Adaptive bistable stochastic resonance and its application in mechanical fault feature extraction

Stochastic resonance (SR) is an important approach to detect weak vibration signals from heavy background noise. In order to increase the calculation speed and improve the weak feature detection performance, a new bistable model has been built. With this model, an adaptive and fast SR method based on dyadic wavelet transform and least square system parameters solving is proposed in this paper. By adding the second-order differential item into the traditional bistable model, noise utilization can be increased and the quality of SR output signal can be improved. The iteration algorithm for implementing the adaptive SR is given. Compared with the traditional adaptive SR method, this algorithm does not need to set up the searching range and searching step size of the system parameters, but only requires a few iterations. The proposed method, discrete wavelet transform and the traditional adaptive SR method are applied to analyzing simulated vibration signals and extracting the fault feature of a rotor system. The contrastive results verify the superiority of the proposed method, and it can be effectively applied to weak mechanical fault feature extraction.     

Wavelet basis selection and feature extraction for shift invariant ultrasound foreign body classification

This paper proposes a novel technique for automatic ultrasound non-destructive foreign body (FB) detection and classification. A signal registration process is introduced to eliminate shift variations commonly encountered in ultrasound signals. Information theory based methods are then developed for wavelet basis selection and feature extraction to facilitate robust FB classification. Probabilistic neural networks are used for FB classification. Experimental results confirm that the wavelet basis selected by the proposed method improves the FB classification accuracy. It is concluded that low order wavelet bases have better ability to distinguish classes with great similarities than their higher order counterparts, while the reverse is true for more divergent classes.     

A wavelet space based approach for Doppler ultrasound blood signals separation

In medical Doppler ultrasound systems, a high-pass filter which is usually employed to filter wall clutter components, will remove the information of the low velocity blood flow. To extract intact Doppler ultrasound blood signals, a novel approach is proposed based on the spatially selective noise filtration. The wall signals are firstly estimated by the spatially selective noise filtration from wavelet spatial correlation property. Then the wall clutters are exactly obtained by a wavelet threshold de-noising technique which eliminates the residual blood flow signals. Finally the intact blood flow signals are achieved by subtracting the wall signals from the mixed signals. This approach is applied to both computer simulated and in vivo carotid Doppler ultrasound signals. The experiment results show that the wavelet space based approach can exactly extract the blood flow signals, and achieve about 45% lower results in the mean absolute error than that of the high-pass filtering. This approach is expected to be an effective method to remove the wall clutters in Doppler ultrasound systems.     

LOW-OSCILLATION COMPLEX WAVELETS

In this paper we explore the use of two low-oscillation complex wavelets—Mexican hat and Morlet—as powerful feature detection tools for data analysis. These wavelets, which have been largely ignored to date in the scientific literature, allow for a decomposition which is more “temporal than spectral” in wavelet space. This is shown to be useful for the detection of small amplitude, short duration signal features which are masked by much larger fluctuations. Wavelet transform-based methods employing these wavelets (based on both wavelet ridges and modulus maxima) are developed and applied to sonic echo NDT signals used for the analysis of structural elements. A new mobility scalogram and associated reflectogram is defined for analysis of impulse response characteristics of structural elements and a novel signal compression technique is described in which the pertinent signal information is contained within a few modulus maxima coefficients. As an example of its usefulness, the signal compression method is employed as a pre-processor for a neural network classifier. The authors believe that low oscillation complex wavelets have wide applicability to other practical signal analysis problems. Their possible application to two such problems is discussed briefly—the interrogation of arrhythmic ECG signals and the detection and characterization of coherent structures in turbulent flow fields.     

