Recurrent Wavelet Neural Network Research Articles

Online-Growing Neural Network Control for Dissolved Oxygen Concentration

The activated sludge method is one of the most commonly used methods for wastewater treatment process (WWTP). Among them, the dissolved oxygen (DO) concentration is a key factor affecting microbial metabolism and wastewater treatment effectiveness. However, due to the nonlinearity and dynamicity of WWTP, it is difficult to precisely control the DO concentration by traditional control methods. To solve this problem, the online-growing pipelined recurrent wavelet neural network (OG-PRWNN) control is proposed to improve the DO control accuracy. First, the online growing mechanism is designed to adjust the number of modules of the controller by measuring the control performance. Then, the structure of the controller is automatically determined to meet the different operating conditions of the WWTP. Second, the online algorithm of parameters incorporating adaptive learning rates is designed to train the OG-PRWNN to meet the control requirements. In addition, the stability of the OG-PRWNN controller is analyzed by the Lyapunov stability theorem. Finally, the performance of the controller is verified by the benchmark simulation model of WWTP. Simulation results show that the OG-PRWNN controller can obtain better control accuracy.

Intelligent Energy Management System Evaluation of Hybrid Electric Vehicle Based on Recurrent Wavelet Neural Network and PSO Algorithm

Intelligent Energy Management System Evaluation of Hybrid Electric Vehicle Based on Recurrent Wavelet Neural Network and PSO Algorithm

Group-based bi-directional recurrent wavelet neural network for efficient video super-resolution (VSR)

Video super-resolution (VSR) is an important technology for enhancing the quality of video frames. The recurrent neural network (RNN)-based approach is suitable for sequential data because it can use accumulated temporal information. However, since existing methods only tend to capture slow and symmetrical motion with low frame rate, there are still limitations to restore the missing details for more dynamic motion. Most of the previous methods using spatial information treat different types of the spatial features identically. It leads to lack of obtaining meaningful information and enhancing the fine details. We propose a group-based bi-directional recurrent wavelet neural network (GBR-WNN) to exploit spatio-temporal information effectively. The proposed group-based bi-directional RNN (GBR) framework is built on the well-structured process with the group of pictures (GOP). In a GOP, we resolves the low-resolution (LR) frames from border frames to center target frame. Because super-resolved features in a GOP are cumulative, neighboring features are improved progressively and asymmetrical motion can be dealt with. Also, we propose a temporal wavelet attention (TWA) adopting attention module for both spatial and temporal features simultaneously based on discrete wavelet transform. Experiments show that the proposed scheme achieves superior performance compared with state-of-the-art methods.

Self-organizing pipelined recurrent wavelet neural network for time series prediction

The pipelined recurrent wavelet neural network (PRWNN) with nested structure has been widely used for time series prediction. However, selecting the appropriate network structure for different problems is still an unsolved problem. To solve this issue, the self-organizing PRWNN (SPRWNN) is proposed. Different from the PRWNN with fixed structure, the number of modules and hidden neurons of SPRWNN can be automatically determined during the training process. Firstly, the number of hidden neurons in each module of the SPRWNN is adjusted by the spiking strength (SS) of nodes. Then, the module growth mechanism is designed by measuring the performance of the SPRWNN. Furthermore, the convergence of SPRWNN is proved theoretically. The experimental results show that the SPRWNN can not only automatically adjust the network structure, but also improves the prediction accuracy by about 40% as compared with the PRWNN with fixed structure.

Speed Control of BLDC Motor With Ripple Effect Reduction Using Recurrent Wavelet Neural Networks

The interest in research and use of the Brushless DC (BLDC) motor has increased significantly over the last few years, due to development of electric vehicles (EV) as well as to its use in other growing areas. With further predicted growth of the EV industry, it is expected that the BLDC motor will become even more significant. BLDC motor is a highly nonlinear, dynamic system, which in terms of its control makes tuning of controllers difficult, especially when high precision is required. Because of its operating principle BLDC motor also has issues with torque ripple which affects the speed response in a speed controlled system. In this paper an intelligent controller that combines a proportional-integral-derivative (PID) controller with a recurrent wavelet neural network is proposed. By using an intelligent controller, as an adition to a standard PID, the approximation ability of neural networks is used not only to deal with the nonlinearity of the system, but also to reduce the torque ripple influence on the speed response. By computer simulations it is verified that the proposed controller is able to completely neutralize the ripple effect and achieve good transient speed response. It is also shown that improved robustness against load disturbance compared to the system containing only a PID controller is achieved.

