基于局域判别基的音频信号特征提取方法

音频特征提取在音频信号分析和处理中起着非常重要的作用。考虑到音频信号的非平稳性,对音频信号进行小波包分解,为了获取健壮的特征,采用改进的局域判别基（LDB）技术对小波包树进行裁剪,提取局域判别基各子空间能量的统计特征组成特征矢量,并利用Fisher准则函数进行特征选择,根据特征矢量设计支持向量机分类器,对三类音频进行分类识别。实验结果表明,该方法提取的特征矢量在音频信号分类中是非常有效的。     

基于MPI的云计算模型

根据消息传递接口（MPI）的特点,提出云计算在MPI领域的应用方法,包括MPI的云计算算法设计模型、云计算原理、核心计算模式、处理流程,并介绍云计算的分布式及并行化特性。理论分析结果表明,该算法是有效可行的,优于传统并行技术,能够为算法分布化及并行化提供新思路。     

基于MPI的背包问题并行程序设计与实现

MPI（Message Passing Interface）是消息传递并行程序设计的标准之一,概述了MPI的概念和组成,着重介绍了支持并行程序设计的消息传递接口（MPI）以及在MPI环境下的并行程序设计方法,并给出一个MPI并行程序设计实例,说明了MPI的程序设计流程和普通串行程序设计之间的关联。     

基于小波分析和HMM的语音识别模型建立与仿真

利用隐马尔可夫模型HMM优异的时序建模能力及小波变换可以对信号进行多尺度分析并有效提取信号的局部信息的特点,建立了混合语音识别模型.在语音信号的识别过程中考虑到了信号的非平稳性,采用并行识别的方法分别获取分类信息,根据混合模型的识别算法做出识别决策,减小了系统对环境的依赖性,提高了其自适应能力.仿真实验结果表明,混合模型识别结果比单一HMM模型或小波模型识别结果更佳,提高了整体的识别速度和识别率.     

基于HMM和WNN的心音信号身份识别研究*

将隐马尔可夫模型（HMM）与小波神经网络（WNN）相结合,提出了一种基于心音信号的身份识别方法。该方法首先利用HMM对心音信号进行时序建模,并计算出待识别心音信号的输出概率评分;再将此识别概率评分作为小波神经网络的输入,通过小波神经网络将HMM的识别概率值进行非线性映射,获取分类识别信息;最后根据混合模型的识别算法得出识别结果。实验采集80名志愿者的160段心音信号对所提出的方法进行验证,并与GMM模型的识别结果进行了对比,结果表明,所选方法能够有效提高系统的识别性能,达到了比较理想的识别效果。     

基于二维细胞自动机和中值运算的图像滤波方法

为滤除手写汉字识别图像预处理中二值图像上的噪声点，结合中值滤波和细胞自动机滤波方法，提出一种二值图像CA-M（Cellular Automata—Median）滤波方法，并给出具体的算法流程．利用该方法进行滤波去噪实验并计算算法效率，对二值图像去噪非常有效．     

基于高阶累积量的多音并行调制信号检测研究

针对多音并行信号的调制识别问题,提出了一种基于四阶累积量的识别方法,用以识别特定通信协议下子载波调制样式为MPSK的多载波调制信号.该方法利用带通MPSK信号的累计量特征,在不考虑相位影响的情况下提出了该多音并行信号近似模型并通过理论计算和仿真验证此模型的正确性.然后利用此模型通过计算仿真找到子载波数与累计量参数的对应关系.根据这种对应关系提出识别算法并进行识别仿真,仿真结果表明在低信噪比(-5dB)下有很好的识别效果.     

基于DHMM的轴承故障音频诊断方法

轴承音频信号包含了大量的运行状态信息,与振动信号相比,音频信号的采集是非接触式的,具有使用方便和成本低廉等优势。通过提取机械轴承音频信号的Mel频率倒谱系数（MFCC）特征参数,并使用具有良好识别和抗噪性能的隐马尔可夫模型（HMM）分析轴承运行状态,首次将HMM对音频信号的分析方法应用于故障诊断。为了能够实现对轴承故障的实时诊断,采用计算量较小的离散HMM（DHMM）模型加快训练和识别速度。实验结果表明,该方法对轴承运行状态的识别正确率接近90%,识别时间约为31 ms,效果较好,有效可行,具有很好的应用前景。     

数据融合系统中并行目标识别的研究与实现

分析了将情报侦察数据融合系统中目标识别中心并行化的可行性，提出了基于航迹的任务划分策略，采取集中式动态负载平衡技术设计并实现了基于消息传递接口（MPI）的并行目标识别中心。同时给出了Dempster-Shafer证据理论在适当条件下的一种简化形式及其在并行环境下的应用方法。最后给出融合系统并行目标识别中心的性能测试结果，在保证识别可信度的基础上大大提高了处理速度，解决了融合系统的性能瓶颈问题。     

