Generation Wavelet Research Articles

Deep second generation wavelet autoencoders based on curvelet pooling for brain pathology classification

Neurological disorders represent anomalies relevant to the human nervous system. They also contain biochemical, anatomical and electrical modifications in the central nervous system, the spinal cord and the brain. These disorders provoke different symptoms. Early diagnosis of such changes is necessary for treatment, to limit disease progression. An accurate and CAD system is introduced in this paper to classify brain magnetic resonance images, which overcomes crucial problems in pattern classification, such as extracting certain features in the training phase. Our contribution is to merge a Second-Generation Wavelet Transform Network (SGWTN) and deep learning architectures, hence suggesting novel supervised feature extraction approaches for pattern classification. Our novel architecture allows the classification of all classes of our dataset by the reconstruction of a deep stacked SGWTN-Autoencoder (SGWTN-AE). Combining Curvelet Pooling (CP) with the Adam gradient calculation method can improve their accuracy. We implement CP with Adam in this work using the Haar curvelet (CurvPool-AH) and the Shannon Curvelet (CurvPool-AS). This Network is obtained after a series of SGWTN-AEs followed by a Softmax classifier at the last layer. We find that CurvPool performs fairly well on all datasets. However, overfitting is an issue with all the approaches considered. CP has the potential to outperform current approaches, particularly when merged with an adaptive gradient and curvelets chosen specifically for the data. Our architecture is tested with various image datasets. These latter are as follows: DS-66, DS-90, DS-160, and DS-255. Based on 5×5 cross-validation, our suggested approach can outperform the state-of-the-art methods in terms of classification accuracy.

Fault data extraction of human-computer interaction interface for music electronic products based on improved second generation wavelet algorithm

Fault data extraction of human-computer interaction interface for music electronic products based on improved second generation wavelet algorithm

Fault data extraction of human-computer interaction interface for music electronic products based on improved second generation wavelet algorithm

Fault data extraction of human-computer interaction interface for music electronic products based on improved second generation wavelet algorithm

A multi-branch redundant adversarial net for intelligent fault diagnosis of multiple components under drastically variable speeds

Intelligent fault diagnosis with small training samples plays an important role in the safety of mechanical equipment. However, affected by sharp speed variation, fault feature is extremely weak, which raises difficulty for fault diagnosis. The mutual coupling of multi-component fault features further increases the difficulty. Considering the ability of redundant second generation wavelet transform in non-stationary feature extraction, a multi-branch redundant adversarial net (RedundancyNet) is proposed to address the above issues. The Net consists of discriminator, the generator based on redundant reconstruction, and the classifier based on redundant decomposition. Firstly, through adversarial training process, the generator fuses multi-scale features to generate the signal with varying speeds, thereby expanding training data. Secondly, through layer-by-layer multi-resolution feature enhancement, the classifier boosts weak fault features of vibration signals at variable speeds. Finally, a multi-branch framework is proposed to realize multi-component fault location and damage identification. The proposed method is validated on two cases. The average classification accuracy in the two cases reach 97.14% and 98.33% respectively. However, other end-to-end intelligent fault diagnosis methods for varying speeds or small samples can only reach the highest classification accuracy of 95.14% in Case 1 and 93.59% in Case2, which is much less than RedundancyNet. The analysis results highlight the effectiveness of the net under drastically variable speeds and small faulty training samples. Besides, the proposed classifier is easy to understand, which reveals the process of feature learning and the extracted feature under varying speeds.

A Robust Passive Islanding Detection Technique With Zero-Non-Detection Zone for Inverter-Interfaced Distributed Generation

In this paper, a new and efficient passive islanding detection technique for a grid-connected hybrid distributed generating system has been proposed using Second Generation Wavelet Transform (SGWT) and Maximum Overlap Discrete Wavelet Transform (MODWT). In this technique, the ripple content of the voltage signal is measured at the Point of Common Coupling (PCC) and it is monitored through the decomposition of the signal by applying SGWT and MODWT technique up to suitable finer levels. This algorithm able to detects all kinds of islanding at nearly 0.3 sec, even with zero power mismatch. It is verified on a wide range of operational conditions and is also validated in various non-islanding conditions. Furthermore, the proposed technique is simple, there is no need for a classifier requirement, not depends on system parameters, zero non-detection zone (NDZ), and its operation is independent of the capacity and nature of distributed generation (DG) connected to the utility grid.

