Showing papers in "Turkish Journal of Electrical Engineering and Computer Sciences in 2016"
TL;DR: In this article, a CNN with fused convolutional sub-sampling architecture was proposed for finger-vein recognition, which achieved an accuracy of 99.38% with an error ratio of 80/20 percent for separation of training and test samples.
Abstract: A novel approach using a convolutional neural network (CNN) for finger-vein biometric identification is presented in this paper. Unlike existing biometric techniques such as fingerprint and face, vein patterns are inside the body, making them virtually impossible to replicate. This also makes finger-vein biometrics a more secure alternative without being susceptible to forgery, damage, or change with time. In conventional finger-vein recognition methods, complex image processing is required to remove noise and extract and enhance the features before the image classification can be performed in order to achieve high performance accuracy. In this regard, a significant advantage of the CNN over conventional approaches is its ability to simultaneously extract features, reduce data dimensionality, and classify in one network structure. In addition, the method requires only minimal image preprocessing since the CNN is robust to noise and small misalignments of the acquired images. In this paper, a reduced-complexity four-layer CNN with fused convolutional-subsampling architecture is proposed for finger-vein recognition. For network training, we have modified and applied the stochastic diagonal Levenberg{Marquardt algorithm, which results in a faster convergence time. The proposed CNN is tested on a finger-vein database developed in-house that contains 50 subjects with 10 samples from each finger. An identification rate of 100.00% is achieved, with an 80/20 percent ratio for separation of training and test samples, respectively. An additional number of subjects have also been tested, in which for 81 subjects an accuracy of 99.38% is achieved.
93 citations
TL;DR: In this article, an analysis of RC and RL electrical circuits described by a fractional difierential equation of Caputo type is provided. The Laplace transform of the fractional derivative is used.
Abstract: This paper provides an analysis of RC and RL electrical circuits described by a fractional difierential equation of Caputo type. The order considered is 0<γ ≤ 1. The Laplace transform of the fractional derivative is used. To keep the dimensionality of the physical quantities, R, C, L, and an auxiliary parameter σ are introduced, characterizing the existence of fractional components in the system. The relationship between γ and σ is reported. The response obtained from the fractional RC and RL circuits exhibits the characteristic behaviors of a cap-resistor, memcapacitor, and memristor, as well as charge-voltage for memcapacitive systems and current-voltage for memristive systems. The relationship between Ohm’s law and Faraday’s laws for the charge stored in a capacitor and induction is reported. Illustrative examples are presented.
59 citations
TL;DR: The proposed CNN architecture exhibits a much reduced design complexity when compared with other CNN solutions applied in pattern recognition, and is able to process and classify a 32 × 32 pixel face image in less than 0.27 ms, which corresponds to a very high throughput of over 3700 images per second.
Abstract: An approach using a convolutional neural network (CNN) is proposed for real-time gender classification based on facial images. The proposed CNN architecture exhibits a much reduced design complexity when compared with other CNN solutions applied in pattern recognition. The number of processing layers in the CNN is reduced to only four by fusing the convolutional and subsampling layers. Unlike in conventional CNNs, we replace the convolution operation with cross-correlation, hence reducing the computational load. The network is trained using a second-order backpropagation learning algorithm with annealed global learning rates. Performance evaluation of the proposed CNN solution is conducted on two publicly available face databases of SUMS and AT&T. We achieve classification accuracies of 98.75% and 99.38% on the SUMS and AT&T databases, respectively. The neural network is able to process and classify a 32 × 32 pixel face image in less than 0.27 ms, which corresponds to a very high throughput of over 3700 images per second. Training converges within less than 20 epochs. These results correspond to a superior classification performance, verifying that the proposed CNN is an effective real-time solution for gender recognition.
55 citations
TL;DR: In this paper, five probability distribution functions are employed to fit the wind speed data from four different geographical locations in the world in a preliminary analysis, and the suitability of PDFs to predict the wind power densities and annual energy production using manufacturers' power curve data at three of the selected sites is analyzed.