Exploiting Wavelet Recurrent Neural Networks for satellite telemetry data modeling, prediction and control

Multidimensional times series prediction is a challenging task. Only recently the increased data availability has made it possible to tackle with such problems. In this work we devised a novel method to exploit the multiple correlated features in the time series. The recurrent neural networks and the wavelet transform have been important innovations in the fields of signal processing and time series prediction. This paper proposes a Wavelet Recurrent Network for multi-steps ahead prediction of multidimensional time series. The proposed model combines these two elements into a neural network that predicts multiple samples in the future that are multiple time steps ahead with respect to the input samples. This Wavelet Recurrent Network carries out a multiresolution decomposition of the input signal through the wavelet transform, predicts the future wavelet coefficients with the recurrent neural network and transforms the output back in the time domain. The proposed model is applied to the prediction of satellite telemetry data, that is composed of readings from multiple sensors which are highly correlated. The prediction of such telemetries can help the engineers to detect anomalies in the system, that, in the context of space missions, are particularly dangerous since they can compromise the entire mission if not handled properly. The results show that the proposed model outperforms the recurrent network without wavelet transform both in terms of accuracy and in the width of the forecast horizon.

Effluent ammonia nitrogen prediction using a phase space reconstruction method combining pipelined recurrent wavelet neural network

In wastewater treatment processes (WWTPs), the effluent ammonia nitrogen (NH4-N) concentration is a significant index to measure effluent quality. Recently, the soft computing methods have been widely used to measure effluent NH4-N concentration. However, the performance of soft computing method is closely related with its input variables, which is difficult to choose. As an alternative, the time series prediction method PSR-PRWNN is proposed, which combines phase space reconstruction (PSR) technique and pipelined recurrent wavelet neural network (PRWNN). Different from soft computing methods, the time series prediction method is a method which predicts the effluent NH4-N concentration by using its history data rather than other variables data. Firstly, the chaotic characteristics of effluent NH4-N time series is proved by using the correlation dimension method. Based on chaotic characteristics, the phase space of effluent NH4-N concentration is reconstructed by PSR technique. Then, the relationship model between inputs and output of the reconstructed phase space is established by PRWNN. Thirdly, the parameters of PRWNN are trained by an online gradient algorithm with adaptive learning rates. Finally, the experimental results indicate that the PSR-PRWNN can obtain better training results and prediction accuracy than other algorithms.

Grey Wolf Optimizer in Design Process of the Recurrent Wavelet Neural Controller Applied for Two-Mass System

In this paper, an adaptive speed controller of the electrical drive is presented. The main part of the control structure is based on the Recurrent Wavelet Neural Network (RWNN). The mechanical part of the plant is considered as an elastic connection of two DC machines. Oscillation damping and robustness against parameter changes are achieved using network parameters updates (online). Moreover, the various combinations of the feedbacks from the state variables are considered. The initial weights of the neural network and the additional gains are tuned using a modified version of the Grey Wolf Optimizer. Convergence of the calculation is forced using a new definition. For theoretical analysis, numerical tests are presented. Then, the RWNN is implemented in a dSPACE card. Finally, the simulation results are verified experimentally.

Segmentation Free Text Recognition for Overlapping Characters using Spectral Features and Bidirectional Recurrent Wavelet Neural Network

Segmentation Free Text Recognition for Overlapping Characters using Spectral Features and Bidirectional Recurrent Wavelet Neural Network

Segmentation free text recognition for overlapping characters using spectral features and bidirectional recurrent wavelet neural network

Segmentation free text recognition for overlapping characters using spectral features and bidirectional recurrent wavelet neural network

Adaptive recurrent NeuroFuzzy control for power system stability in smart cities

A smart city is a dynamic and sustainable urban system that provides a great quality of service to its residents by optimally managing its resources. In smart cities, low-frequency oscillations are a serious concern to the power system as they adversely affect system’s stability. Moreover, power system stabilizers are inefficient due to their fixed-parameter architecture. Flexible ac transmission system controller performs effective damping of low-frequency oscillations when provided with suitable supplementary damping control like a Static synchronous series compensator. In this context, the paper proposes an adaptive neuro-fuzzy recurrent wavelet control for smart cities that uses the recurrent Gaussian membership function and recurrent wavelet neural network in antecedent and consequent parts respectively. The paper applies the gradient descent optimization with a back-propagation algorithm to update the parameters of suggested adaptive neuro-fuzzy recurrent wavelet control. Simulations are performed on two test systems, both showing the effectiveness of the recommended control scheme as compared to traditional lead-lag control and artificial neuro-fuzzy Takagi Sugeno Kang control. Calculations of performance index for various fault scenarios lead to the conclusion that the control scheme can damp oscillations effectively and enhances the power system’s transient stability.

