Function Neural Network Research Articles

SPATIOTEMPORAL SEGMENTATION OF SATELLITE IMAGE TIME SERIES USING SELF-ORGANIZING MAP

Abstract. Nowadays, researchers have free access to an unprecedentedly large amount of remote sensing images collected by satellites and sensors with different spatial, temporal, and spectral resolutions. This scenario has promoted the use of satellite image time series for spatiotemporal analysis. This paper presents a methodology for spatiotemporal segmentation of satellite image time series. Spatiotemporal segmentation finds homogeneous regions in space and time from remote sensing images based on spectral features. The proposed approach is unsupervised based on the self-organizing map (SOM) neural network and hierarchical clustering algorithm. It was implemented and applied to a region in the Mato Grosso state, Brazil. The results were evaluated using qualitative and quantitative approaches. In the qualitative approach, visual analysis was performed based on the land use and land cover map of the TerrraClass Cerrado project. In the quantitative approach, supervised and geometric metrics were used to analyze the quality of the produced segments. The results obtained are promising since the segments produced were homogeneous and with a low occurrence of over-segmentation.

LSTM-Based RNN Framework to Remove Motion Artifacts in Dynamic Multicontrast MR Images with Registration Model

Today, many people under the age of 10 are being examined for brain-related issues, including tumours, without displaying any symptoms. It is not unusual for children to develop brain-related concerns such as tumours and central nervous system disorders, which may affect 15% of the population. Medical experts believe that the irregular eating habits (junk food) and the consumption of pesticide-tainted fruits and vegetables are to blame. The human body is naturally resistant to harmful gears, but only up to a point. If it exceeds the limit, a cell manipulation process is automatically initiated that can remove dangerous inactive tissues from the cell membrane and later grows into tumour blockage in the human body. Thus, the adoption of an advanced computer-based diagnostic system is highly recommended in order to generate visually enhanced images for anomaly identification and infectious tissue segmentation. In most cases, an MR image is chosen since it is easier to distinguish between affected and nonaffected tissue. Conventional convolution neural network (CCNN) mapping and feature extraction are difficult because of the vast volume of data. In addition, it takes a lengthy time for the MRI scanning process to obtain diverse positions for anomaly identification. Aside from the discomfort, the patient may experience motion abnormalities. Recurrent neural network (RNN) classifies tumour regions into several isolated portions much faster and more accurately, so that it can be prevented. To remove motion artefacts from dynamic multicontrast MR images, a novel long short-term memory- (LSTM-) based RNN framework is introduced in this research. With this method, the MR image’s visual quality is improved over CCNN while simultaneously mapping a larger volume and extracting more quiet characteristics than CCNN can. DC-CNN, SMSR-CNN, FMSI-CNN, and DRCA-CNN results are compared. For both low and high signal-to-noise ratios, the suggested LSTM-based RNN framework has gained reasonable feature intelligibility (SNRs). In comparison to previous approaches, it requires less computing and has higher accuracy when it comes to detecting infected portions.

A Full-Sample Clustering Model Considering Whole Process Optimization of Data

With the continuous increase of data volume and data dimensions, it becomes more and more difficult to improve the accuracy and interpretability of the algorithm only from the clustering algorithm itself. In order to improve the accuracy of the clustering algorithm and improve the interpretability of the clustering results, we propose an Improved feature selection and combined clustering model considering whole process optimization. In this model, we processed the data from the whole process of data mining and carried out clustering analysis. Firstly, we started data preprocessing, and then used the feature selection algorithm of text weight + principal component analysis (PCA) to reduce the feature dimension and obtain important features and data sets for clustering. Secondly, we used the improved Self organizing maps (SOM) neural network and K-means clustering combination model to perform clustering analysis and established clustering algorithm evaluation indicators. Thirdly, we used collaborative filtering to cluster data sets that included missing data to ensure that all sample data can obtain results. Finally, through case analysis, it was verified that the model proposed in this paper had high clustering accuracy and interpretability.