Abstract: In this paper, five probability distribution functions are employed to fit the wind speed data from four different geographical locations in the world in a preliminary analysis. These wind regimes are selected such that they represent wide ranges of mean wind speeds and present different shapes of wind speed histograms. The wind speed data used for modelling consist of 10-min average SCADA data from three US wind farms and hourly averages recorded at a weather station in Canada. Out of the five, three functions, namely Weibull, Rayleigh, and gamma, which provide a better fit to the data, are selected to carry out further analyses. This study investigates the ability of these functions to match different statistical descriptions of wind regimes. Parameter estimation is done by the method of moments, and models are evaluated by root mean square error and R square methods. The suitability of PDFs to predict the wind power densities and annual energy production using manufacturers’ power curve data at three of the selected sites is analysed. Power curves extracted from actual data of one wind farm using novel fourand five-parameter logistic approximations are also introduced here for energy analyses.
53 citations
TL;DR: This study proposes that the BBO algorithm has a high-quality solution and stable convergence characteristics, and thus it improves the transient response of the controlled system, and compares it with the ABC algorithm, particle swarm optimization algorithm, and differential evolution algorithm.
Abstract: A self-tuning method to determine the appropriate parameters of a proportional-integral-derivative controller for an automatic voltage regulator (A VR) system using a biogeography-bas ed optimization (BBO) algorithm is proposed in this study. The BBO algorithm was developed based on the theory of biogeography, which describes migration and its results. We propose that the BBO algorithm has a high-quality solution and stable convergence characteristics, and thus it improves the transient response of the controlled system. The performance of the BBO algorithm depends on the transient response, root locus, and Bode analysis. Robustness analysis is done in the A VR system, which is tuned by an articial bee colony (ABC) algorithm in order to identify its response to changes in the system parameters. We compare the BBO algorithm with the ABC algorithm, particle swarm optimization algorithm, and differential evolution algorithm. The results of this comparison show that the BBO algorithm has a better tuning capability than the other optimization algorithms.
48 citations
TL;DR: The KPSO algorithm is compared with traditional clustering techniques such as the low energy adaptive clustering hierarchy (LEACH) protocol and K-means clustering separately and can increase the network lifetime while minimizing the energy consumption.
Abstract: Energy saving in wireless sensor networks (WSNs) is a critical problem for diversity of applications. Data aggregation between sensor nodes is huge unless a suitable sensor data ow management is adopted. Clustering the sensor nodes is considered an effective solution to this problem. Each cluster should have a controller denoted as a cluster head (CH) and a number of nodes located within its supervision area. Clustering demonstrated an effective result in forming the network into a linked hierarchy. Thus, balancing the load distribution in WSNs to make efficient use of the available energy sources and reducing the traffic transmission can be achieved. In solving this problem we need to �nd the optimal distribution of sensors and CHs; thus, we can increase the network lifetime while minimizing the energy consumption. In this paper, we propose our initial idea on providing a hybrid clustering algorithm based on K-means clustering and particle swarm optimization (PSO); named KPSO to achieve efficient energy management of WSNs. Our KPSO algorithm is compared with traditional clustering techniques such as the low energy adaptive clustering hierarchy (LEACH) protocol and K-means clustering separately.
43 citations
TL;DR: The idea of the recursive estimation of KF is used to propose two recursive updating rules for the process and observation covariances respectively designed based on the covariance matching principles and the proposed adaptive Kalman filter AKF proved itself to have an improved performance over the conventional KF.
Abstract: Kalman filter (KF) is used extensively for state estimation. Among its requirements are the process and observation noise covariances which are unknown or partially known in real life applications. Biased initializations of the covariances result in performance degradation of KF or divergence. Therefore, an extensive research is carried on to improve its performance however, relying on a moving window, heavy computations, and the availability of the exact model are the fundamental problems in most of the proposed techniques. In this paper, we are using the idea of the recursive estimation of KF to propose two recursive updating rules for the process and observation covariances respectively designed based on the covariance matching principles. Each rule is a tuned scaled version of the previous covariance in addition to a tuned correction term derived based on the most recent available data. The proposed adaptive Kalman filter AKF avoided the aforementioned problems and proved itself to have an improved performance over the conventional KF. The results show that the AKF estimates are more accurate, have less noise and more stable against biased covariance initializations.
41 citations
TL;DR: This paper proposes a novel mechanism for discriminating DDoS and flash crowds based on the combination of the parameters reflecting their behavioral differences and shows that the proposed mechanism does effective detection with fewer false positives and false negatives.
