度补偿,可以更准确地预测复杂的交通演化场景的短期交通流。

基于轮轨定位数据的有轨电车区间驾驶特征分析

为分析人工驾驶条件下有轨电车区间速度及可靠性特征,基于轮轨定位数据,计算有轨电车在加速段、巡航段、制动段和交叉口的运行特征指标,分析人工驾驶决策对各指标的影响机制;并建立区间运行速度的多因素回归分析模型及概率分布模型。结果表明:由于人工驾驶的模糊控制特点,司机无法实现充分加减速;终点速度和制动系数对区间运行速度贡献度总占比达57%,是驾驶行为优化的重点;区间运行速度呈高斯混合分布(Gaussian Mixture Model,GMM),对常见绿波带宽有较高的偏出率,是造成线路时间可靠性低的重要原因。

基于高斯混合模型的碳酸盐岩储层物性与孔隙参数地震岩石物理同步反演

碳酸盐岩储层是我国油气资源增储上产的重要领域,然而此类储层通常发育复杂的孔隙结构,极易影响物性参数地震预测结果的精度.本文提出一种基于高斯混合模型的物性与孔隙参数地震岩石物理同步反演方法.基于微分有效介质模型和Gassmann方程,推导定量关联孔隙度、流体饱和度及孔隙纵横比与弹性参数的线性正演算子.引入了高斯混合模型表征物性与孔隙参数的联合先验概率分布,进而表征岩相统计差异特征.基于贝叶斯线性反演理论及线性正演算子,构建目标参数后验概率的解析表达式,然后根据测井数据反演井旁孔隙纵横比,以提供可靠的孔隙参数先验约束,并利用迭代贝叶斯反演算法,提高线性正演算子的模拟精度.本方法将孔隙纵横比作为反演目标参数,表征碳酸盐岩孔隙结构的空间变化特征,结合贝叶斯线性反演方法,以提高物性参数地震反演结果的精度与运算效率.数据测试表明,所提出的方法物性参数反演结果对比常规方法精度有明显改善,且在一定程度上,新方法受初始模型和弹性参数的影响较小.三维地震数据的应用验证了本方法的有效性,其孔隙度反演结果能够有效的指示有利储层的空间分布.

基于区域上下文感知的前景提取

为降低图像颜色分布和弱边缘对前景提取的负面影响,提出一种基于区域上下文感知的前景提取模型。结合图像亮度梯度和颜色信息将图像分成互不相交的区域,区域内颜色具有同质性,近邻区域间存在显著差异;依据颜色分布设计区域相似性测度,构建区域上下文关系和感知空间;在感知空间中将前景提取转化为二分类问题。实验结果表明,相对于传统模型,该算法提高了弱边缘、颜色非均匀分布图像的前景提取效果。

Identification of the state-space model and payload mass parameter of a flexible space manipulator using a recursive subspace tracking method

The on-orbit parameter identification of a space structure can be used for the modification of a system dynamics model and controller coefficients. This study focuses on the estimation of a system state-space model for a two-link space manipulator in the procedure of capturing an unknown object, and a recursive tracking approach based on the recursive predictor-based subspace identification(RPBSID) algorithm is proposed to identify the manipulator payload mass parameter. Structural rigid motion and elastic vibration are separated, and the dynamics model of the space manipulator is linearized at an arbitrary working point(i.e., a certain manipulator configuration).The state-space model is determined by using the RPBSID algorithm and matrix transformation. In addition, utilizing the identified system state-space model, the manipulator payload mass parameter is estimated by extracting the corresponding block matrix. In numerical simulations, the presented parameter identification method is implemented and compared with the classical algebraic algorithm and the recursive least squares method for different payload masses and manipulator configurations. Numerical results illustrate that the system state-space model and payload mass parameter of the two-link flexible space manipulator are effectively identified by the recursive subspace tracking method.