A Dynamic ECMS Method Considering Vehicle Speed Pattern and Minimum Engine Operation Time for a Range-Extender Electric Vehicle (Jan. 2022)

A new equivalent consumption minimization strategy (ECMS) is proposed for a hybrid electric vehicle in which the equivalent factor varies dynamically with the current speed and the minimum engine operation time that leads to improvement in vehicle fuel economy and battery performance. To construct the strategy, a self-organizing map neural-network was utilized to recognize and group the speed patterns. The driving cycles corresponding to these speed patterns were reconstructed for solving the corresponding optimal equivalent factors. When applying the strategy on the road, the current vehicle speed pattern is recognized by using the historical vehicle speed data, and the corresponding optimal equivalent factor is adopted based on that pattern. Furthermore, to constrain the battery state of charge (SOC) to avoid exceeding the allowable values and to consider the hysteresis of engine starting, a fuzzy logic controller was used to dynamically adjust the minimum engine operation time. The simulated results indicated that the proposed ECMS improved the vehicle fuel economy by 7.87% and 6.23% over the conventional ECMS and the rule-based single-point strategy, respectively. Furthermore, the peak battery charging-current was reduced by 42.94% and 27.73%, and the peak discharging-current by 62.19% and 56.97%, respectively. In addition, the proposed strategy could lead to better SOC evolution of the battery.

Performance Analysis of Chaotic Neural Network and Chaotic Cat Map Based Image Encryption

Nowadays, chaotic systems are used quite often, especially in image encryption applications. Hypersensitivity to the initial conditions, limited field-changing signs and irregular movements make these systems one of the critical elements in scientific matters such as cryptography. These systems are divided in two parts as discrete time and continuous time in terms of their dimensions and properties. Gray level image encryption applications generally use one-dimensional and color image encryption applications generally use multi-dimensional chaotic systems. In this study, Tent Map, Cat Map, Lorenz, Chua, Lu chaotic systems were used for chaotic neural network based image coding application and Logistic Map and 3D Cat Map chaotic systems were used for 3D chaotic Cat Map based image encryption application. The encrypted image and the original image were examined by various analysis methods. As a result of these analyzes, it was seen that both applications gave very successful results. Analyzes have shown that the chaotic neural network based image coding algorithm is more secure and successful.

Evsel Elektriksel Yük Profilleri için Bulanık Mantık ve Yapay Sinir Ağları Teknikleri ile İki-Kademeli Kümeleme Yaklaşımı

In this paper, household electricity load profile (LP) clustering problem is addressed. LP clustering analysis has been utilized as predicted end-user LPs for demand or supply management strategies to maintain the stability of the power systems. The consumption dynamics of the LPs are formed by the combinations of technical and social factors. Hence, discovering the dynamic patterns of the LPs has been a challenging problem. For this problem, we have offered successive applications of Sugeno fuzzy-logic (SFL) and self-organizing map neural network (SOMNN) techniques. Firstly, the data sets of the LPs are clustered by fuzzy logic approach by the reference models which are generated with the common family-types per persons. Then, considering the extra input of the weighted occupancy profiles, SOMNN is performed to improve the clustering result according to the dataset. The proposed strategy has been simulated by MATLAB® and the related results are presented.

Optimization of the Battery Pack Heat Dissipation Structure of a Battery-Type Loader

The development of a battery-type loader is an important research direction in the field of industrial mining equipment. In the energy system, the battery will inevitably encounter the problem of heat dissipation when using high-power electricity. In this study, we took the power battery pack of a 3 m3 battery-type underground loader as the research object. The influence of single factors, such as the position of the air outlet of the battery pack, the size of the air outlet, the width of the separator, and the reverse plate, on the heat dissipation characteristics of the battery pack were studied. Then, a prediction model between the structural parameters and temperature was established using a radial basis function (RBF) neural network. This prediction model was then used as an adaptation evaluation model for global optimization through the multiobjective particle swarm optimization (PSO) algorithm, using which the optimal combination of structural parameters was obtained. The maximum temperature of the battery pack after optimization was reduced by 22%, compared to that before optimization, and the average temperature was reduced by 12.5%. Overall, the heat dissipation effect significantly improved. The optimization results indicate that the method proposed in this paper is feasible for use in optimizing battery heat dissipation systems in electric vehicles, thus providing a reference for research related to battery pack heat dissipation.