Abstract: Distributed denial of service (DDoS) attacks are ever threatening to the developers and users of the Internet. DDoS attacks targeted at the application layer are especially difficult to be detected since they mimic the legitimate users' requests. The situation becomes more serious when they occur during flash events. A more sophisticated algorithm is required to detect such attacks during a flash crowd. A few existing works make use of flow similarity for differentiating flash crowds and DDoS, but flow characteristics alone cannot be used for effective detection. In this paper, we propose a novel mechanism for discriminating DDoS and flash crowds based on the combination of the parameters reflecting their behavioral differences. Flow similarity, client legitimacy, and web page requested are identified as the principal parameters and are used together for effective discrimination. The proposed mechanism is implemented on resilient proxies in order to protect the server from direct flooding and to improve the overall performance. The real datasets are used for simulation, and the results are presented to evaluate the performance of the proposed system. The results show that the proposed mechanism does effective detection with fewer false positives and false negatives.
39 citations
TL;DR: In this paper, an optimal tuning of the controller parameters of a proportional-int (PID) controller for an automatic voltage regulator (A VR) system using a heuristic gravitational search algorithm (GSA) based on mass interactions and Newton's law of gravity is presented.
Abstract: This paper presents optimal tuning of the controller parameters of a proportional-int (PID) controller for an automatic voltage regulator (A VR) system using a heuristic gravitational search algorithm (GSA) based on mass interactions and Newton's law of gravity. The determination of optimal controller parameters is considered an optimization problem in which different performance indexes and a performance criterion in the time domain have been used as objective functions to test the performance and effectiveness of the GSA. In the determining process of the parameters, the designed PID controller with the proposed approach is simulated under different conditions and the performance of the controller is compared with those reported in the literature. From the numerical simulation results it is clear that the GSA approach is successfully applied to reveal the performance and the feasibility of the proposed controller in the A VR system.
32 citations
TL;DR: Comparison of forecasting performances in terms of relative absolute error (RAE) and root relative squared error (RRSE) measurements shows that the SVR model gives a better performance than the multiple linear regression model.
Abstract: In this study multiple linear regression, multilayer perceptron (MLP) regression, and support vector regression (SVR) are used to make multivariate tourism forecasting for Turkey. This paper is a comparative study of data mining techniques based on multivariate regression modelling with monthly data points to forecast tourism demand; it focuses on Turkey. Both MLP and SVR methods are widely employed in the variety forecasting problems. Most of the previous research on tourism forecasting used univariate time series or a limited number of variables with mostly yearly or quarterly, and rarely monthly frequencies. However, the application of data mining techniques for multivariate forecasting in the context of tourism demand has not been widely explored. This paper differs from earlier research in two ways: 1) it proposes multivariate regression modelling with monthly data points to forecast tourism demand; and 2) it focuses on Turkey by using a dataset with the most recently accumulated (between January 1996 and Dec 2013) 67 time series with respect to Turkey and its top 26 major tourism clients. Comparison of forecasting performances in terms of relative absolute error (RAE) and root relative squared error (RRSE) measurements shows that the SVR model, with RAE = 12.34% and RRSE = 14.02%, gives a better performance. The results obtained in this study provide information for researchers interested in applying data mining techniques to tourism demand forecasting and help policy makers, government bodies, investors, and managers for their regularization, planning, and investments by way of accurate tourism demand forecasting.
31 citations
TL;DR: A multiterminal DC compact node consisting of voltage-source converters and DC-DC converters as a promising arrangement to feed electric vehicle (EV) charging stations with renewable power sources is described.
Abstract: This paper describes a multiterminal DC compact node consisting of voltage-source converters and DC-DC converters as a promising arrangement to feed electric vehicle (EV) charging stations with renewable power sources. This research analyzed the behavior of the node currents under DC faults, and more specically their natural response, which can produce challenging electrical protection requirements. An electrical protection system was thus designed based on the use of protective devices that included protective relays, solid state circuit breakers and hybrid circuit breakers. These protective devices monitor local readings to detect and isolate DC faults as quickly as possible. This highlighted the fact that there are devices, available on the market, that comply with the fast response requirement to prevent damage or destruction of the converters andlter capacitors of the node.
TL;DR: Considering geographic information and traffic density, an optimization overture for optimal siting and sizing of a rapid CS (RCS) is proposed in this paper, which aims to minimize the daily total cost (which includes the cost of substation energy loss, traveling cost of EVs to the CS, and investment, variable, and operational costs of the stations simultaneously) while maintaining system constraints.