Wavelet filter-based weak signature detection method and its application on rolling element bearing prognostics

De-noising and extraction of the weak signature are crucial to fault prognostics in which case features are often very weak and masked by noise. The wavelet transform has been widely used in signal de-noising due to its extraordinary time-frequency representation capability. In this paper, the performance of wavelet decomposition-based de-noising and wavelet filter-based de-noising methods are compared based on signals from mechanical defects. The comparison result reveals that wavelet filter is more suitable and reliable to detect a weak signature of mechanical impulse-like defect signals, whereas the wavelet decomposition de-noising method can achieve satisfactory results on smooth signal detection. In order to select optimal parameters for the wavelet filter, a two-step optimization process is proposed. Minimal Shannon entropy is used to optimize the Morlet wavelet shape factor. A periodicity detection method based on singular value decomposition (SVD) is used to choose the appropriate scale for the wavelet transform. The signal de-noising results from both simulated signals and experimental data are presented and both support the proposed method.     

改进的经验模态分解法分离超声多普勒血流与管壁信号

超声多普勒血流信号常包含管壁信号的干扰,准确分离二者对提高血流检测的精度具有重要作用。本文提出两种改进的经验模态分解(EMD)方法,先将含管壁信号的超声多普勒信号分解成多层本征模态函数(IMF),然后根据血流信号与管壁信号的不同特性,对既含管壁信号又含血流信号的IMF分量进行分离处理,最后将各层IMF分量中的管壁成分叠加得到管壁信号的估计,而血流信号可通过原信号减去估计的管壁信号而得到。将本方法用于计算机仿真信号和人体实测的超声多普勒信号,并与高通滤波器法、空间选择性降噪法和原EMD法进行比较,结果表明:本文提出的两种方法能在较大的管壁搏动速度范围内准确地分离血流信号和管壁信号,其平均相对误差比高通滤波器的结果降低了约52%和57%。可见,本文提出的两种方法有望用于血流信号与管壁信号的准确分离。      

倒谱参数稀疏分解下的汉语音谎言检测

为了提高汉语语音的谎言检测准确率,提出了一种对信号倒谱参数进行稀疏分解的方法。首先,采用小波包滤波器组对语音信号进行多频带划分,求得子频带对数能量并进行离散余弦变换以提取小波包频带倒谱系数,结合梅尔频率谱系数得到倒谱参数;其次,依据K-奇异值分解方法分别利用说谎和非说谎两种状态下的语音倒谱参数集训练得到过完备混合字典,在此字典上根据正交匹配追踪算法对参数集进行稀疏编码提取稀疏特征;最终进行多种分类模型下的识别实验·实验结果表明,稀疏分解方法相比传统参数降维方法具有更好的优化性能,本文推荐的稀疏谱特征最佳识别率达到78.34%,优于其他特征参数,显著提高了谎言检测识别准确率。      

气体栓子超声多普勒信号的仿真研究

通过分析研究气体栓子的超声多普勒信号的特征及其成因,建立气体栓子的超声多普勒仿真模型,为后续的气体栓子与固体栓子的分类提供理论基础。先利用超声辐射力和液体黏滞阻力的模型计算气体栓子所受的超声辐射力和黏滞阻力,然后计算气体栓子在血管内的径向加速度和轴向加速度,并从加速度得到气体栓子在血管内的运动轨迹,最后建立气体栓子的计算机仿真模型。实验中仿真合成气体栓子和固体栓子的超声多普勒信号,对信号的分析结果表明,相比于固体栓子,仿真的气体栓子信号受到作用力而产生的加速度将导致气栓信号在多普勒声谱图上频域的展宽,当气体栓子由低速区域加速到高速区域再由高速区减速至低速区将会在声谱图上体现为"V"型,若只有单一的加速或减速运动将只表现为斜线型。将仿真的气栓信号与临床采集的气栓信号进行比较,发现两者的特征是吻合的,说明气栓超声多普勒信号的仿真方法是合理的。      

De-noising of wayside acoustic signal from train bearings based on variable digital filtering