Smooth-Switching Gain Based Adaptive Neural Network Control of n-Joint Manipulator with Multiple Constraints

Modeling errors, external loads and output constraints will affect the tracking control of the n-joint manipulator driven by the permanent magnet synchronous motor. To solve the above problems, the smooth-switching for backstepping gain control strategy based on the Barrier Lyapunov Function and adaptive neural network (BLF-ANBG) is proposed. First, the adaptive neural network method is established to approximate modeling errors, unknown loads and unenforced inputs. Then, the gain functions based on the error and error rate of change are designed, respectively. The two gain functions can respectively provide faster response speed and better tracking stability. The smooth-switching for backstepping gain strategy based on the Barrier Lyapunov Function is proposed to combine the advantages of both gain functions. According to the above strategy, the BLF-ANBG strategy is proposed, which not only solves the influence of multiple constraints, unknown loads and modeling errors, but also enables the manipulator system to have better dynamic and steady-state performances at the same time. Finally, the proposed controller is applied to a 2-DOF manipulator and compared with other commonly used methods. The simulation results show that the BLF-ANBG strategy has good tracking performance under multiple constraints and model errors.

A nonlinear process monitoring strategy with a 2-phase fault diagnosis approach

The article delves into the development of Statistical Fault Detection and Diagnosis Strategies for an Integrated Steel Plant (ISP) taking into account the nonlinear relationship amongst the monitored Process and Feedstock Characteristics. The strategies being devised is based on Neural Network Fitting model cum Principal Component Analysis based technique (NNF-PCA) and Kernel Principal Component Analysis (KPCA) based technique. For detection of fault(s) Hotelling T2 chart based on (Principal Component Analysis) PCA and KPCA scores were employed and an ensemble strategy amalgamating KPCA and Self Organizing Map neural network has also been proposed for the detection of the out-of-control observations or faults. The article also proposes a 2-phase Fault Diagnosis approach christened as Preliminary Diagnosis phase and Specific Diagnosis phase. The Preliminary Diagnosis phase is based on Pattern Analysis of the control chart monitoring statistic observations and the Specific Diagnosis phase is based on the employment of appropriate Fault Diagnostic Statistic. The Preliminary Diagnosis reveals the broader source of assignable cause for the onset of the fault(s) and the Specific Diagnosis reveals the relative contribution of the individual Process and Feedstock characteristics. An in-depth comparative analysis between the NNF-PCA based strategy and KPCA based strategy w.r.t. three comparative aspects and four comparative parameters were carried out with their findings being duly highlighted which revealed the slight effectiveness of the KPCA based strategy with respect to the NNF-PCA based counterpart.

An ANN-Fuzzy Cognitive Map-Based Z-Number Theory to Predict Flyrock Induced by Blasting in Open-Pit Mines

Blasting is widely employed as an accepted mechanism for rock breakage in mining and civil activities. As an environmental side effect of blasting, flyrock should be investigated precisely in open-pit mining operations. This paper proposes a novel integration of artificial neural network and fuzzy cognitive map (FCM) with Z-number reliability information to predict flyrock distance in open-pit mine blasting. The developed model is called the artificial causality-weighted neural networks, based on reliability (ACWNNsR). The reliability information of Z-numbers is used to eliminate uncertainty in expert opinions required for the initial matrix of FCM, which is one of the main advantages of this method. FCM calculates weights of input neurons using the integration of nonlinear Hebbian and differential evolution algorithms. Burden, stemming, spacing, powder factor, and charge per delay are used as the input parameters, and flyrock distance is the output parameter. Four hundred sixteen recorded basting rounds are used from a real large-scale lead–zinc mine to design the architecture of the models. The performance of the proposed ACWNNsR model is compared with the Bayesian regularized neural network and multilayer perceptron neural network and is proven to result in more accurate prediction in estimating blast-induced flyrock distance. In addition, the results of a sensitivity analysis conducted on effective parameters determined the spacing as the most significant parameter in controlling flyrock distance. Based on the type of datasets used in this study, the presented model is recommended for flyrock distance prediction in surface mines where buildings are close to the blasting site.