Abstract: Recently, electric vehicles (EVs) have been seen as a felicitous option towards a less carbon-intensive road transport. The key issue in this system is recharging the EV batteries before they are exhausted. Thus, charging stations (CSs) should be carefully located to make sure EV users can access a CS within their driving range. Considering geographic information and traffic density, this paper proposes an optimization overture for optimal siting and sizing of a rapid CS (RCS). It aims to minimize the daily total cost (which includes the cost of substation energy loss, traveling cost of EVs to the CS, and investment, variable, and operational costs of the stations simultaneously) while maintaining system constraints. The binary gravitational search algorithm, genetic algorithm, and binary particle swarm optimization algorithm were employed to optimize the daily total cost by finding the best location and sizing of the RCS in a given metropolitan area in Malaysia. The results show that the proposed methods can find optimal locations and sizing of a RCS that can benefit EV users, CS developers, and the power grid.
TL;DR: A system is proposed to solve mobile robot navigation by opting for the most popular two RL algorithms, Sarsa and Q, and uses state and action sets, defined in a novel way, to increase performance.
Abstract: In recent decades, reinforcement learning (RL) has been widely used in different researchelds ranging from psychology to computer science. The unfeasibility of sampling all possibilities for continuous-state problems and the absence of an explicit teacher make RL algorithms preferable for supervised learning in the machine learning area, as the optimal control problem has become a popular subject of research. In this study, a system is proposed to solve mobile robot navigation by opting for the most popular two RL algorithms, Sarsa( ) and Q( ) . The proposed system,
TL;DR: In this paper, the authors used a TRNSYS model for estimating electricity yields from a fixed slope photovoltaic (PV) panel at a test site in Cape Town, South Africa.
Abstract: TRNSYS stands for transient system simulation software. This paper describes a procedure that was used to validate a TRNSYS model for estimating electricity yields from a fixed slope photovoltaic (PV) panel. The objective was to find how close to reality predicted energy yield for a specified panel can be, at a location near one of the weather stations listed in the software’s database. The software was used to predict daily total incident radiation on a horizontal plane and electrical energy yields from a 90 Wp panel when sloped at 34◦ facing north at a test site in Cape Town, South Africa. The panel and other system components were then installed and tested to give actual electrical energy yields. The site was 5 km from a TRNSYS listed weather station. A local weather station logging 10-min data of actual total incident radiation on a horizontal plane enabled comparison with the model’s estimate. Analysis of electrical energy yield gave statistical kappa values of 0.722 and 0.944 at actual to model acceptance ratio levels of 90% and 80%, respectively. Regression analysis of measured and model incident horizontal plane energy gave a coefficient of 0.782 across the year. It was thus concluded that within limits of meteorological phenomena behaviour, TRNSYS modelling reliably predicted energy yields from the PV panel installed in the neighbourhood of one of the software’s listed stations.
Journal Article•
TL;DR: The results indicate that the proposed method reduces the amount of training time and has a considerable success in removing noisy data from the training dataset, and can achieve a higher generalization performance in comparison with the other methods in large, real-world datasets.
Abstract: Classifying large and real-world datasets is a challenging problem in machine learning algorithms. Among the machine learning methods, the support vector machine (SVM) is a well-known approach with high generalization ability. Unfortunately, while the number of training data increases and the data contain noise, the performance of SVM significantly decreases. In this paper, a fast and de-noise two-stage method for training SVMs to deal with large, real-world datasets is proposed. In the first stage, data that contain noises or are suspected to be noisy are identified and eliminated from the genuine training dataset. The process of elimination and identification is based on the movement of the center of the convex hull data in the training dataset. The convex hull data are computed via the QHull algorithm. On the other hand, the well-known fuzzy clustering method (FCM) is applied to compress and reduce the size of the training dataset. Finally, the reduced and purified cluster centers are used for training the SVM. A set of experiments is conducted on the four benchmarking datasets of the UCI database. Moreover, the amount of training time and the generalization of the proposed approach are compared with FCM-SVM and normal SVM. The results indicate that the proposed method reduces the amount of training time and has a considerable success in removing noisy data from the training dataset. Therefore, the proposed method can achieve a higher generalization performance in comparison with the other methods in large, real-world datasets.