基于多态并行处理器的生物计算并行实现

针对传统的生物计算中DNA序列保守序列的识别（模体识别）和最长公共子序列计算需要较大的数据量、计算量,以及功耗大等问题,文中提出了两种基于PAAG多态并行处理器的并行算法,该并行处理器能够支持数据、线程、指令多种并行。通过编程在PAAG多态并行处理的处理单元（ PE）上开发了相应的串行和并行程序,将计算的不同过程分派到不同的处理单元（ PE）上进行处理,实现了不同粒度算法的并行。实验结果表明,文中提出的并行算法使模体识别和最长公共子序列的计算效率得到明显提高。     

无线传感网络中特征序列构造与识别算法

无线传感网络中设备之间的干扰逐步成为制约网络性能提高的瓶颈,但干扰并不会导致信号的所有信息被完全淹没. 为准确识别并发通信的链路,需要为每条链路分配唯一的特征序列. 传统的MAC地址在干扰背景下难以被识别. 本文利用信号的特征信息,提出了一种基于多哈希的特征序列的分布式构造与识别策略. 这种构造与识别策略利用干扰信息提高无线信号的传输效率、提升并发传输的特征序列识别能力,并显著降低识别特征序列的处理时间开销. 实验利用软件无线电试验环境USRP ＋ GNU Radio构建实测平台进行对所设计的算法进行评估,实验结果充分验证了特征序列的构造与识别算法的高效性.     

Aircraft class identification based on take-off noise signal segmentation in time

Aircraft noise is one of the most uncomfortable kinds of sounds. That is why many organizations have addressed this problem through noise contours around airports, for which they use the aircraft type as the key element. This paper presents a new computational model to identify the aircraft class with a better performance, because it introduces the take-off noise signal segmentation in time. A method for signal segmentation into four segments was created. The aircraft noise patterns are extracted using an LPC (Linear Predictive Coding) based technique and the classification is made combining the output of four parallel MLP (Multilayer Perceptron) neural networks, one for each segment. The individual accuracy of each network was improved using a wrapper feature selection method, increasing the model effectiveness with a lower computational cost. The aircraft are grouped into classes depending on the installed engine type. The model works with 13 aircraft categories with an identification level above 85% in real environments.     

BP神经网络和支持向量机相结合的电容器介损角辨识

针对电力电容器介质损耗的计算方法稳定性较差,频率波动对介损角的辨识有较大影响的问题,提出了BP神经网络和支持向量机（support vector machine, SVM）相结合（BP-SVM）的辨识方法,并且首次应用于电容器介损角的辨识。在辨识过程中,首先,对电容器工作一段时间的信号进行采样和预处理,预处理后的信号作为训练集训练BP-SVM模型;然后,使用训练好的BP-SVM模型对预处理后新的采样信号进行辨识,判断介损角的变化量。此外,给出了基于BP-SVM模型的介损角表示信号D_δ（t）的计算过程,同时分析了在讨论域内信号D_δ（t）的幅值即是介损角δ。仿真分析结果表明,提出的BP神经网络和SVM相结合的电容器介损角辨识方法比基于深度学习的辨识方法具有更高的辨识准确率,并且频率变化对BP-SVM方法的辨识性能无明显影响。     

Improvements and analysis of nonlinear parallel behavioral models

This article performs an analysis of current limitations regarding the extraction of parallel behavioral models to reproduce the power amplifier (PA) nonlinear behavior and its dynamics. To overcome these limitations, a general preprocessing block that clearly improves the identification capabilities shown by classical parallel structures is proposed. It follows the principle of separating both static and dynamic nonlinear behavior of the PA to obtain a better identification performance. A comparison with common parallel configurations using linear estimation is performed, to highlight the benefits of using the preprocessing structure. Furthermore, a new nonlinear parallel structure using sub‐band filtering techniques is also proposed. For the models extraction and comparison, four types of noise‐free simulated data presenting different levels of nonlinearities and memory, as well as a measured signal obtained from a laboratory amplifier have been considered. © 2009 Wiley Periodicals, Inc. Int J RF and Microwave CAE, 2009.     

Subspace identification by data orthogonalization and model decoupling

It has been observed that identification of state-space models with inputs may lead to unreliable results in certain experimental conditions even when the input signal excites well within the bandwidth of the system. This may be due to ill-conditioning of the identification problem, which occurs when the state space and the future input space are nearly parallel.We have in particular shown in the companion papers (Automatica 40(4) (2004) 575; Automatica 40(4) (2004) 677) that, under these circumstances, subspace methods operating on input-output data may be ill-conditioned, quite independently of the particular algorithm which is used. In this paper, we indicate that the cause of ill-conditioning can sometimes be cured by using orthogonalized data and by recasting the model into a certain natural block-decoupled form consisting of a “deterministic” and a “stochastic” subsystem. The natural subspace algorithm for the identification of the deterministic subsystem is then a weighted version of the PI-MOESP method of Verhaegen and Dewilde (Int. J. Control 56 (1993) 1187-1211). The analysis shows that, under certain conditions, methods based on the block-decoupled parametrization and orthogonal decomposition of the input-output data, perform better than traditional joint-model-based methods in the circumstance of nearly parallel regressors.     