基于拓展控制集的PMSM有限控制集无模型预测电流控制策略

永磁同步电机(permanent magnet synchronous motor,PMSM)具有高效率、高功率密度与高可靠性等优势,已在工业界得到广泛应用。文中针对PMSM驱动系统,提出基于拓展控制集的有限控制集无模型预测电流控制(finite-control-set model-free predictive current control,FCS-MFPCC)。首先,分析PMSM系统的数学模型并详述有限控制集模型预测电流控制(finite-control-set model predictive current control,FCS-MPCC)的原理。其次,介绍基于线性扩张状态观测器(linear extended state observer,LESO)的传统FCS-MFPCC。针对传统FCS-MFPCC稳态性能不足的问题,采用基于离散空间矢量调制(discrete space vector modulation,DSVM)的控制集拓展方案,将控制集的电压矢量数目拓展至25。然后,为解决拓展控制集带来的高计算量问题,提出一种快速寻优策略,阐述该策略的实施原理与流程。最后,基于一台500 W PMSM实验平台,对比传统FCS-MFPCC与所提FCS-MFPCC的控制性能,验证所提算法的有效性与优越性。实验结果表明,所提算法能够有效提升系统稳态性能,且定子绕组电流总谐波畸变率由10.07%降低至6.48%。

基于能量场的城市道路车辆交互强度研究

由于交通环境的复杂性、车辆之间的交互多变性,无人驾驶的行为决策研究一直存在应用灵活性较差、精度不高的瓶颈。因此,以探究车辆之间的交互机理为目标,深入研究车流运行的微观规律,为无人驾驶的行为决策研究提供理论依据。首先,基于能量场的物理性质,将车辆类比于引力场中的场源,建立车辆的行车交互场模型,并依据能量场中的质量和车辆物理特性定义了交互虚拟质量,例如速度、加速度、车型。采用行车交互力模型量化车辆之间的交互程度,依据临界车头间距与多普勒效应分别重新定义交通环境中的距离和速度。其次,利用实车在上海浦东新区同顺大道采集速度、相对距离等相关数据,通过视频处理软件Kinovea提取行车记录仪中的数据,采用高斯混合模型验证行车交互力与实际驾驶行为之间的关系。最后,通过K-means算法对行车交互力聚类分析,量化车辆行驶风险等级。结果表明:通过高斯混合模型计算出的跟驰、换道分类结果与实际分类结果的误差分别为1.12%和9.1%,说明行车交互力对车辆交互的量化描述能力较好;同时,在行车交互力模型的基础上将行车风险分为4级,有效评估了行车风险。本研究提出的行车交互力模型不仅可以拓展以往行车安全场的应用范围,也为驾驶行为决策研究提供一定的理论支撑。

基于自适应GMM阶数与混合特征的说话人识别研究

针对高斯混合模型(GMM)阶数选取缺陷和说话人特征信息不足的问题,提出了基于自适应GMM阶数和多种语音特征融合的说话人识别算法。首先,通过提取梅尔频率倒谱系数(MFCC)和线性预测梅尔频率倒谱系数(LPMFCC),并根据Fisher准则得到一个17维的MFCC和LPMFCC参数组合的混合特征参数,以增强说话人的特征信息。然后,根据自适应思想,在K-means聚类算法中计算簇内误差平方和(SSE)。最后,通过肘部法则自适应调整K值,以获得一个最优GMM阶数,使得系统在已有的声纹特征下获得最优的识别效果。结果表明,该算法不仅完善了说话人的特征信息,并且克服了对GMM阶数选取的缺陷。最终结合LPCC和MFCC两种特征算法,融合得到的混合特征LPMFCC+MFCC的识别率相比于LPCC和MFCC提升了26.34%和12.34%。

基于并发模型的LAC在线故障检测

针对线性模拟电路的在线故障检测进行了研究。首先,通过对并发测试问题的分析及建模要求,从线性模拟电路的网表描述得到可用于残差生成的扩展状态空间模型,这种扩展的状态空间模型定义了电路的变量/方程依赖矩阵即因果依赖矩阵,进而从因果依赖矩阵可得到电路系统的标称模型和未知输入及故障模型。然后,基于无差拍观测器方法采用扩展的状态模型来生成最优故障检测器的电路方程,从而实现故障的在线残差检测。仿真实验结果表明,与传统的线性模拟电路的离线和在线故障检测方法相比,所提出的方案具有更好的故障检测能力,适用于任何线性模拟电路。

Monte Carlo Likelihood Estimation of Mixed-Effects State Space Models with Application to HIV Dynamics

The statistical inference for generalized mixed-effects state space models （MESSM） are investigated when the random effects are unknown. Two filtering algorithms are designed both of which are based on mixture Kalman filter. These algorithms are particularly useful when the longitudinal ts are sparse. The authors also propose a globally convergent algorithm for parameter estimation of MESSM which can be used to locate the initial value of parameters for local while more efficient algorithms. Simulation examples are carried out which validate the efficacy of the proposed approaches. A data set from the clinical trial is investigated and a smaller mean square error is achieved compared to the existing results in literatures.