TL;DR: This paper presents a new feature set for the problem of recognizing pulse repetition interval (PRI) modulation patterns based upon the features extracted from the multiresolution decomposition of different types of PRI modulated sequences and shows that the proposed feature set is highly robust and separates jittered, stagger, and other modulation patterns very well.
Abstract: This paper presents a new feature set for the problem of recognizing pulse repetition interval (PRI) modulation patterns. The recognition is based upon the features extracted from the multiresolution decomposition of different types of PRI modulated sequences. Special emphasis is placed on the recognition of jittered and stagger type PRI sequences due to the fact that these types of PRI sequences appear predominantly in modern electronic warfare environments for some specic mission requirements and recognition of them is heavily based on histogram features. We test our method with a broad range of PRI modulation parameters. Simulation results show that the proposed feature set is highly robust and separates jittered, stagger, and other modulation patterns very well. Especially for the stagger type of PRI sequences, wavelet-based features outperform conventional histogram-based features. Advantages of the proposed feature set along with its robustness criteria are analyzed in detail.
Journal Article•
TL;DR: This research presents a probabilistic procedure for estimating the intensity of the response of the H2O signal in the E-modulus of the blood supply to the brain.
Abstract: Chunbiao LI1,2,3,∗, İhsan PEHLİVAN, Julien Clinton SPROTT School of Electronic & Information Engineering, Nanjing University of Information Science & Technology, Nanjing, P.R. China Department of Physics, University of Wisconsin-Madison, Madison, WI, USA Engineering Technology Research and Development Center of Jiangsu Circulation Modernization Sensor Network, Jiangsu Institute of Commerce, Nanjing, P.R. China Department of Electrical and Electronics Engineering, Faculty of Technology, Sakarya University, Esentepe Campus, Serdivan, Sakarya, Turkey
TL;DR: In this article, an artificial neural network (ANN), multiple linear regression (MLR), and autoregressive integrated moving average (ARIMA) models are employed to forecast and track furnace flame temperature selecting the most appropriate inputs that affect this process parameter.
Abstract: The blast furnace (BF) is the heart of the integrated iron and steel industry and used to produce melted iron as raw material for steel. The BF has very complicated process to be modeled as it depends on multivariable process inputs and disturbances. It is very important to minimize operational costs and reduce material and fuel consumption in order to optimize overall furnace efficiency and stability, and also to improve the lifetime of the furnace within this task. Therefore, if the actual flame temperature value is predicted and controlled properly, then the operators can maintain fuel distribution such as oxygen enrichment, blast moisture, cold blast temperature, cold blast flow, coke to ore ratio, and pulverized coal injection parameters in advance considering the thermal state changes accordingly. In this paper, artificial neural network (ANN), multiple linear regression (MLR), and autoregressive integrated moving average (ARIMA) models are employed to forecast and track furnace flame temperature selecting the most appropriate inputs that affect this process parameter. All data were collected from Erdemir Blast Furnace No. 2, located in Eregli, Turkey, during 3 months of operation and the computational results are satisfactory in terms of the selected performance criteria: regression coefficient and root mean squared error. When the proposed model outputs are considered for the comparison, it is seen that the ANN models show better performance than the MLR and ARIMA models.
TL;DR: The experimental results demonstrated that ASM could be even used both in the estimation of HRV signals and to detect the peaks from raw and noisy PPG signals without a preprocessing method.
Abstract: The purpose of this paper is twofold. The first purpose is to detect M-peaks from raw photoplethysmography (PPG) signals with no preprocessing method applied to the signals. The second purpose is to estimate heart rate variability (HRV) by finding the peaks in the PPG signal. HRV is a measure of the fluctuation of the time interval between heartbeats and is calculated based on time series between strokes derived from electrocardiogram (ECG), arterial pressure (AP), or PPG signals, separately. PPG is a method widely used to measure blood volume of tissue on the basis of blood volume change in every heartbeat. In the estimation of the HRV signal from the PPG signal, HRV is calculated by measuring the time intervals between the peak values in the PPG signal. In the present paper, a novel peak detection algorithm was developed for PPG signals. Finding peak values correctly from PPG signals, the HRV signal can be estimated. This peak detection algorithm has been called an adaptive segmentation method (ASM). In this method, the PPG signals are first separated into segments with sample sizes and then the peak points in these signals are detected by comparing with maximum points in these segments. To evaluate the estimated pulse rate and HRV signals from PPG, Poincare plots and time domain features including minimum, maximum, mean, mode, standard deviation, variance, skewness, and kurtosis values were used. Our experimental results demonstrated that ASM could be even used both in the estimation of HRV signals and to detect the peaks from raw and noisy PPG signals without a pre-processing method.
