基于禁止搜索的非线性时间匹配优化算法

动态时间规整算法DTW(Dynamic Time Warping)作为一种非线性时间匹配技术已成功地应用于语音识别系统中。DTW算法使用动态规划技术来搜索两个时间序列的最优规整路径,虽然这种算法计算量小,运算时间较短,但只是一种局部优化算法。禁止搜索TS(Tabu Search)算法是一种具有短期记忆的广义启发式全局搜索技术,适用于解决许多非线性优化问题。本文将该技术用于语音识别系统中,提出了基于禁止搜索的非线性时间规整的优化算法TSTW,使得时间规整函数尽可能逼近全局最优。仿真结果表明,TSTW比DTW有更高的识别率,且运行时间比遗传时间规整算法GTW大大减少。     

DTW Based Classification of Diverse Pre-Processed Time Series Obtained from Handwritten PIN Words and Signatures

Personal identity verification by means of signature handwriting dynamics is a widely researched aspect of behavioral biometrics. The Dynamic Time Warping (DTW) technique has been successfully used for accessing the similarity of time series of handwritten objects by minimizing non-linear time distortions. Generally, in DTW based classifiers, the sequences are normalized in time and amplitude domains. In the paper, different length and amplitude normalization techniques are applied on signatures and handwritten PIN word sequences and their influence on accuracy of recognition are examined. A special approach to amplitude normalization based on reference level assigned Dynamic Time Warping (DTW) technique is presented. The standard deviation values calculated from the time series are used as so called bio-reference levels to improve the performance of classification. For this, they are added to the time series of query and sample datasets prior to DTW matching. The acquisition of online data is carried out by a digital pen equipped with pressure and inclination sensors. The time series obtained from the pen during handwriting provide valuable insight into the unique characteristics of the writers. Experimental results show that with the help of proposed length and amplitude normalizations of sequences including the bio-reference levels, the computational time is reduced and false acceptance rates are decreased.     

基于自适应遗传算法的星载SAR天线方向图与模糊综合

该文针对改善星载合成孔径雷达(SAR)的模糊特性,提出了一种自适应遗传算法。该算法同时对模糊和方向图进行优化。首先确定模糊区域,然后以天线方向图的主瓣宽度和副瓣电平(包括星载SAR模糊区域的副瓣电平)为目标函数,应用自适应遗传算法对天线方向图进行综合。为了避免早熟的现象,在该算法中,交叉概率、变异概率和变异范围同时进行了自适应的变化。和非自适应遗传算法相比较,该算法迭代步骤少,收敛速度快。仿真结果表明,模糊度得到了很好的抑制,对星载SAR系统设计具有实际意义。     

HMM在语音识别系统中的应用

介绍语音识别技术的应用状况与发展,对基于动态时间伸缩技术、隐含马尔科夫模型及人工神经网络的3种不同的语音识别系统进行了比较,重点介绍了隐含马尔科夫模型(HMM)在语音识别系统中的应用。其中基于HMM的语音识别系统是在UniSpeech芯片上实现基于DHMM的识别系统,然后又在同一平台上实现了基于CHMM的识别系统。     

融合边界信息的语音样例快速检索

提出了一种融合音素边界信息的语音样例快速检索方法。该方法首先提取查询样例和测试集的音素后验概率;然后,运用层次凝聚聚类算法将音素后验概率序列分段（即音素边界检测）,计算每个分段的平均向量并将其分别组成新查询和新索引,再运用动态时间规整进行语音样例的检索;最后,使用虚拟相关反馈技术对检索结果进行修正。实验结果表明:尽管此方法的检索精度略低于直接运用动态时间规整进行检索的检索精度,但其检索速度大大优于后者,且与其他相关文献提出的方法相比,此方法在检索速度方面也具有明显优势。      

Inverse Double NARX Fuzzy Modeling for System Identification

In this paper, a novel inverse double nonlinear autoregressive with exogenous input (NARX) fuzzy model is applied to simultaneously model and identify both joints of the prototype two-axis pneumatic artificial muscle (PAM) robot arm's inverse dynamic model. Highly nonlinear features of both joints of the nonlinear manipulator system are identified by the proposed inverse double NARX fuzzy (IDNF) model based on experimental input–output training data. The modified genetic algorithm (GA) optimally generates the appropriate fuzzy if–then rules to perfectly characterize the dynamic features of the two-axis PAM manipulator system. The evaluation of different IDNF models with various ARX model structures will be discussed. For the first time, the nonlinear IDNF model of the two-axis PAM robot arm is investigated. The results show that the nonlinear IDNF model that is trained by GA performs better and has a higher accuracy than the conventional inverse fuzzy model.      

