Trained Wavelet Neural Network Research Articles

Airborne transient electromagnetic imaging method based on wavelet neural network

The airborne transient electromagnetic method is a crucial technique for near-surface exploration in rugged terrains. In this study, we introduce a multi-input, single-output double-hidden wavelet neural network design for airborne transient electromagnetic pseudo-resistivity imaging. We construct uniform half-space, stratified stratum, and three-dimensional geoelectric models. By evaluating various performance indicators of neural networks that employ different wavelet basis functions as activation functions, we identify the most suitable wavelet basis functions. The quasi-resistivity is computed using both the wavelet neural network and the backpropagation neural network and then juxtaposed with traditional apparent resistivity. Our findings indicate that the wavelet neural network's quasi-resistivity aligns more closely with the resistivity of the real model. It is also more responsive to low resistivity anomalies than the conventional apparent resistivity translation algorithm. The wavelet approach moderates the undershoot or overshoot occurrences during abrupt stratum changes, offering a more accurate representation of subterranean electrical properties. When compared to the backpropagation neural network, the wavelet neural network provides a superior fit for the model, rendering a smoother quasi-resistivity depth curve. Therefore, it stands out as an improved method for pseudo-resistivity imaging. By processing survey data via the trained wavelet neural network, we find that the all-time apparent resistivity and quasi-resistivity align well with real-world situations. The wavelet neural network prominently showcases the low-resistance calculation results, offering a broader resistivity range. This clarity enhances anomaly detection, confirming the wavelet neural network's applicability and practicality for airborne transient electromagnetic imaging.

SEMG-Based Identification of Hand Motion Commands Using Wavelet Neural Network Combined With Discrete Wavelet Transform

Surface electromyogram (sEMG) signals can be applied in medical, rehabilitation, robotic, and industrial fields. As a typical application, a myoelectric prosthetic hand is controlled by the sEMG signals of the amputee’s residual muscles. To improve the dexterity of the myoelectric prosthetic hand, additional hand motion commands need to be classified. The more sEMG sensors are used, the more hand motion commands can be classified. However, the amputee’s residual muscles are limited. In order to improve the practicability of the myoelectric prosthetic hand, it is critical to investigate the effective pattern recognition algorithms to deal with the sEMG signals detected by fewer sensors, while identifying as many hand motion commands as possible. Current pattern recognition algorithms for sEMG signals are challenged by limited recognition patterns and unsteady classification accuracy rates. To solve these dilemmas, we employed discrete wavelet transform (DWT) and wavelet neural network (WNN) algorithms to improve the pattern recognition effects of sEMG signals. In addition, the back propagation and gradient descent algorithms were utilized to train WNN. In this work, we only used three sEMG sensors to classify and recognize six kinds of hand motion commands. The maximum identification accuracy rate is 100%, and an average classification accuracy rate of the proposed WNN is 94.67%, which is substantially better than the artificial neural network (ANN) algorithm.

Rolling Bearing Fault Diagnosis based on Time-frequency Feature Parameters and Wavelet Neural Network

In this paper, we present a novel method in fault diagnosis of rolling bearing based on time-frequency feature parameters and wavelet neural network. First, the time feature parameters are extracted from the vibration signal. Then the empirical mode decomposition (EMD) is used to decompose the signals of rolling bearings into a number of intrinsic mode functions (IMFs), and then the IMF energy-torques could be calculated through the vibration signal. Finally, those time-frequency feature parameters are taken as fault samples to train wavelet neural network (WNN)．The analysis results from the experimental show that the time- frequency feature parameters and WNN is effective in rolling bearing fault diagnosis. This paper provides the theoretical foundation for fault diagnosis in rotary machines.

Design of an Intelligent Controller Based on Wavelet Neural Network to Improve the Stability of Power Systems

To damp the oscillations in a power system, a new intelligent controller is proposed. This controller is an online trained wavelet neural network controller (OTWNNC) in which adaptive learning rates derived by the Lyapunov stability are employed to guarantee the convergence of the proposed controller. During the online control process, the identification of system is not necessary, because of learning ability of the proposed controller. One of the proposed controller features is robustness to different operating conditions and disturbances. Moreover, the Prony method is used to obtain the exponential damping of power system oscillations in this paper. The test power system is a two-area four-machine system power. The simulation results show that the oscillations are satisfactorily damped out by the OTWNNC.