Nonlinear modeling of industrial boiler NOx emissions

ABSTRACT Pollutant emissions into the atmosphere are recognized as a significant problem in fossil fuel combustion. The pollution emission measurement in industrial boilers is difficult and expensive but fundamental for monitoring and controlling. Frequently continuous emissions monitoring (CEM) system is out of service or useless due to obsolescence, high maintenance cost, and so on or simply is not installed. When a system for measuring pollutant emissions is not available, an alternative method must be employed to get the pollutant emission value. According to the black-box model approach, this article describes the nonlinear modeling of NOx emissions from a utility boiler. Bayesian-Gaussian (BG), multilayer perceptron (MLP), and Volterra polynomial basis functions (VPBF) neural networks are developed for model benchmarking. Experimental data from a utility boiler was acquired in order to model definition and evaluation. The models process three boiler variables oxygen excess, fuel mass flow and flue gas recirculation gates for NOx emission estimation. Models with BG show better performance than models with MLP and VPBF for NOx prediction. Implications: The technology to control NOx emissions generated by combustion operates under strict regulations. In order to reduce NOx emissions, theoretical models of NOx generation have been studied extensively, including nitrogen chemistry and the dynamic flow of gas particles which is very complex. The new technology trends would require the continuous measurement of high precision NOx emissions to achieve further reductions in NOx emissions. Currently, NOx emissions are measured by a Continuous Emission Monitoring system, which turns out to be extremely expensive and difficult to maintain, so alternative low-cost solutions are desirable. Our contribution shows how algorithms based on different artificial intelligence techniques are viable and quality alternatives for the measurement of continuous NOx emissions. The NOx emissions models based on IA algorithms are viable alternatives that have versatility and self-tuning capacity due to the fact that they are based on boiler operation parameters which have valuable information few explored nowadays.

Design of Mayer Wavelet Neural Networks for Solving Functional Nonlinear Singular Differential Equation

In the present work, an advance computational intelligence paradigm based on functional Mayer artificial neural network (FM-ANN) is accessible for solving the singular nonlinear functional differential equation (NFDE) numerically. The solution of singular NFDE is performed by using the artificial neural networks (ANNs) optimized with global search genetic algorithm (GA) enhanced by local refinements of sequential quadratic (SQ) programming and the hybrid of GASQ programming. The proposed scheme is applied for solving three types of second-order singular NFDEs. In order to validate the correctness of the designed scheme, the comparison of the proposed and exact solutions has been performed. Moreover, the statistical interpretations are used to prove the worth, convergence, accuracy, stability, and robustness of FM-ANN-GASQP for the solution of singular NFDEs.

Research on an Improved SOM Model for Damage Identification of Concrete Structures

In order to solve the problem of intelligent detection of damage of modern concrete structures under complex constraints, an improved self-organizing mapping (SOM) neural network model algorithm was proposed to construct an accurate identification model of concrete structure damage. Based on the structure and algorithm of the SOM network model, the whole process of the core construction of the concrete structure damage identification network model is summarized. Combined with the damage texture characteristics of concrete structures, through the self-developed 3D laser scanning system, an improved method based on a small number of samples to effectively improve the effectiveness of network input samples is proposed. Based on the principle of network topology map analysis and its image characteristics, a SOM model improvement method that can effectively improve the accuracy of the network identification model is studied. In addition, based on the reactive powder concrete bending fatigue loading test, the feasibility and accuracy of the improved method are verified. The results show that the improved SOM concrete structure damage identification model can effectively identify unknown neuron categories in a limited sample space, and the identification accuracy of the SOM network model is improved by 4.69%. The proposed improved SOM model method fully combines the network topology and its unique image features and can accurately identify structural damage. This research contributes to the realization of high-precision intelligent health monitoring of damage to modern concrete structures. In addition, it is of great significance for the timely detection, identification and localization of early damage to structures.