基于混合遗传算法的唯相位直接数据域算法

提出一种基于混合遗传算法的唯相位直接数据域最小二乘算法.通过采用标准遗传算法与Neider-Mead单纯形法相结合的混合遗传算法,提高了优化效率和运算速度.首先根据标准直接数据域算法推导得出目标函数,继而将目标函数作为适应度函数,将所有自适应权值的未知相位作为决策变量,通过混合遗传算法进行非线性优化,从而求得各个自适应权值的优化解.作为一种唯相位自适应算法,它在硬件实现上比传统算法更具简单性.同时,它只对单快拍数据进行处理,避免了样本协方差矩阵的构造以及矩阵求逆运算,更适合于实时处理.仿真结果表明,算法具有良好的信号恢复和干扰置零性能,比基于非线性其轭梯度法的唯相位直接数据域算法性能更优.     

A robust phase-only Direct Data Domain algorithm based on generalized rayleigh quotient optimization using hybrid Genetic Algorithm

A robust phase-only Direct Data Domain Least Squares （D3LS） algorithm based on generalized Rayleigh quotient optimization using hybrid Genetic Algorithm （GA） is presented in this letter. The optimization efficiency and computational speed are improved via the hybrid GA composed of standard GA and Nelder-Mead simplex algorithms. First, the objective function, with a form of generalized Rayleigh quotient, is derived via the standard D3LS algorithm. It is then taken as a fitness function and the unknown phases of all adaptive weights are taken as decision variables Then, the nonlinear optimization is performed via the hybrid GA to obtain the optimized solution of phase-only adaptive weights. As a phase-only adaptive algorithm, the proposed algorithm is simpler than conventional algorithms when it comes to hardware implementation. Moreover, it proc- esses only a single snapshot data as opposed to forming sample covariance matrix and operating matrix inversion. Simulation results show that the proposed algorithm has a good signal recovery and interferences nulling performance, which are superior to that of the phase-only D3LS algorithm based on standard GA.     

双稳系统中非周期随机共振的数值仿真

本文采用互相关方法研究了噪声和随机二进制信号同时激励双稳系统(施密特触发器和双势阱系统)时的输出响应，观察到非周期随机共振：利用双稳系统中的非周期随机共振效应，可以减小随机信号传输中的噪声水平，改善输出信号质量，这在数字通信领域具有十分重要的意义。     

改进的高效动态时间规整算法语音识别系统

动态时间规整算法是结合了动态时间规整(DTW)技术和距离测度计算技术的一种非线性规整算法,在语音识别模板匹配中有重要的应用。为此提出一种改进的高效动态时间规整算法,其能有效加快搜索路径的寻找。基于Matlab实现了隐马尔科夫算法、高效动态时间规整算法和改进的高效动态时间规整算法的语音识别系统,同时进行了算法的仿真实验。实验结果表明,基于改进高效动态时间规整算法的训练速度远大于基于隐马尔可夫算法和高效动态时间规整算法的训练速度,而识别率下降很小,对于小词汇量非连续语音识别中高效动态时间规整算法的识别率为97.56%,隐马尔可夫算法的识别率为97.14%,改进高效动态时间规整算法的识别率为96.43%。     

Upper Bound Kullback–Leibler Divergence for Transient Hidden Markov Models

This paper reports an upper bound for the Kullback–Leibler divergence (KLD) for a general family of transient hidden Markov models (HMMs). An upper bound KLD (UBKLD) expression for Gaussian mixtures models (GMMs) is presented which is generalized for the case of HMMs. Moreover, this formulation is extended to the case of HMMs with nonemitting states, where under some general assumptions, the UBKLD is proved to be well defined for a general family of transient models. In particular, the UBKLD has a computationally efficient closed-form for HMMs with left-to-right topology and a final nonemitting state, that we refer to as left-to-right transient HMMs. Finally, the usefulness of the closed-form expression is experimentally evaluated for automatic speech recognition (ASR) applications, where left-to-right transient HMMs are used to model basic acoustic-phonetic units. Results show that the UBKLD is an accurate discrimination indicator for comparing acoustic HMMs used for ASR.      

Design and analysis of planar monopole antennas using a genetic algorithm approach

The application of genetic algorithm (GA) optimization to the design and analysis of planar monopole antennas is presented. GA is first used to optimize the impedance matching bandwidth of two particular planar element shapes, the bow-tie (BT) and reverse bow-tie (RBT). The results of this study indicate that the RBT can achieve a significantly wider bandwidth with a much smaller size than the traditional BT. In a follow-on study, GA is used to generate arbitrarily shaped planar monopole designs, which exhibit improved broadband performance and/or reduced size compared with the RBT. The designs generated by the GA demonstrate a better tradeoff between matching bandwidth and electrical size compared with planar monopole designs previously characterized in the literature. Analysis of results from simulation and measurement are presented, which provide insight into the operation of these antennas as well as the key parameters that lead to improved performance. Finally, a performance bound is generated to relate the bandwidth limitation of planar monopoles to size.     