TL;DR: It is shown that the overall channel exploitation is increased by utilizing the spectrum holes without interfering with the PUs' transmissions, and new equations for the throughput of the proposed CR network are obtained.
Abstract: In this paper, a slotted ALOHA-based cognitive radio (CR) network is proposed and the throughput performance of the proposed CR network model under Rayleigh fading channels is examined. Our CR network contains two special groups of users, primary users (PUs) and CR users (CRUs), and they are considered to be sharing a time- slotted-based common communication channel. While PUs can access the channel at any time owing to their legal right, CRUs can only access the channel when it is not occupied by the PUs. In the network model developed, PUs access the channel utilizing time division multiple access as a medium access control technique, and CRUs can access the channel by exploiting slotted ALOHA as a random access scheme when the channel is idle. In the proposed network model additive white Gaussian noise and Rayleigh channels is considered for comparison reasons. Taking capture effect into account in Rayleigh fading channels, we have obtained new equations for the throughput of the proposed CR network. We have also developed, modeled, and simulated a sample networking scenario by using MATLAB with the aim of validating the analytical throughput results. Simulation results of the proposed network model precisely match with the analytical results obtained under different network load conditions. Furthermore, it is shown that the overall channel exploitation is increased by utilizing the spectrum holes without interfering with the PUs' transmissions.
TL;DR: Two CPW-fed compact printed slot antennas for UWB applications are proposed in this paper and the measured impedance bandwidths (S 11 < {10 dB) achieved by the antennas are very compact in size and designed on low-cost FR4 substrate.
Abstract: Two CPW-fed compact printed slot antennas for UWB applications are proposed in this paper. In both antennas, the slot in the ground plane is of octagonal shape, while the patch is rectangular in one of the antennas and hexagonal in the other. Thin metallic stubs of different lengths are added to the ground plane and protrude into the slot. These stubs help to improve the impedance matching over a broader bandwidth. The antennas are very compact in size and designed on low-cost FR4 substrate. The measured impedance bandwidths (S 11 < {10 dB) achieved by the
TL;DR: A knowledge-based genetic algorithm (KBGA), which combines the image characteristics and knowledge about image compressibility with CGA operators such as initialization, selection, crossover, and mutation for searching for the optimal quantization table, is proposed.
Abstract: JPEG has played an important role in the image compressioneld for the past two decades. Quantization tables in the JPEG scheme is a key factor that is responsible for compression/qual trade-off. Finding the optimal quantization table is an open research problem. Studies recommend the genetic algorithm to generate the optimal solution. Recent reports revealed optimal quantization table generation based on a classical genetic algorithm (CGA). Although the CGA produces better results, it shows inefficiency in terms of convergence speed and productivity of feasible solutions. This paper proposes a knowledge-based genetic algorithm (KBGA), which combines the image characteristics and knowledge about image compressibility with CGA operators such as initialization, selection, crossover, and mutation for searching for the optimal quantization table. The experimental results show that the optimal quantization table generated using the proposed KBGA outperforms the default JPEG quantization table in terms of mean square error (MSE) and peak signal-to-noise ratio (PSNR) for target bits per pixel. The KBGA was also tested on a variety of images in three different bits values per pixel to show its strength. The proposed KBGA produces an average PSNR gain of 3.3% and average MSE gain of 20.6% over the default JPEG quantization table. The performance measures such as average untness value, likelihood of evolution leap, and likelihood of optimality are used to validate the efficacy of the proposed KBGA. The novelty of the KBGA lies in the number of generations used to attain an optimal solution as compared to the CGA. The validation results show that this proposed KBGA guarantees feasible solutions with better quality at faster convergence rates.
TL;DR: In this work, two new equations are proposed to find the ideality factor and shunt resistance and to obtain accurate MPP of the PV model under varying environmental conditions.