Accurate prediction of water quality in urban drainage network with integrated EMD-LSTM model

Quickly and accurately grasping the water quality in the drainage network is essential for the management and early warning of the urban water environment. Modeling-based detection methods enable fast and reagent-free water quality detection based on inexpensive multi-source data, which is cleaner and more sustainable than traditional chemical-reaction-based detection methods. But the unsatisfactory accuracy limits their practical application. This study proposes an integrated EMD-LSTM model that combines the data preprocessing module centered on empirical mode decomposition (EMD) and the long short-term memory (LSTM) neural network prediction module to improve the accuracy of the modeling-based detection methods. In the integrated EMD-LSTM model, EMD allows retaining outliers and utilizing data on non-aligned moments, which contributes to capturing data patterns, while powerful nonlinear mapping and learning ability of LSTM neural network enables the time series prediction of water quality. As a result, the EMD-LSTM has achieved the highest R2 values (0.961, 0.9384, 0.9575, 0.9441, 0.9502) and the lowest RMSE values (8.3112, 6.7795, 0.2691, 2.6239, 1.4894) in the prediction of COD, BOD5, TP, TN, NH3–N when compared with the integrated models formed by combining other preprocessing procedures (i.e., traditional operation, short-time Fourier transform) and data-driven forecasting algorithms (i.e., partial least squares regression, gradient boosting regression, deep neural network). This study provides enlightenment for improving the accuracy of modeling-based detection methods, which has driven the development of water quality detection technology towards cleaner and more sustainable.

Design Method of Intelligent Ropeway Type Line Changing Robot Based on Lifting Force Control and Synovial Film Controller

Aiming at the problems of low efficiency, reliability, and safety of manual construction for demolition of old lines, a design method of an intelligent ropeway type line changing robot based on lifting force control and synovial film controller is proposed. First, the mechanical model of robot load and line sag is established, and the sag of the overhead line where the robot is located is used to calculate the jacking force that the jacking device needs to provide to the robot. Then, by introducing the radial basis function (RBF) neural network adaptive algorithm into the synovial controller, an adaptive sliding mode position control algorithm based on the RBF neural network is designed to achieve high-precision motion control of the robot in complex operating environments. Finally, based on the compactness, weight, and reliability of the robot, the optimal design is carried out from four aspects of topology, size, shape and morphology, and the design scheme of the robot for wire removal is proposed, and the robot is produced. The developed robot and the other three robots are compared and analyzed under the same conditions through simulation experiments. The results show that the maximum operating time, maximum climbing angle, and maximum traveling speed of the robot developed in this study are all optimal, which are 45 min, 10°, and 1 m/s respectively, and the performance is better than the other three comparison algorithms.

Oil Family Typing Using a Hybrid Model of Self-Organizing Maps and Artificial Neural Networks.

Identifying the number of oil families in petroleum basins provides practical and valuable information in petroleum geochemistry studies from exploration to development. Oil family grouping helps us track migration pathways, identify the number of active source rock(s), and examine the reservoir continuity. To date, almost in all oil family typing studies, common statistical methods such as principal component analysis (PCA) and hierarchical clustering analysis (HCA) have been used. However, there is no publication regarding using artificial neural networks (ANNs) for examining the oil families in petroleum basins. Hence, oil family typing requires novel and not overused and common techniques. This paper is the first report of oil family typing using ANNs as robust computational methods. To this end, a self-organization map (SOM) neural network associated with three clustering validity indexes was employed on oil samples belonging to the Iranian part of the Persian Gulf oilfields. For the SOM network, at first, 10 default clusters were selected. Afterward, three effective clustering validity coefficients, namely, Calinski–Harabasz (CH), Silhouette (SH), and Davies–Bouldin (DB), were studied to find the optimum number of clusters. Accordingly, among 10 default clusters, the maximum CH (62) and SH (0.58) were acquired for 4 clusters. Similarly, the lowest DB (0.8) was obtained for four clusters. Thus, all three validation coefficients introduced four clusters as the optimum number of clusters or oil families. According to the geochemical parameters, it can be deduced that the corresponding source rocks of four oil families have been deposited in a marine carbonate depositional environment under dysoxic–anoxic conditions. However, oil families show some differences based on geochemical data. The number of oil families identified in the present report is consistent with those previously reported by other researchers in the same study area. However, the techniques used in the present paper, which have not been implemented so far, can be introduced as more straightforward for clustering purposes in oil family typing than those of common and overused methods of PCA and HCA.