An integrated feedforward-feedback control structure utilizing a simplified global gravitational search algorithm to control nonlinear systems

This paper presents an integrated feedforward-feedback control structure to control nonlinear dynamical systems. This intelligent control system exploits a modified recurrent wavelet neural network (MRWNN) in the feedforward (FF) and the feedback (FB) loops of the control structure. Specifically, the MRWNN is proposed to boost the approximation performance of a previously reported network by employing two amendments to the original structure. To optimize the parameters of both the FF and the FB controllers, an enhanced version of the gravitational search algorithm (GSA) is developed to improve the searching capability of the original algorithm. In particular, two modifications were adopted, including the removal of two control parameters related to the gravitational constant in the original algorithm and the utilization of the global best solution to constitute the next generation of agents. Hence, the proposed algorithm is called the simplified global gravitational search algorithm (SGGSA), which has demonstrated better optimization performance compared to those of other techniques, including the original GSA. By conducting several evaluation tests using different nonlinear time-variant dynamical systems, the effectiveness of the proposed control structure was confirmed in terms of control precision and robustness against external disturbances. In addition, the MRWNN has exhibited a superior control performance compared with other related controllers.

Stochastic recurrent wavelet neural network with EEMD method on energy price prediction

Novel hybrid neural network prediction model (denoted by E-SRWNN) is formed by combining ensemble empirical mode decomposition (EEMD) and stochastic recurrent wavelet neural network (SRWNN), in order to improve the precision of energy indexes price forecasting. Energy index price series are non-stationary, nonlinear and random. EEMD method is utilized to decompose the closing prices of four energy indexes into subsequences with different frequencies, and the SRWNN model is composed by adding stochastic time effective function and recurrent layer to the wavelet neural network (WNN). Stochastic time effective function makes the model assign different weights to the historical data at different times, and the introduction of recurrent layer structure will enhance the data learning. In this paper, E-SRWNN model is compared with other WNN-based models and the deep learning network GRU. In the error evaluation, the general standards, such as linear regression analysis, mean absolute error and theil inequality coefficient, are utilized to compare the predicted effects of different models, and then multiscale complexity-invariant distance is applied for further analysis. Empirical research illustrates that the proposed E-SRWNN model displays strong forecasting ability and accurate forecasting results in energy price series forecasting.

Learning traffic as a graph: A gated graph wavelet recurrent neural network for network-scale traffic prediction

Network-wide traffic forecasting is a critical component of modern intelligent transportation systems for urban traffic management and control. With the rise of artificial intelligence, many recent studies attempted to use deep neural networks to extract comprehensive features from traffic networks to enhance prediction performance, given the volume and variety of traffic data has been greatly increased. Considering that traffic status on a road segment is highly influenced by the upstream/downstream segments and nearby bottlenecks in the traffic network, extracting well-localized features from these neighboring segments is essential for a traffic prediction model. Although the convolution neural network or graph convolution neural network has been adopted to learn localized features from the complex geometric or topological structure of traffic networks, the lack of flexibility in the local-feature extraction process is still a big issue. Classical wavelet transform can detect sudden changes and peaks in temporal signals. Analogously, when extending to the graph/spectral domain, graph wavelet can concentrate more on key vertices in the graph and discriminatively extract localized features. In this study, to capture the complex spatial-temporal dependencies in network-wide traffic data, we learn the traffic network as a graph and propose a graph wavelet gated recurrent (GWGR) neural network. The graph wavelet is incorporated as a key component for extracting spatial features in the proposed model. A gated recurrent structure is employed to learn temporal dependencies in the sequence data. Comparing to baseline models, the proposed model can achieve state-of-the-art prediction performance and training efficiency on two real-world datasets. In addition, experiments show that the sparsity of graph wavelet weight matrices greatly increases the interpretability of GWGR.

Wind speed forecasting based on wavelet transformation and recurrent neural network

AbstractWith the increase in power demand, renewable energy resources such as wind energy have been developed as one of the fastest growing energy sources. However, the wind power generation system depends on the availability of wind flow. The intermittence nature of wind speed is the most serious concern in wind energy application into the existing power system. Hence, wind speed forecasting approaches are proposed in order to deal with these problems. In this paper, a hybrid model of wind speed forecasting is developed. The proposed hybrid model consists of two steps: the first one is decomposition of wind speed sample data by wavelet technique, and the second step uses these decomposed data to estimate wind speed through recurrent wavelet neural network (RWNN). To validate the proposed model, it is compared with the conventional recurrent neural network (RNN) prediction structure. The obtained results based on real data provide the effectiveness of the proposed model in terms of mean absolute error and the rate of convergence parameter.