Abstract: The objective of this paper is to estimate the values of five parameters (A , Rse , Rsh , ILG , and Isat) of a PV module more accurately and to extract the maximum power point (MPP) accurately under varying environmental conditions. Suitable new equations are proposed, by which the values of the series and shunt resistances are initialized in order to obtain good convergence speed in the Gauss–Seidel method. In this work, two new equations are proposed to find the ideality factor and shunt resistance and to obtain accurate MPP of the PV model under varying environmental conditions. The proposed PV model is validated at standard test conditions and under varying environmental conditions. Current–voltage and power–voltage characteristics of different PV modules are simulated using MATLAB. Accuracy of the proposed model is validated by comparing with the results of an adaptive neuro-fuzzy inference system and experimental data taken under varying environmental conditions.
TL;DR: Numerical results indicate that the performance of the proposed receiver is very close to the Rayleigh theoretical bound, and it is shown that the neural network-based receiver may be used for channel estimation and equalization over Rayleigh channels.
Abstract: In communication systems, the channel noise is usually assumed to be white and Gaussian distributed. Therefore, an optimum receiver structure designed for the additive white Gaussian noise (AWGN) channel is employed in applications. However, in wireless communication systems, noise is often caused by strong interferences. Moreover, there are other effects such as phase offset that degrade the performance of the receiver. Designing the optimum receiver for different channel models is difficult and not reasonable because channel model and channel statistics are not known at the receiver. In this paper, we propose a neural network-based approach to demodulate the transmitted signal over unknown channels. Naturally, the collection of the training data, design and training of the neural network, and finally reconfiguration of the system according to the designed neural network are implemented on software-defined digital signal processing facilities. In particular, we show that the proposed receiver is capable of jointly canceling the strong interferences and phase offset. Simulation results in various signal environments are presented to illustrate the performance of the proposed system. It is shown that the proposed approach has the same performance as the correlation demodulator structure for AWGN channels, while it has a clear advantage for unknown channel models. Moreover, it is shown that the neural network-based receiver may be used for channel estimation and equalization over Rayleigh channels. Numerical results indicate that the performance of the proposed receiver is very close to the Rayleigh theoretical bound.
TL;DR: The presented multiplier is expected to be useful in the design of various analog signal processing applications such as modulators and frequency doublers, as illustrated in this paper.
Abstract: In this paper, a new CMOS four-quadrant analog multiplier circuit is proposed, based on a pair of dual- translinear loops. The signicant features of the circuit are its high accuracy and high linearity as well as its body effect-free operation, owing to the fact that the circuit relies on a new dual-translinear topology. In addition, harmonic distortions are precisely discussed due to their conceivable mismatches, including transconductance and threshold voltage of the transistors. HSPICE postlayout simulation results are presented to verify the validity of the theoretical analysis, where under a supply voltage of 2.8 V, the bandwidth of the proposed multiplier is 137 MHz, and the corresponding maximum linearity error remains as low as 1.12%. Moreover, the power dissipation of the proposed circuit is found to be 521 W. The presented multiplier is expected to be useful in the design of various analog signal processing applications such as modulators and frequency doublers, as illustrated in this paper.
Journal Article•
TL;DR: In this paper, the authors investigated the problem of high torque ripple in switched reluctance motor (SRM) drives and proposed a method for below the base speed operation of SRM that determines both the turnoff and the turn-on angles for reducing motor torque ripple.
Abstract: This paper investigates the problem of high torque ripple in switched reluctance motor (SRM) drives. A method is proposed for below the base speed operation of SRM that determines both the turn-off and the turn-on angles for reducing motor torque ripple. Determination of the turn-off angle is an offline process performed through solving a multiobjective optimization function consisting of two criteria: torque ripple and copper loss. Turn-on angle adjustment, however, is an online process based on the intersection approach of consecutive phase currents, particularly proposed in this paper. Simulation and experimental results are presented to validate the reduction in torque ripple gained from the proposed angles control scheme.
TL;DR: A CSP-based moving window technique to obtain the most distinguishable CSP features and increase the classifier performance by finding the best time segment of electroencephalogram trials is developed.