Research on Fractional-Order Global Fast Terminal Sliding Mode Control of MDF Continuous Hot-Pressing Position Servo System Based on Adaptive RBF Neural Network

In this paper, a novel fractional-order global fast terminal sliding mode control (FGFTSMC) strategy based on an adaptive radial basis function (RBF) neural network is proposed to improve the performance of a medium density fiberboard (MDF) continuous hot-pressing position servo system with parameter perturbation and external load disturbance. Primarily, the mathematical model of the MDF continuous hot-pressing position servo system is constructed based on the dynamic equation of the hydraulic system. Then, a FGFTSMC is designed to speed up the convergence rate of the system, in which an adaptive law is used to estimate the upper bound of the unknown parameters to overcome the existing parameter perturbation of the system. In addition, an RBF neural network is introduced to approximate the external load disturbance of the system. The stability of MDF continuous hot-pressing position servo system based on the control scheme developed in this paper is proven using the Lyapunov theory. Finally, the simulation results show that the presented control scheme can effectively ensure the tracking accuracy of the system and enhance the robustness of the system.

Adaptive tracking control of time‐varying delay switched nonstrict‐feedback nonlinear systems with unknown backlash‐like hysteresis

SummaryIn this article, the adaptive tracking control issue is addressed for uncertain switched nonstrict‐feedback nonlinear systems with unknown time‐varying delay and backlash‐like hysteresis under neural networks (NNs) estimation framework. In contrast to most existing results of nonlinear systems with time‐delay, the time‐varying delay we consider depends entirely on the number of switching subsystems. Based on the inherent properties of radial basis function NNs, the nonstrict‐feedback structure is effectively compensated. Meanwhile, with the help of the combination of Razumikhin function method, adaptive backstepping scheme, and dynamic surface control approach, the quadratic common Lyapunov function is established for all subsystems, and then an adaptive common control strategy is developed. For arbitrary switching signals, it is shown that all the signals in the closed‐loop system can be remained bounded and the system output can follow a desired signal under the designed controller. Finally, the validity of the presented strategy is verified via two simulation examples.

Design of the MVT RBF neural network robotic manipulator control system based on model block approximation

Due to the uncertain dynamic characteristics, the requirements for robotic manipulator control are increasingly complex. The traditional radial basis function (RBF) neural network has a good generalization ability, but its redundant and tedious training process cannot meet the “Intelligent” control requirement of robotic manipulator. This study designs a new valve-regulated memristive RBF neural network, which adopts the model block approximation control strategy to estimate the three coefficient matrices of the robotic manipulator and uses the memristor with voltage threshold (MVT) as an electronic synapse to provide connections between neurons for the neural network and store information. This study adopts the design idea of software hardening and replaces the updated neural network weight with the change of the memristance value in the MVT network (crossed array), which can effectively improve the control performance of the traditional RBF neural network and can also provide analytical data for the fault detection of the subsequent control system. A simulation analysis is conducted with a single-joint robotic manipulator as the control object, and the results verify the rationality and feasibility of the proposed control algorithm.

Design of extended Kalman filtering neural network control system based on particle swarm identification of nonlinear U-model

This paper studies the modelling of a class of nonlinear plants with known structures but unknown parameters and proposes a general nonlinear U-model expression. The particle swarm optimization algorithm is used to identify the time-varying parameters of the nonlinear U-model online, which solves the identification problem of the nonlinear U-model system. Newton iterative algorithm is used for nonlinear model transformation. Extended Kalman filter (EKF) is used as the learning algorithm of radial basis function (RBF) neural network to solve the interference problem in a nonlinear system. After determining the number of network nodes in the neural network, EKF can simultaneously determine the network threshold and weight matrix, use the online learning ability of the neural network, adjust the network parameters, make the system output track the ideal output, and improve the convergence speed and anti-noise capability of the system. Finally, simulation examples are used to verify the identification effect of the particle swarm identification algorithm based on the U-model and the effectiveness of the extended Kalman filtering neural network control system based on particle swarm identification.