In the wayside Acoustic Defective Bearing Detector (ADBD) system, the recorded signal usually includes both the sound from train bearings and the other disturbance sources. The fact of heavy noise corruption and the Doppler Effect of multi-source acoustic signals would badly reduce the effectiveness of online defect detection of the ADBD system. In order to extract useful information from the multi-source signal with Doppler Effect, this paper proposes an effective de-noising method based on the variable digital filter (VDF) for the ADBD system. Specifically, the ridge extraction based on Short-Time Fourier Transform (STFT) is applied to estimate the instantaneous frequencies (IFs), with which the fitting IF curves based on the Morse theory of theoretical acoustics could be achieved by using the nonlinear curve-fitting so that the parameters of the initial position of the acoustic sources could be calculated. By the aid of these parameters, the IFs according to the target train bearing could be then extracted. After that, the FIR variable digital filters could be designed with all the IFs which match the Morse theory with Doppler Shift so that the noise from the other parts could be effectively restrained after filtering the original signal. The effectiveness of this method is verified by means of a simulation with multi-frequency signals and applications to diagnosis of train roller bearing defects. Results indicate that the proposed method is effective.     

Deceptive Chinese speech detection based on sparse decomposition of cepstral feature

In order to improve the performance of deception detection based on Chinese speech signals, a method of sparse decomposition on spectral feature is proposed. First, the wavelet packet transform is applied to divide the speech signal into multiple sub-bands. Band cepstral features of wavelet packets are obtained by operating the discrete cosine transform on loga?rithmic energy of each sub-band. The cepstral feature is generated by combing Mel Frequency Cepstral Coefficient and Wavelet Packet Band Cepstral Coefficient. Second, K-singular value decomposition algorithm is employed to achieve the training of an over-complete mixture dictionary based on both the truth and deceptive feature sets, and an orthogonal matching pursuit algorithm is used for sparse coding according to the mixture dictionary to get sparse feature.Finally, recognition experiments axe performed with various classified modules. Experimental results show that the sparse decomposition method has better performance comparied with con?ventional dimension reduced methods. The recognition accuracy of the method proposed in this paper is 78.34%, which is higher than methods using other features, improving the recognition ability of deception detection system significantly.     

基于频率切片小波变换和支持向量机的癫痫脑电信号自动检测

实现癫痫脑电信号的自动检测对癫痫的临床诊断和治疗具有重要意义.本文提出先使用频率切片小波变换分离出5个不同频段的节律信号,再分别计算每个节律信号的近似熵和相邻节律的波动指数,最后使用遗传算法优化的支持向量机进行分类.实验结果表明,所提出的方法能够对正常、癫痫发作间期和癫痫发作期三种脑电信号进行准确分类,分类准确率为98.33%.     

In vitro microemboli classification using neural network models and RF signals

Emboli classification is of high clinical importance for selecting appropriate treatment for patients. Several ultrasonic (US) methods using Doppler processing have been used for emboli detection and classification as solid or gaseous matter. We suggest in this experimental study exploiting the Radio-Frequency (RF) signal backscattered by the emboli since they contain additional information on the embolus than the Doppler signal. The aim of the study is the analysis of RF signals using Multilayer Perceptron (MLP) and Radial-Basis Function Network (RBFN) in order to classify emboli.Anthares scanner with RF access was used with a transmit frequency of 1.82 MHz at two mechanical indices (MI) 0.2 and 0.6. The mechanical index is given as the peak negative pressure (in MPa) divided by the square root of the frequency (in MHz). A Doppler flow phantom was used containing a 0.8 mm diameter vessel surrounded by a tissue mimicking material. To imitate gas emboli US behaviour, Sonovue microbubbles were injected at two different doses (10μl and 5μl) in a nonrecirculating at a constant flow. The surrounding tissue was assumed to behave as a solid emboli. In order to mimic real clinical pathological situations, Sonovue concentration was chosen such that the fundamental scattering from the tissue and from the contrast were identical. The amplitudes and bandwidths of the fundamental and the 2nd harmonic components were selected as input parameters to the MLP and RBFN models. Moreover the frequency bandwidths of the fundamental and the 2nd harmonic echoes were approximated by Gaussian functions and the coefficients were used as a third input parameter to the neural network models. The results show that the Gaussian coefficients provide the highest rate of classification in comparison to the amplitudes and the bandwidths of the fundamental and the 2nd harmonic components. The classification rates reached 89.28% and 92.85% with MLP and RBFN models respectively.This short communication demonstrates the opportunity to classify emboli based on a RF signals and neural network analysis.     