Accurate modeling of switched reluctance machine based on hybrid trained WNN

According to the strong nonlinear electromagnetic characteristics of switched reluctance machine (SRM), a novel accurate modeling method is proposed based on hybrid trained wavelet neural network (WNN) which combines improved genetic algorithm (GA) with gradient descent (GD) method to train the network. In the novel method, WNN is trained by GD method based on the initial weights obtained per improved GA optimization, and the global parallel searching capability of stochastic algorithm and local convergence speed of deterministic algorithm are combined to enhance the training accuracy, stability and speed. Based on the measured electromagnetic characteristics of a 3-phase 12/8-pole SRM, the nonlinear simulation model is built by hybrid trained WNN in Matlab. The phase current and mechanical characteristics from simulation under different working conditions meet well with those from experiments, which indicates the accuracy of the model for dynamic and static performance evaluation of SRM and verifies the effectiveness of the proposed modeling method.

Intelligent control of SVC using wavelet neural network to enhance transient stability

In order to enhance transient stability in a power system, a new intelligent controller is proposed to control a Static VAR compensator (SVC) located at center of the transmission line. This controller is an online trained wavelet neural network controller (OTWNNC) with adaptive learning rates derived by the Lyapunov stability. During the online control process, the identification of system is not necessary, because of learning ability of the proposed controller. One of the proposed controller features is robustness to different operating conditions and disturbances. The test power system is a two-area two-machine system power. The simulation results show that the oscillations are satisfactorily damped out by the OTWNNC.

Speech Enhancement Using Wavelet Neural Network with Sub-Band Adaptive Matched Filter

Wavelet Neural Network (WNN) Is a Time-frequency Analysis Method, which Detects the Subtle Small Changes in the Signal Frequency Domain. Adaptive Filter Provides a Kind of Simple and Applied Method for Processing Signals in Noise. in this Paper, we Proposed a New Speech Enhancement Technique which Is Based on Wavelet Neural Network Using Adaptive Matched Filter Adjusting Weight. we Choose the Signal with Noise Pollution as the Input Signal and then Put it to the Trained Wavelet Neural Network. Wavelet Decomposition and Wavelet Neural Network Weights Processing Adopt Signal Sub-band Adaptive Matched Filter, the Output Signal of Wavelet Neural Network Is an Approximation Form of Original Signal. the Results Show that the WNN Is a Quite Effective Method for the Speech Enhancement and Improving the Ration of Signal to Noise.

Bacterial foraging trained wavelet neural networks: application to bankruptcy prediction in banks

This paper proposes a modified bacterial foraging technique (BFT) to train wavelet neural network (WNN) in order to predict bankruptcy in banks. The BFT is modified in that the swarming step that implements cell-to-cell interaction is deleted. The parameters translation, dilation and the weights connecting different layers in WNN are updated using the BFT. The resulting neural network is called BFTWNN. The effectiveness of BFTWNN is tested on bankruptcy prediction as well as benchmark datasets. We employed ten-fold cross validation in the study. Numerical experiments suggested that the BFTWNN outperformed threshold accepting trained wavelet neural network (TAWNN) (Vinaykumar et al., 2008) and WNN in benchmark datasets by wide margin while it yielded results comparable to that of differential evolution wavelet neural network (DEWNN) (Chauhan et al., 2009) in terms of area under the receiver-operating characteristic curve (AUC). Of particular significance is the superiority of the BFTWNN over the original WNN on all but one datasets.

Differential evolution trained radial basis function network: application to bankruptcy prediction in banks

In this paper, we propose differential evolution (DE) to train the supervised part of the radial basis function (RBF) network in the soft computing paradigm. Here the unsupervised part of the RBF is taken care of by K-means clustering. The new network is named as differential evolution trained radial basis function (DERBF) network. The efficacy of DERBF is tested on bank bankruptcy datasets viz. Spanish banks, Turkish banks, US banks and UK banks as well as benchmark datasets such as iris, wine and Wisconsin breast cancer. The performance of DERBF is compared with that of differential evolution trained wavelet neural networks (DEWNN) (Chauhan et al., 2009), threshold accepting trained wavelet neural network (TAWNN) (Vinaykumar et al., 2008) and wavelet neural network with respect to the criterion area under receiver operating characteristic curve. The results showed that DERBF is very good at generalisation in the ten-fold cross validation for all datasets.

