Search

The paper is focused on the determination of mechanical fracture parameters of alkali-activated concrete (AAC) specimens of nominal dimensions 80 × 80 × 240 mm with central notches of the following types: straight-through notch, chevron notch with 90° notch angle and chevron notch with 120° notch angle. In accordance with the results of the experimental research, a numerical simulation of the fracture test of the specimens tested in three-point bending was carried out. The 3D FEM computational model is based on the configuration of the laboratory test and was developed in ATENA-GiD software. Artificial neural network (ANN)-based inverse analysis was used to identify the mechanical fracture parameters of studied composite. A neural network ensemble was created, with one robust network for each type of notch, covering the entire domain of identified parameters. The main goal of this paper is to present the results of the identification of the mechanical fracture parameters of the AAC and their comparison with the values obtained by direct evaluation of the experiments. Another objective is to investigate the influence of the notch type on the mechanical fracture parameters of the AAC. [ABSTRACT FROM AUTHOR]

The rise of cyber-attacks such as Distributed Denial of Service (DDoS) and SQL injection has become a significant concern for organizations and individuals who rely on the internet. Traditional detection methods have become increasingly ineffective in addressing these challenges, necessitating the development of new and innovative solutions. This paper proposes using Convolutional Neural Networks (CNNs) to detect DDoS and SQL injection attacks. Our paper proposes a Convolutional Neural Network model with a multi-layer neural and relu activation function optimizer that reaches a higher degree of precision than previous Deep Learning models. In this study, we tested the relatively new dataset CIC-IDS-2018, which contains different types of attacks. With this dataset, our model achieves an unprecedented accuracy of>96%, minimizing computational time. [ABSTRACT FROM AUTHOR]

This paper describes an Artificial Neural Network (ANN) based speed estimation of BLDC drive using the three-phase current of the motor. The star point of three parallel resistors is used as the artificial star point of the motor. The comparators are connected to motor terminals and star point of resistors. This circuit determines the zero crossing of the back-EMF of the motor phases. Using zero crossing detection of the back-EMF with the estimated speed from the ANN, precise control of the BLDC motor is achieved, without the need for a sensor. The ANN is an offline trained network, trained using data obtained from simulations. The network is validated against different Inputs. The network is precise but at the same time simple enough to be implemented on a low-cost processor. The results are validated with simulations done in MATLAB using the trained neural network The findings and applications of neural networks to control the speed of PMSM are very well presented in the referenced papers, but the same for BLDC motors, especially those used in electric vehicles, have not been explored yet. The presented paper attempts to address the challenges and advantages of adopting this sensor-free approach for speed control in EVs. [ABSTRACT FROM AUTHOR]

Road traffic crashes result in a significant number of fatalities and economic losses worldwide. Rear-end collisions, caused by insufficient distance between vehicles, are a major contributor to these accidents. To address this issue, the use of Dedicated Short-Range Communication (DSRC) technology has been developed prominence in last decade. However, uncertainties associated with DSRC-based systems, such as GPS positioning errors and transmission delays, pose challenges to the development of robust collision warning models. In this paper, a design of three DSRC-based rear-end collision warning models to mitigate the impact of uncertainties. The models include the vehicle kinetic model (VK), the maximum compensation vehicle kinetic model (MC VK), and the artificial neural network model (NN). The VK model shows both a high possibility of false warning and missing. The MC VK model shows a 0% possibility in missing warning, which can be determined as the early warning boundary. The NN model shows low probability in both false and missing warning, and 80% correct warning makes it a good candidate for emergency warning boundary. [ABSTRACT FROM AUTHOR]