Wayside acoustic defective bearing detection based on improved Dopplerlet transform and Doppler transient matching

The diagnosis of train bearing defects plays a significant role in maintaining the safety of railway transport. However, the phenomenon of Doppler Effect in the acoustic signal recorded by the wayside Acoustic Defective Bearing Detector (ADBD) system leads to the difficulty for fault diagnosis of train bearings with a high moving speed. This paper proposes a double-searching solution based on improved Dopplerlet transform and Doppler transient matching to overcome the difficulty in wayside acoustic bearing diagnosis. In the solution, the first searching procedure is to extract necessary parameters of Doppler Effect under the situation with very low signal-to-noise ratio (SNR) based on an improved Dopplerlet transform. Using the obtained parameters, the Doppler Effect can be embedded into the constructed periodic Laplace wavelet transient models. Subsequently, the second searching procedure is conducted to search fault impact period of the defective bearing through an operation, called Doppler transient matching, which is to calculate the correlation coefficient between the Doppler transient model and the filtered raw signal with the Doppler Effect. The proposed double-searching algorithm can adapt to the real Doppler Effect situation and extract the exact fault impact period from the Doppler distorted signal, and thus shows powerful capability to analyze wayside acoustic signals from train bearings. The proposed wayside acoustic diagnostic scheme is verified by means of a simulated Doppler distorted signal with a very low SNR (−20 dB) and the experiments conducted on train bearings. The results indicate that the proposed algorithm is effective and has obvious advantages for ADBD system.     

基于小波变换的压缩感知理论对水质检测紫外-可见光谱数据的去噪研究

消除噪声影响对提高直接光谱法水质检测系统的测量稳定性和精度都具有重要意义。直接光谱法在线水质检测系统通常采用长寿命、无需预热的脉冲氙灯和适用于复杂检测环境的工业级光谱探测装置。针对整个光谱探测系统受到光源、光路和光电转换器件的严重影响,测定的光谱数据含有大量噪声这一实际问题,提出了基于小波变换的压缩感知去噪算法,并与传统小波阈值去噪方法进行了比较实验。针对化学需氧量为200 mg·L-1的邻苯二甲酸氢钾标液的紫外-可见光谱数据进行去噪处理,采用压缩感知去噪算法,将信号在小波域内分解,得到含噪高频系数;采用随机高斯矩阵作为压缩感知算法的观测矩阵,压缩比设置为2,对高频系数进行观测;选择正交匹配追踪算法恢复高频小波系数的稀疏性,从而达到去噪目的。同时针对传统的小波阈值去噪,采用daubechies4作为小波基的软阈值滤波方法对光谱数据进行去噪处理。为验证去噪算法的可行性,采集某溪水和城市生活污水的光谱信号分别采用以上两种方法进行去噪处理,实验结果表明：基于小波变换的压缩感知去噪算法适用于紫外-可见光谱法在线水质检测系统,该方法能在保留水样原始光谱信号的吸收特征的前提下有效地去噪,且去噪效果优于小波阈值去噪算法。与小波阈值去噪算法相比,信噪比提高了12.201 5 dB,均方根误差减小了0.009 3,峰值信噪比增加了5.299 dB。不仅避免了小波阈值去噪过程中阈值的选取依靠主观判断问题,而且在重构过程中有效地抑制了噪声,为直接光谱法检测水质参数提供了一种新的解决方案。     

一种用于光纤链路振动信号模式识别的规整化复合特征提取方法

相位光时域反射链路监测系统是一种利用光纤作为传感介质的传感系统, 能够监测、定位、识别入侵信号.模式识别模块是其重要组成部分, 实时智能区分安全扰动和危险入侵.本文提出一种用于光纤链路振动信号模式识别的复合特征提取方法.利用改进的双门限方法确定有效信号段的起止位置, 结合最大能量与最高信噪比挑选出采样周期内主要入侵扰动的特征段.综合利用特征段时域持续时间和小波包能量谱提取复合特征向量, 使用支持向量机进行模式识别.实验表明, 基于本文提出的规整化特征提取方法的模式识别准确率有了显著提高.     