Aging state detection of viscoelastic sandwich structure using redundant second generation wavelet packet transform and fuzzy support vector data description

Viscoelastic sandwich structure plays an important role in mechanical equipment, nevertheless viscoelastic material inevitably suffers from gradual aging. For guaranteeing the operation safety of mechanical equipment, it is urgent to perform the aging state detection of viscoelastic sandwich structure with vibration response signal analysis. However, the structural vibration response signal is non-stationary and its variation caused by the structural aging state change is very puny, and the abnormal state samples is lacking. The vibration-based structural aging state detection has become a challenging task. Therefore, a novel method based on redundant second generation wavelet packet transform (RSGWPT) and fuzzy support vector data description (FSVDD) is proposed for this task. For extracting sensitive aging feature information, RSGWPT is introduced to process the structural vibration response signal, and multiple energy features are extracted from the frequency-band signals to reflect structural aging state change. For accurate and automatic aging state identification, by fusing fuzzy theory, FSVDD only uses the normal state samples for training and can identify the abnormal severity degrees is developed to identify the structural aging states. The proposed method is applied on a viscoelastic sandwich structure to validate its effectiveness, and different structural aging states are created through the accelerated aging of viscoelastic material. The analysis results show the outstanding performance of the proposed method.

Recognition and Simulation of Exercise Mode Based on Energy Consumption Model

Sports energy consumption is a quantitative reflection of physical exercise effect. Combined with different sports modes and students’ physical characteristics, the calculation model of sports energy consumption is put forward. Firstly, the relationship between students’ age, height, weight, gender, and energy consumption is analyzed by using multiple linear regression method, and a linear acceleration model is proposed by combining different exercise methods. The relationship between the integral value of acceleration and energy consumption is analyzed, and a linear integral model based on different motion modes is proposed. Based on the kinetic energy theorem, the student movement energy expenditure is estimated. This paper proposes a human movement recognition method based on hybrid features, which mostly can represent the curve of the second generation wavelet transform edge thinning, and from the edge and texture features of the optimal said human posture, the statistical characteristic of the second generation wavelet transform is subtly trained as image characteristics, learning and recognition of human movement. Then, the motion recognition algorithm is tested, which can effectively identify the common movement patterns of primary and middle school students. Finally, the linear relationship between the estimation results of the model and the calculation results of Meijer is analyzed. The analysis results show that the linear acceleration model proposed in this paper can estimate the energy consumption of primary and middle school students’ motion relatively accurately.

Crack Detection under Coating of Flow Passage Parts based on Adaptive Second Generation Wavelet

Cracks under coating of flow passage parts of hydropower station is a very serious fault, which can directly lead to fatigue fracture. This will directly affect the safety of the entire hydropower unit. The array eddy current testing method has been studied, and crack signals are found with a lot of noise. In order to improve the signal quality, adaptive second generation wavelet transform is applied to the array eddy current signal analysis to reduce the signal noise. A test block with some cracks under the coating was tested, and the tiny crack defect with size of 0.13mm × 0.5mm × 5mm was identified. The experimental results show that the array eddy current detection method based on adaptive second generation wavelet transform can effectively identify the cracks under the coating of flow passage components of hydropower station, with high detection accuracy and strong adaptability, and has good engineering application value.