一种改进的语音识别词错误率评估算法

在建立语音识别系统的过程中错误率评估起着非常重要的作用,传统的词错误率算法仅仅是基于最小错误率,具有显著的缺陷,因而不能准确评估系统的错误率。提出一种改进的基于最小错误率和时间信息的词错误率评估算法,能够准确评估系统的错误率,为声学模型的优化提供指导。同时列举了该评估算法在建立语音识别系统过程中的应用。     

SIMD-systolic architecture and VLSI chip for computing the dynamic time-warping algorithm

A VLSI architecture, which exhibits both SIMD and systolic behaviour for computing the dynamic time-warping (DTW) algorithm is presented. Such an architecture is well-suited for VLSI implementation because of its regular structure and small number of input/output. Currently, based on a 1-2 µm CMOS technology, a SIMD-systolic data-path chip has been designed and fabricated for computing the DTW algorithm. It is functionally correct and packaged as a 68-pin PGA chip. With such a chip, a 20000-word real-time DTW-based speech recognition system is achievable.     

DTW的ASIC实现算法研究

通过分析DTW算法，提出了一种适合ASIC实现的心动阵列结构。仿真结果表明，该并行VLSI处理器阵列系统能够在N M-1个时钟周期内计算出两个模板的匹配加权距离。较之基于通用处理器串行实现的DTW算法需要的3pMN／2个时钟周期，该算法节省了大量的运算时间。     

传感器网络基于DTW的多小波压缩算法

提出传感器网络环境下基于DTW的多小波数据压缩算法。首先研究汇聚节点中异步数据点—点对的对应关系,设计迭代算法求出具有最大相关性的DTW弯曲路径。接着提出最佳匹配点选择算法,通过DTW弯曲路径中一对一数据点—点对来预测异步数据向量间的函数关系,获取最佳匹配点,得到具有最大相关性的传感数据矩阵。然后设计多小波变换,利用传感数据矩阵的相关性来压缩数据,同时解决数据矩阵的行列不对称问题。实验结果表明,所提出的算法在能量聚集比、重构精度和运行时间等压缩性能指标上优于经典的分布式小波压缩算法。     

基于蚁群DTW算法的哼唱搜索研究

虽然传统DTW算法在模糊匹配上具有很好的性能,但是DTW算法通过局部最优化算出最佳路径的最小累计距离,计算量较大,搜索效率较低。蚁群DTW算法,结合蚁群算法的正回馈机制,搜索语音信号之间匹配的一条全局最优路径,既利用了语音信号的全局特征又考虑了其局部信息,与传统DTW算法相比,能大大提高哼唱搜索效率。     

基于DTW/SVM的语音识别系统在DSP中的实现

采用动态时间归正算法(DTW)和支持向量机(SVM)相结合产生一个新的基于径向基函数的DTW核函数实现语音识别,该方法在小词汇量及孤立词识别方面相对传统的隐马尔可夫模型有较大优势。为了满足语音识别系统对实时性和便携性的要求,提出了基于DTW/SVM的混合方法在TMS320C6711DSP芯片中实现的应用研究;给出了语音识别系统的原理框图,其中采用Mel倒谱系数为语音特征参数,应用了可变窗长端点检测技术;阐述了DSP设计中系统的软硬件设计方案及具体的接口电路,该系统使得语音识别更为快速便捷,并且具有一定的通用性。     

Research on point stabilization of a wheeled mobile robot using fuzzy control optimized by GA

A point stabilization scheme of a wheeled mobile robot(WMR)which moves on uneven surface is presented by using fuzzy control.Taking the kinematics and dynamics of the vehicle into account,the fuzzy controller is employed to regulate the robot based on a kinematic nonlinear state feedback control law.Herein,the fuzzy strategy is composed of two velocity control laws which are used to adjust the speed and angular velocity,respectively.Subsequently,genetic algorithm(GA)is applied to optimize the controller parameters.Through the self-optimization,a group of optimum parameters is gotten.Simulation results are presented to show the effectiveness of the control strategy.     

基于动态时间规整和主动外观模型的动态表情识别

针对静态表情特征缺乏时间信息,不能充分体现表情的细微变化,该文提出一种针对非特定人的动态表情识别方法:基于动态时间规整(Dynamic Time Warping, DTW)和主动外观模型(Active Appearance Model, AAM)的动态表情识别。首先采用基于局部梯度DT-CWT(Dual-Tree Complex Wavelet Transform)主方向模式(Dominant Direction Pattern, DDP)特征的DTW对表情序列进行规整。然后采用AAM定位出表情图像的66个特征点并进行跟踪,利用中性脸的特征点构建人脸几何模型,通过人脸几何模型的匹配克服不同人呈现表情的差异,并通过计算表情序列中相邻两帧图像对应特征点的位移获得表情的变化特征。最后采用最近邻分类器进行分类识别。在CK+库和实验室自建库HFUT-FE(HeFei University of Technology-Face Emotion)上的实验结果表明,所提算法具有较高的准确性。