Assessment of Aerial Target Threat Based on Genetic Algorithm Optimizing Fuzzy Recurrent Wavelet Neural Network

In target threat assessment of air combat, the evaluation system model is usually nonlinear and the assessment which is difficult to obtain also has some uncertain information. In order to effectively solve these problems, the Single-hidden-layer Fuzzy Recurrent Wavelet Neural Network Optimized by Genetic Algorithm (GA-SLFRWNN) is presented in this paper. In this new method, the influence factors for assessment and the ambiguity of their information are first analyzed. The RWNN are embed in the back part of FNN (fuzzy neural network) for the purpose of enhancing self-learning ability. Then GA is used to optimize the initial parameters of the model and the optimal learning rate based on Lyapunov theory is proposed. The simulation results show that the proposed algorithm improves the stability of the evaluation system, accelerates the convergence speed and enhances the prediction accuracy compared with the FNN and SLFRWNN.

Adaptive recurrent neural network motion control for observation class remotely operated vehicle manipulator system with modeling uncertainty

Precise motion control of remotely operated vehicles plays an important role in a great number of submarine missions. However, the high-performance operations are difficult to realize due to the uncertainty in system modeling with self-disturbance. On the basis of the multibody system dynamics, self-disturbances from the tether and manipulator have been systematically analyzed in order to transform them into observed forces. A novel S surface–based adaptive recurrent wavelet neural network control system has been proposed on the nonlinear control of underwater vehicles, with its recurrent wavelet neural network structure designed for the approximation of the uncertain dynamics. Moreover, a robust function has been proposed to improve system robustness and convergence. The comparison shows that the remotely operated vehicle operation performance including the three-dimensional path following and vehicle-manipulator coordinate control has been greatly improved.

Motion control of linear induction motor based on optimal recurrent wavelet neural network-PID controller

This paper contains a proposed controller based on optimal recurrent wavelet neural network (RWNN) with PID controller to control the velocity and hence the stator current as well as the developed thrust of three phase linear induction motor (LIM) which consider the end effects. A vector control represented by indirect field oriented control (IFOC) technique is appointed to attain velocity and flux control for different loading conditions. Moreover, a voltage source inverter based on space vector pulse width modulation (SVPWM) is utilized to give the required stator voltage of LIM. A well-known particle swarm optimization (PSO) algorithm is employed for online tuning of the proposed controller. The computer simulation results show that this controller is effective and gives preferable and rigorous performance compared with a results obtained from conventional wavelet neural network (WNN) and PID controllers.

In-move aligned SINS/GNSS system using recurrent wavelet neural network (RWNN)-based integration scheme

Advances in micro-electro mechanical system (MEMS) technology bring about revolutionary changes in autonomous vehicle navigation. As a new development, strap-down inertial navigation system (SINS) is effectively combined with global navigation satellite system (GNSS) to construct an integrated SINS/GNSS system. However, time-growing navigation error is the main challenge of using MEMS-grade inertial measurement unit (IMU) in the SINS/GNSS system. Failure of un-accounted inertial sensor error causes a rapid degradation in the overall performance of low-cost SINSs. This paper aims to enhance the long-term performance of low-cost MEMS-grade SINS/GNSS navigation system. A new integration scheme is presented for in-move aligned SINS/GNSS system. Un-modeled nonlinearities in the SINS dynamics as well as error uncertainties in the measurements of MEMS-grade IMU motivate using a robust data fusion algorithm for the proposed integration scheme. Considering these facts, a new recurrent wavelet neural network (RWNN)-based algorithm is designed for data fusion in the proposed in-move aligned SINS/GNSS system. Several vehicular field tests have been carried out to assess the long-term performance and accuracy of the proposed navigation algorithm.

Forecasting energy fluctuation model by wavelet decomposition and stochastic recurrent wavelet neural network

Abstract Forecasting the fluctuations of global energy markets has become a focus of economic and energy research. In this paper, in an attempt to improve the prediction accuracy of energy prices, a novel hybrid neural network is developed through combining discrete wavelet transform (DWT) and stochastic recurrent wavelet neural network (SRWNN). The DWT is utilized as a processing technique to decompose subseries with different frequency, and the SRWNN model is established based on the randomization of wavelet neural network (WNN), which considers the memory of historical events and the weights of historical data depending on their occurrence time. The empirical experiments are performed in the prediction of four energy market prices, and the effectiveness of proposed DWT-SRWNN model is presented through contrastive results of the different predictive models. Further, a novel approach called multi-scale composite complexity synchronization (MCCS) is applied to display and evaluate the predictive effect. The empirical results demonstrate a higher accuracy of the proposed hybrid model in global energy price series forecasting.