Performance analysis of biorthogonal filter design using the lifting-based scheme for medical image transmission

To minimise the storage space and for fast transfer of the digital images, it is necessary for the medical images to undergo image compression. There are various techniques in which the images are being diagnosed, based on that the image compression being performed. The choice of the filters in the image compression is an issue which affects the quality of the image. Hence, a novel biorthogonal filter using the lifting scheme has been developed. The proposed architecture gives the same characteristics of second generation wavelets. The proposed architecture is designed using MATLAB for different medical images and the PSNR, SNR, MSE, BPP having a compression of 96.85%. The verified model is synthesised using Xilinx Spartan-6 FPGA with the speed of −3. The results show that the designed method works at a frequency of 252.89 MHz, and absorbs the power 0.036 W for 2.5 volts with temperature at 25.2°C.

A novel genetic algorithm with CDF5/3 filter-based lifting scheme for optimal sensor placement

The generic algorithm has been receiving significant attention due to the node placement problem in the field of sensor application in terms of machine learning. Sensor deployment is able to provide maximum coverage and maximum connectivity with less energy consumption to sustain the network lifetime. The maximum quality coverage problem has been solved successfully by an evolutionary algorithm while placing nodes in optimal position. In evolutionary algorithms, genetic algorithm (GA) plays an important technique for deploying the sensor in the form of population matrix. However, the existing techniques are unable to place sensor position perfectly. In this paper, a novel genetic algorithm with second generation wavelet transform (SGWT) is proposed for identifying optimal potential position for node placement. In order to improve the quality of population matrix, bi-orthogonal Cohen-Daubechies-Feauveau wavelet (CDF 5/3) has been employed. The proposed method is performed primarily to generate sensor position with different populations. Subsequently, it can extend to CDF5/3 filter-based lifting scheme to adjust the sensor position. The proposed method has been compared with random deployment, genetic algorithm and GA with CDF5/3 wavelets in terms of target to cover by the sensor. The result of the proposed method affirms better optimisation as compared to the state-of-art techniques.

Compound fault identification of rolling element bearing based on adaptive resonant frequency band extraction

The high frequency resonance (HFR) technique is regarded as a powerful tool for fault diagnosis of rolling element bearings. Different from the usage of the HFR in single fault, the determination of multiple resonant frequency bands under the compound faults and extraneous random impulses is still a challenging task. This paper develops a novel compound fault identification method based on adaptive resonant frequency band extraction. The improved redundant second generation wavelet packet transform is first presented to decompose vibration signal into various narrow bands for providing a fine separation of fault signatures. Then the squared envelope spectrum sparsity criteria is designed to quantify fault characteristics buried in narrow frequency bands. Consequently, the squared envelope spectrum sparsogram is constructed to highlight optimal resonant bands, and the compound faults can be well detected by band-pass filtering and envelope analysis. The numerical and experimental results confirm effectiveness and superiority of the proposed method, which is more sensitive to fault-related impulses and robust to extraneous interferences.

Application of AI failure identification techniques in condition monitoring using wavelet analysis

In the context of Industry 4.0, Condition Based Maintenance (CBM) for complex systems is essential in order to identify failures and mitigate them. After the identification of a sensor set that guarantees the system monitoring, three main problems must be addressed for effective CBM: i) collection of the right data; ii) choice of the optimal technique to identify the specific data-set; iii) correct classification of the results. The solutions currently used are typically data driven and, therefore, the results are variable, as it is sometimes challenging to identify a pattern for all specific failures. This paper presents a solution that combines a data driven approach with an in-depth knowledge of the mechanical system’s behaviour. The choice of the right sensor set is calculated with the aid of the software MADe (Maintenance Aware Design environment), whereas the optimal data-set identification technique is pursued with a second tool called Syndrome Diagnostics. After an overview of such methodology, this work also presents RSGWPT (Redundant Second Generation Wavelet Packaged Transform) analysis to show different possible outcomes depending on the available sensor data and to tailor a detection technique to a given data set. Supervised and unsupervised learning techniques are tested to obtain either an anomaly detection or a failure identification depending on the chosen sensor set. By using the described method, it is possible to identify potential failures in the system with sufficient notice to implement the optimal maintenance actions.