Sound Signal Based Fault Classification System in Motorcycles Using Hybrid Feature Sets and Extreme Learning Machine Classifiers

Vehicles generate dissimilar sound patterns under different working environments. These generated sound patterns signify the condition of the engines, which in turn is used for diagnosing various faults. In this paper, the sound signals produced by motorcycles are analyzed to locate various faults. The important attributes are extracted from the generated sound signals based on time, frequency and wavelet domains which clearly describe the statistical behavior of the signals. Further, various types of faults are classified using the Extreme Learning Machine (ELM) classifier from the extracted features. Moreover, the improved classification performance is obtained by the combination of feature sets in different domains. The simulation results clearly demonstrate that the proposed hybrid feature set together with the ELM classifier gives more promising results with higher classification accuracy when compared with the other conventional methods.     

Experimental research into time–frequency characteristics of cavitation noise using wavelet scalogram

The cavitation has become the main cause of the damage to the hydraulic machine. Cavitation detection is very important to guarantee the safe running of the hydraulic machine. The sound, especially the audible sound, based methods are becoming attractive due to their simplicity and logicality in the application. However, the cavitation noise is easy to be contaminated by the background noise. In order to understand the characteristics of the cavitation noise deeply, using the wavelet scalogram analysis, this paper presents an experimental study to investigate the time–frequency characteristics of the cavitation noise of various cavitation states and the relation between the cavitation noise and the cavitation process. In addition, the method of parameters optimization for the wavelet is used to improve the transform performance of the wavelet scalogram. The results show that: the cavitation noise is composed of the components with wide band frequency and obvious impulse feature; but the cavitation noise of different cavitation stages has different time–frequency characteristics and compositions; in addition, the cavitation noise can be distinguished from the background noise because they have totally different frequency characteristic. This study validates that the cavitation noise can be used to detect the cavitation state and monitor the cavitation process.     

Internal tissue displacement measurement based on ultrasonic wave Doppler signal digital detection and its application to fetal movement monitoring

We have developed a fetal movement monitoring system based on small displacement measurement of internal tissues. When ultrasonic pulses are transmitted to the fetus, the reflected ultrasonic waves which have a Doppler frequency shift due to the fetal movements are detected by using an ultrasonic pulsed Doppler technique. In this paper, we propose a displacement measurement method for internal tissues which is based on the Doppler signal digital detection technique. In the method, the received ultrasonic RF signals are sampled with a sampling frequency of four times higher than the centre frequency of the ultrasonic waves; the Doppler frequency shift signals are derived using digital signal processing. From the detected signals, the internal displacements are estimated using the arc-tangent method. The basic algorithm of the detection method has already been used in the area of blood flow sensing, however, we apply the algorithm to the displacement measurement of internal tissues. The comparison between the proposed method and the conventional method is presented. The fetal movement quantitative monitoring system based on the method which has been constructed is shown.     

Detection on micro-Doppler effect based on 1550 nm laser coherent radar

A laser coherent detection system of 1550 nm wavelength was presented, and experimental research on detecting micro-Doppler effect in a dynamic target was developed. In this paper, the return signal in the time domain is decomposed into a set of components in different wavelet-scales by multi-resolution analysis, and the components are associated with the vibrational motions in a target. The micro-Doppler signatures are extracted by applying the reconstruction (inverse wavelet transform). During the course of the final data processing frequency analysis and time–frequency analysis are applied to analyze the vibrational signals and estimate the motion parameters successfully. The experimental results indicate that the micro-Doppler information in a moving can be effectively detected, and tiny vibrational signatures also can be acquired effectively by wavelet multi-resolution analysis and time–frequency analysis.     

基于听觉谱特征的变压器绕组状态检测研究

绕组松动是变压器常见故障之一,对变压器的安全运行产生巨大威胁.故对其进行精准的监测,对提高电力系统的安全稳定性具有十分重要的意义.基于声信号的变压器绕组松动检测,由于其具有无损检测和不需停运变压器等优点,成为近年来研究的热点.但声信号检测存在故障特征提前复杂和易受噪声干扰等缺陷,限制了其工程应用.该文提出了一种基于声信...