Utilizing bio-oils in liquid pumping unites are considered as a one of the most important parts in the water filtration and pumping stations especially in the area of automatic control. In this paper using bio-oils are being investigated which have high degradability and it is environmentally friendly. The thermal performance features of the liquid pumping unit lubricated with bio sunflower seed vegetable oil was investigated and compared with the performance features with mineral oil (MO10W-30). The study includes, oil temperature, density, power and tourq under different rational speed (1200-1500 rpm) for the pump. From experimental investigations carried out in this study, it can be concluded that biodegradable vegetable oils are adequate alternative to the mineral oil as a renewable bio cooling and lubricate. An automatic pumping unit was implemented (three tanks one as a source and two auxiliary) based on two types of controllers. These two are: PID controller and artificial neural networks. The results in neural back propagation that more accurate than PID controller. The over shoot is improving in artificial network is zero where compare with PID result 0.8 m3/s flow rate [ABSTRACT FROM AUTHOR]

An adaptive artificial neural network model of a fuzzy neural inference system is evaluated in this paper to predict the performance of a nanofluidic evacuated tube solar collector. Variable intake temperatures and weight percentages of the aluminum oxide nanofluid (1 percent, 3 percent, 5 percent), as well as different flow rates (3, 9, 15) l/min were investigated. In this study, three ANFIS inputs were used: inlet temperature, solar radiation rate, and ambient temperature. Where one output is obtained from the artificial neural network, which is the values of the angles of the reflective electronic curtain made of acrylic mirrors, which enhance the performance of the solar collector as well as control the temperature of the solar collector when the curtain is closed. We can note that the practical results are consistent with the prediction results by predicting the performance of the solar collector produced by ANFIS. [ABSTRACT FROM AUTHOR]

In this paper, the author analyzes the use of a trained artificial neural network to precisely evaluate hybrid rocket engine performance in chemical equilibrium condition, avoiding the use of an high fidelity chemical model. A three-stage small satellite launcher is considered as test case, and the optimization of the engine design and the corresponding ascent trajectory is performed by means of an in-house coupled optimization procedure. The net evaluates specific heat ratio, characteristic velocity and thrust coefficient for given mixture ratio, combustion chamber pressure and nozzle expansion area ratio. The results show that the use of the artificial neural network can improve model accuracy at an affordable computational cost. [ABSTRACT FROM AUTHOR]

The paper introduces an original permutation approach to assessment of the predictive ability of models based on artificial neural networks. Three models based on artificial neural networks (a multilayer perceptron, a network with the radial basis function, and a neural network with the generalized regression) were implemented to illustrate the permutation approach. Data on the spatial distribution of copper in the upper soil layer of Novy Urengoy (Yamalo-Nenets Autonomous Okrug, Russia) were used for modeling. To evaluate the performance of the models, three different methods were used: error indices estimation, a graphical approach (the Taylor diagram), and a randomization estimation of the probability of obtaining the divergence between the observed and predicted series under assumption that both these datasets are taken from the same population. In the permutation approach, two statistics were used: the difference in means and the correlation coefficient. The permutation approach proved to be productive, as it allowed assessing the significance of the divergence between the observed and predicted datasets. [ABSTRACT FROM AUTHOR]

Waste rubber tires have been used in building materials to support the environment and green construction. There is a growing demand for rubberized materials as they are cost-effective and useful from a sustainable standpoint. There are several properties of the rubber tire that could be applied usefully such as low density, and waterproofing properties. Rubberized concrete has composed of waste rubber as natural aggregate and is an alternative solution to the use of tire rubber particles in the production of concrete. It has been proven that the addition of waste rubber tires to concrete starts to low strength and this restriction its application in structural elements but benefits to enhance the ductility, impact resistance, thermal conductivity, and acoustic properties. This paper presents a review of the recent studies on the application and development of artificial neural networks (ANNs) to predict the compressive strength of rubberized concrete. From this review, predicting the compressive strength of rubberized concrete by ANNs is generally more accurate, and their development is inexpensive and time-saving. In this review, the advantages, and limitations of ANNs to predict the compressive strength of rubberized concrete, as well as some new research trends have been addressed. [ABSTRACT FROM AUTHOR]