Application of Random Forest and Hidden Markov Models for Automatic and Fast Classification of Power Quality Signals

In this paper, wavelet transform, namely the maximal overlap discrete Wavelet Transform (MODWT) and the second generation Wavelet Transform (SGWT) have been implemented. These wavelet transforms are applied to get selected features of the signals. Features are used as inputs to two types of classifiers namely, Hidden Markov Model (HMM) classifiers and the Random Forest (RF) classifier in the both absence and presence of Noise to evaluate the efficiency. The classification accuracy (CA) calculated using these classifiers clearly shows that the RF classifiers is a better classifier then the HMM classifier as it possess higher recognition rate at all levels of noise along with the pure PQ signals. Another important property of RF classifier is the proper classification of large number of class of both slow and the fast disturbances.

Data processing of pulsation signals based on local wave decomposition and teager energy

Nonstationary and nonlinear signals are often encountered in the research and development of turbomachinery. One such example is the pulsating strain signal measured during engine ramping to find the maximum resonant strain in the application. Because the pulse signal may come from different interference sources, it is difficult to detect weak useful signal in the background of noise. In order to solve this problem, a new method based on local wave decomposition (LWD) and Teager energy is proposed. According to the local characteristics of the vibration signal, the optimal prediction operator of the transformed sample is constructed by selecting the appropriate square error minimization criterion, so that the second generation wavelet basis function can fit the local characteristics of the vibration signal. The adaptive second generation wavelet is used as the prefilter to improve the effect of LWD decomposition. Then the correlation kurtosis is used to select the sensitive internal model function (IMFs). Finally, the Teager energy operator algorithm is applied to the selected sensitive IMF to identify the characteristic frequency. The validity of the method is verified by the measured strain signal of a turbocharger turbine as an example.

A MATLAB Suite for Second Generation Wavelets on an Interval and the Corresponding Adaptive Grid

A collection of Matlab routines for the second generation wavelet transformation and inverse wavelet transformation on the space $\mathcal{L}_{2}([a,b])$ is presented. These wavelet transforms are further used for computing the wavelet function and scaling function values ( $\psi (x)$ and $\phi (x)$ respectively). At the end a procedure for generating an adaptive grid using second generation wavelet is provided. In the collection of the Matlab routines, we are providing three Matlab functions namely, Reconstruction_testing.m, AdaptiveGrid_standard_testing.m and AdaptiveGrid_modified_testing.m for directly testing the results claimed in the paper.

Centralized Wavelet Multiresolution for Exact Translation Invariant Processing of ECG Signals

Dyadic wavelet transform is useful in analyzing electrocardiogram (ECG) signals due to its fast computation and its multiresolution ability. In order to improve the feature extraction performance of dyadic wavelet transform, a new construction example of centralized multiresolution (CMR) is proposed. The proposed CMR example consists of two elements, namely, a dyadic part and a non-dyadic part. The dyadic part, based on the maximal overlap second generation wavelet packet transform (SGWPT), generates dyadic wavelet packets. The non-dyadic part engenders ensemble wavelet packets by postprocessing on the dyadic part. The produced wavelet packets and ensemble wavelet packets are combined to realize continued spectral refinement around fixed central analysis frequencies. Numerical simulation and a case study of ECG signal decomposition are utilized to validate the enhancements of the proposed CMR example. The processing results of the CMR example are compared with those of the dual tree complex wavelet transform and the conventional SGWPT. It is validated this CMR example achieves better feature extraction performances due to the presence of the exact translation invariance property.