Abstract: Feature extraction is one of the most crucial stages in the field of brain computer interface (BCI). Because of its ability to directly influence the performance of BCI systems, recent studies have generally investigated how to modify existing methods or develop novel techniques. One of the most successful and well-known methods in BCI applications is the common spatial pattern (CSP). In existing CSP-based methods, the spatial filters were extracted either by using the whole data trial or by dividing the trials into a number of overlapping/nonoverlapping time segments. In this paper, we developed a CSP-based moving window technique to obtain the most distinguishable CSP features and increase the classifier performance by finding the best time segment of electroencephalogram trials. The extracted features were tested by using support vector machines (SVMs). The performance of the classifier was measured in terms of classification accuracy and kappa coefficient (κ) . The proposed method was successfully applied to the two-dimensional cursor movement imagery data sets, which were acquired from three healthy human subjects in two sessions on different days. The experiments proved that instead of using the whole data length of EEG trials, extracting CSP features from the best time segment provides higher classification accuracy and κ rates.
TL;DR: A kind of half-bridge-based railway power quality compensator system (HBRPQC) that can compensate negative sequence currents, harmonics, and reactive power simultaneously simultaneously is discussed.
Abstract: The development of electrical railway systems leads to critical power quality problems in the power grid. This paper discusses a kind of half-bridge-based railway power quality compensator system (HBRPQC) that can compensate negative sequence currents, harmonics, and reactive power simultaneously. In order to keep the HBRPQC performance efficient for the different kinds of transformers used in traction power supply substations, a new multifunctional control strategy that performs better than previous methods is proposed. Due to the fast dynamicity of traction loads, a recessive self-tuning PI controller based on fuzzy logic is adopted in the current control system. The output control variables are integrated with carrier-based pulse width modulation techniques to generate the pulse signals of HBRPQC switches. The performance of the proposed control strategy is verified for V/V, Yd11, and Scott transformers by simulation and the results prove the effectiveness of the strategy.
TL;DR: A new method based on optimization methods to simultaneously generate appropriate fuzzy membership and solve the model selection problem for the SVM family in linear/nonlinear and separable/nonseparable classification problems is proposed.
Abstract: The support vector machine (SVM) is a powerful tool for classification problems. Unfortunately, the training phase of the SVM is highly sensitive to noises in the training set. Noises are inevitable in real-world applications. To overcome this problem, the SVM was extended to a fuzzy SVM by assigning an appropriate fuzzy membership to each data point. However, suitable choice of fuzzy memberships and an accurate model selection raise fundamental issues. In this paper, we propose a new method based on optimization methods to simultaneously generate appropriate fuzzy membership and solve the model selection problem for the SVM family in linear/nonlinear and separable/nonseparable classification problems. Both the SVM and least square SVM are included in the study. The fuzzy memberships are built based on dynamic class centers. The firefly algorithm (FA), a recently developed nature-inspired optimization algorithm, provides variation in the position of class centers by changing their attributes’ values. Hence, adjusting the place of the class center can properly generate accurate fuzzy memberships to cope with both attribute and class noises. Furthermore, through the process of generating fuzzy memberships, the FA can choose the best parameters for the SVM family. A set of experiments is conducted on nine benchmarking data sets of the UCI data base. The experimental results show the effectiveness of the proposed method in comparison to the seven well-known methods of the SVM literature.
TL;DR: A new unsupervised framework, UHAD, which uses a two-step strategy to cluster the log events and then uses altering threshold to reduce the volume of events for analysis, which detected the majority of anomalies by relating the events from heterogeneous logs.
Abstract: Log analysis is a method to identify intrusions at the host or network level by scrutinizing the log events recorded by the operating systems, applications, and devices. Most work contemplates a single type of log for analysis, leading to an unclear picture of the situation and difficulty in deciding the existence of an intrusion. Moreover, most existing detection methods are knowledge-depend i.e. using either the characteristics of an anomaly or the baseline of normal traffic behavior, which limits the detection process to only anomalies based on the acquired knowledge. To discover a wide range of anomalies by scrutinizing various logs, this paper presents a new unsupervised framework, UHAD, which uses a two-step strategy to cluster the log events and then uses altering threshold to reduce the volume of events for analysis. The events from heterogeneous logs are assembled together into a common format and are analyzed based on their features to identify anomalies. Clustering accuracy of K-means, expectation-maxi and farthest �rst were compared and the impact of clustering was captured in all the subsequent phases. Even though log events pass through several phases in UHAD before being concluded as anomalous, experiments have shown that the selection of the clustering algorithm and theltering threshold signicantly inuences the decision. The framework detected the majority of anomalies by relating the events from heterogeneous logs. Specically, the usage of K-means and expectation- maximization supported the framework to detect an average of 87.26% and 85.24% anomalous events respectively with various subsets.