Parallel compressive sampling matching pursuit algorithm for compressed sensing signal reconstruction with OpenCL

Compressive sensing (CS) is a new signal processing method, which was developed recent years. CS can sample signals with a frequency far below the Nyquist frequency. CS can also compress the signals while sampling, which can reduce the usage of resources for signal transmission and storage. However, the reconstruction algorithm used in the corresponding decoder is highly complex and computationally expensive. Thus, in some specific applications, e.g., remote sensing image processing for disaster monitoring, the CS algorithm usually cannot satisfy the time requirements on traditional computing platforms. Various studies have shown that many-core computing platforms such as OpenCL are among the most promising platforms that are available for real-time processing because of their powerful floating-point computing capabilities. In this study, we present the design and implementation of parallel compressive sampling matching pursuit (CoSaMP), which is an OpenCL-based parallel CS reconstruction algorithm, as well as some optimization strategies, such as access efficiency, numerical merge, and instruction optimization. Based on experiments using remote sensing images with different sizes, we demonstrated that the proposed parallel algorithm can achieve speedups of about 41 times and 58 times on AMD HD7350 and NVIDIA K20Xm platforms, respectively, without modifying the application code.     

The identification of electric load simulator for gun control systems based on variable-structure WNN with adaptive differential evolution

Owing to the complex nonlinearities of the electric load simulator (ELS) for the gun control system (GCS), the surplus torque plays a great negative impact on the performance of the loading system. This paper proposes a variable-structure wavelet-neural-network (VSWNN) identification strategy based on adaptive differential evolution (ADE). First of all, a mathematical model is established based on the structure and the working principle of the ELS. Then an intelligent identification method is applied, where the wavelet function is chosen as the excitation function, which improves the generalization and approximation ability of the neural network. The ADE is used to optimize the parameters, which solves the difficulty of determining the structure of the WNN. In order to reduce the computation complexity and speed up the convergence of the identification system, the adaptive laws of the pitch adjusting rate (PAR), band width (BW) and variable numbers of neurons are proposed. Finally, a pseudo random multilevel signal and a linear frequency modulation signal are chosen as input signals for the hardware-in-the-loop simulation. The test results show that the proposed ADE-VSWNN algorithm has superior validity and practicability, especially when the identification algorithm is used in the working circumstances with different inertial torque. Further, the high precision and strong robustness of the identification algorithm are further verified.     

基于小波分析的分数阶系统辨识信号降噪的变尺度阈值方法

在目前愈来愈被关注的分数阶控制研究中,系统辨识的分数阶理论与方法是一个重要方向,其中,辨识实验检测数据的降噪是必须关注的课题。基于小波分析理论与方法,首先对系统辨识中常用的以伪随机二进制序列(PRBS)激励的分数阶系统输出信号及其干扰噪声的特性进行分析讨论,在此基础上,为克服常规阈值降噪法的局限性,提出了针对多层小波分解系数进行非线性变尺度量化改造的算法,进而形成了一种分数阶系统辨识信号降噪的变尺度阈值方法。仿真实验表明,该方法能够将噪声干扰削减到满意的水平,对于不同的信噪比情形具有很好的适用性。该研究旨在为进一步的辨识算法设计提供参考,以提高辨识精度。     

Sparse representation-based classification for breath sample identification

It has been discovered that some compounds in human breath can be used to detect some diseases and monitor the development of the conditions. A sensor system in tandem with certain data evaluation algorithm offers an approach to analyze the compositions of breath. Currently, most algorithms rely on the generally designed pattern recognition techniques rather than considering the specific characteristics of data. They may not be suitable for odor signal identification. This paper proposes a Sparse Representation-based Classification (SRC) method for breath sample identification. The sparse representation expresses an input signal as the linear combination of a small number of the training signals, which are from the same category as the input signal. The selection of a proper set of training signals in representation, therefore, gives us useful cues for classification. Two experiments were conducted to evaluate the proposed method. The first one was to distinguish diabetes samples from healthy ones. The second one aimed to classify these diseased samples into different groups, each standing for one blood glucose level. To illustrate the robustness of this method, two different feature sets, namely, geometry features and principle components were employed. Experimental results show that the proposed SRC outperforms other common methods, such as K Nearest Neighbor (KNN), Linear Discriminant Analysis (LDA), Artificial Neural Network (ANN), and Support Vector Machine (SVM), irrespective of the features selected.