Fractal Lifting Wavelets for Machine Fault Diagnosis

Fault diagnosis is of vital importance in safety and reliable operations of modern electromechanical systems. Advanced signal processing techniques are indispensable for extracting incipient features from measured dynamical signals. For discrete wavelet analysis, shift-invariance and proper frequency-scale configuration are both necessary for effective investigation of incipient fault features. In this paper, a novel fractal lifting scheme is proposed based on redundant second generation wavelet packet decomposition (RSGWPD). Implicit wavelet packets (IWPs), generated via fractal lifting scheme, can realize a novel centralized multiresolution. It is demonstrated that each IWP inherits the property of exact shift-invariance originated from redundant lifting scheme. In addition, a novel concept of nested centralized wavelet packet cluster is introduced for explaining merits provided by sets composed of IWPs. The numerical simulations were employed to validate the benefits of exact shift-invariance in a multiscale analysis of discrete time series. RSGWPD and IWPs are combined to conduct multiscale expansion of vibration measurement. To further explore optimal features, an indicator of spatial-spectral ensemble kurtosis is utilized to select optimal analysis parameters. The proposed technique was successfully applied to case studies of gearbox fault diagnosis as well as bearing fault diagnosis. The comparisons were made between results by the proposed technique and those provided by some other mainstream adaptive signal decomposition methodologies. It is verified that the combination of exact shift-invariance and centralized multiresolution significantly enhances the performance of incipient fault feature extraction of nonstationary vibration signals.

LiftingNet: A Novel Deep Learning Network With Layerwise Feature Learning From Noisy Mechanical Data for Fault Classification

The key challenge of intelligent fault diagnosis is to develop features that can distinguish different categories. Because of the unique properties of mechanical data, predetermined features based on prior knowledge are usually used as inputs for fault classification. However, proper selection of features often requires expertise knowledge and becomes more difficult and time consuming when volume of data increases. In this paper, a novel deep learning network (LiftingNet) is proposed to learn features adaptively from raw mechanical data without prior knowledge. Inspired by convolutional neural network and second generation wavelet transform, the LiftingNet is constructed to classify mechanical data even though inputs contain considerable noise and randomness. The LiftingNet consists of split layer, predict layer, update layer, pooling layer, and full-connection layer. Different kernel sizes are allowed in convolutional layers to improve learning ability. As a multilayer neural network, deep features are learned from shallow ones to represent complex structures in raw data. Feasibility and effectiveness of the LiftingNet is validated by two motor bearing datasets. Results show that the proposed method could achieve layerwise feature learning and successfully classify mechanical data even with different rotating speed and under the influence of random noise.

An improved empirical mode decomposition method using second generation wavelets interpolation

Empirical mode decomposition (EMD) may generate undesirable intrinsic mode functions (IMFs) under low sampling rate, which can significantly affect the results of decomposition. In this paper, an improved EMD method using second generation wavelets interpolation is presented which can eliminate undesirable IMFs and reduce the scale mixing effectively under low sampling rate. Firstly, the original signal under low sampling rate is reconstructed by inverse process of second generation wavelets lifting algorithm. Secondly, the location algorithm of extrema using second generation wavelets is given to obtain the accurate position of extrema. Finally, five examples are demonstrated to justify the effectiveness. Numerical simulation and experimental results are attained to show the effectiveness of the proposed method in eliminating undesirable IMFs and reducing the scale mixing, thereby making the proposed improved EMD a promising method for improving the performance of EMD under low sampling rate.

Analyzing the Vibration Signals for Bearing Defects Diagnosis Using the Combination of SGWT Feature Extraction and SVM

The aim of this paper is to introduce a multi-step vibration-based diagnostic algorithm to automatically diagnose bearings faults. The proposed diagnostic scheme extracts the informative features from each component by resorting to the second generation wavelet transform. Undoubtedly, a large dimension of features brought more challenges to detect healthy and defective bearings. In this regard, the dimensionality reduction phase makes use of linear discriminant analysis that aims to obtain a low dimensional representation of high dimensional data as well as achieves maximum separability between different classes. Furthermore, self-organizing maps (SOM) helps in evaluating and facilitating visual comprehension of the extracted features. In the following step, support vector machine (SVM) is used for identifying faulty and fault-free bearings. Finally, the performance of the proposed technique is compared with the previous works.