SOFTWARE RELIABILITY PREDICTION USING INTELLIGENT TECHNIQUES: APPLICATION TO OPERATIONAL RISK PREDICTION IN FIRMS

In this paper, operational risk arising from the technological dimension is effectively modeled by efficiently forecasting software reliability. We propose the use of wavelet neural networks (WNN) to predict software reliability. Two kinds of wavelets were employed in WNN as transfer functions, viz. Morlet wavelet and Gaussian wavelet, thus giving rise to two variants of WNN. The effectiveness of WNN is demonstrated on a data set taken from literature. Its performance is compared with that of multiple linear regression, multivariate adaptive regression splines, back propagation trained neural network, threshold-accepting trained neural network, threshold accepting trained wavelet neural network, pi-sigma network, general regression neural network, dynamic evolving neuro-fuzzy inference system and TreeNet in terms of normalized root mean square error obtained on test data. Based on the experiments performed, it is observed that the WNN-based models outperformed all the other techniques.

Differential evolution trained wavelet neural networks: Application to bankruptcy prediction in banks

In this study, differential evolution algorithm (DE) is proposed to train a wavelet neural network (WNN). The resulting network is named as differential evolution trained wavelet neural network (DEWNN). The efficacy of DEWNN is tested on bankruptcy prediction datasets viz. US banks, Turkish banks and Spanish banks. Further, its efficacy is also tested on benchmark datasets such as Iris, Wine and Wisconsin Breast Cancer. Moreover, Garson’s algorithm for feature selection in multi layer perceptron is adapted in the case of DEWNN. The performance of DEWNN is compared with that of threshold accepting trained wavelet neural network (TAWNN) [Vinay Kumar, K., Ravi, V., Mahil Carr, & Raj Kiran, N. (2008). Software cost estimation using wavelet neural networks. Journal of Systems and Software] and the original wavelet neural network (WNN) in the case of all data sets without feature selection and also in the case of four data sets where feature selection was performed. The whole experimentation is conducted using 10-fold cross validation method. Results show that soft computing hybrids viz., DEWNN and TAWNN outperformed the original WNN in terms of accuracy and sensitivity across all problems. Furthermore, DEWNN outscored TAWNN in terms of accuracy and sensitivity across all problems except Turkish banks dataset.

Application of bacterial foraging technique trained artificial and wavelet neural networks in load forecasting

A new load forecasting (LF) approach using bacterial foraging technique (BFT) trained wavelet neural network (WNN) is proposed in this paper. Artificial neural network (ANN) is combined with wavelet transform called wavelet neural network is applied for LF. The parameters of translation and dilation in the wavelet nodes and the weighting factors in the weighting nodes are tuned using BFT optimization. With the advantages of global search abilities of BFT as well as the multiresolution and localizing natures of wavelets, the networks are constructed which identifies the inherent non-linear characteristics of power system loads. The proposed approach is validated with Tamil Nadu Electricity Board (TNEB) system, India. The comparison of Delta Rule and BFT-based LF for different periods are depicted with their mean absolute percentage errors (MAPE).

Image-based robot motion simulation

This paper describes an image-based robot animation technique. The technique utilizes eigen space method to generate compact representations of images from a set of reference images. A wavelet neural network is used to learn the relationship between robot joints and image representations. The learning algorithm is Quasi-Newton algorithm with Levenberg–Marquardt modifications. The trajectory in the joint space is first planned to generate a joint sequence. A corresponding sequence of representations can be obtained through the trained wavelet neural network and the image sequence can be synthesized using the result of eigen space method. No a priori models of robot and camera or any calibrations are needed. The virtual motion of the robot UP6 is demonstrated.

Intelligent control of induction servo motor drive via wavelet neural network

Abstract This study presents an intelligent control system for an induction servo motor drive to track periodic commands using a wavelet neural network (WNN). With the field orientation mechanism, the dynamic behavior of the induction servo motor drive system is rather similar to a linear system. However, the uncertainties, such as mechanical parametric variation, external disturbance, unstructured uncertainty due to nonideal field orientation in transient state, and unmodelled dynamics in practical applications influence the control performance. Therefore, an intelligent control system that is an on-line trained WNN controller with adaptive learning rates is proposed to control the rotor position of the induction servo motor drive. The adaptive learning rates are derived in the sense of discrete-type Lyapunov stability theorem, so that the convergence of the tracking error can be guaranteed in the closed-loop system. In the whole design process, the strict constrained conditions and prior knowledge of the controlled plant are not necessary according to the powerful learning ability of the intelligent control system. With the proposed intelligent control system, the controlled induction servo motor drive possesses the advantages of good tracking control performance and robustness to uncertainties under wide operating ranges. The effectiveness of the proposed control scheme is verified by both simulated and experimental results.