An original infection has spread over the global village in form of pandemic and the regular lifestyle of human beings got stuck with misfortunes. As it has been declared by the global organization of health sector namely WHO, this Covid got initiated from southern city of China during the last weeks of the year 2019. Based on the multiple observations, it was declared as most infectious as well contagious spread through wind or by coming in close contact with a tainted individual. Various methods have been implemented to cease off the fast spreading infectious disease. For example, keeping a social separation, that is to say, keeping a legitimate actual separation among individuals and reducing close contact with one another, and wearing a facial covering to stay away from the beads that are passed by means of wind air. Consequently, this paper put forth the observation made to explore the initiatives for maintaining physical distance in the public domain and to identify whether a Face Mask is worn or not. The architectural framework provides the exact location and position of an object along with an acknowledgment about facial coverage as per the WHO norms under the moving images of the persons who are into public domain. Our suggested model implements the existing procedures such as dual shot face detector, You look only once and finally ResNet algorithms. Irrespective of the origin location of an individual and their interaction among themselves taking into consideration the space between them and the facial expressions in a continuous form were recognized. A general outcomes dashboard has displayed the result composing the measures with regard to the abuses of an individual person and breaking the protocol of maintaining separate distance measures. Subsequent to carrying out and conveying the procedures of process, the above said exploration model accomplished an accuracy of maximum performance. Subsequently, this exploration project had come up with demonstrated realities that maintaining physical distance along with the usage of facial mask decreases the escalation of the disease, which lead to consequently constructs a procedures to assist with identifying the mentioned actions. [ABSTRACT FROM AUTHOR]

Accident prediction and Black Spot identification have a significant role in improving traffic safety and urban traffic management. This paper deals with developing the model to identify black spot locations using the Classical approach that relies on historical accident occurrence and the Computational Intelligence approach. A hybrid approach with a combination of the best fit Artificial Neural Network (ANN) model and classical Accident frequency method results in 97.72% accurate prediction of the black spots. Combining the best fit ANN model and the Accident severity method results in 93.18% accuracy in predicting black spots. It concluded that combining Computational Intelligence and a Classical approach ensures better accuracy in identifying black spots. The uniqueness of the applied model in this research also includes combining Classical methods and Fuzzy C-means clustering to identify the road segments recognized as black spots. It also identifies the road segments categorized in the medium possibility range expected to be promoted to highly vulnerable groups. [ABSTRACT FROM AUTHOR]

Due to the high labour costs in industrialised countries and the increasing demands on the quality of plastic moulded parts produced by the injection moulding process, companies are challenged to increase their production efficiency. Data-based approaches provide an opportunity for systematic and objective process set-up and control. Especially the process set-up can be a tedious, iterative and complex task which can consume a significant amount of time in the preparation of mass production. Artificial neural networks (ANN) can be used by evolutionary algorithms to search for suitable machine setting parameters for the desired part and process quality features. This often results in locally optimal solutions, which do not guarantee the best possible machine settings. Invertible neural networks (INN) could be a solution as with the help of these models, globally optimal solutions may be identified for definitive optimisation as the entire solution space for given quality parameters can be mapped in short time. In our approach, data for the implementation of the INN is collected in practical tests. The resulting weight and several geometric dimensions of the parts are determined as quality parameters. Based on these data, ANN are trained and utilized for the generation of synthetic data for the training of the INN. The creation of synthetic data is necessary because INN have millions of trainable parameters and therefore degrees of freedom which need to be determined using a bigger amount of training data in comparison to traditional feedforward ANN. In this paper, two different databases with a size of 10,000 samples each are being investigated. The two databases differ in the distributions of holding pressure and cooling times in order to investigate the influenceability of the prediction quality of the INN by training data. The sample settings for the injection velocity, packing pressure, mould temperature and melt temperature are uniformly distributed. Two different types of conventional and conditional invertible neural networks are implemented based on literature recommendations. Measured on validation data, the model quality (R² value) for the prediction of the part weight is 0.995 and 0.975 for the part dimensions. The machine setting parameters proposed by the INN can be analysed in regard to their impact. The holding pressure time is identified as the machine setting with the greatest influence on both part weight and part dimension. The predictions of the INN are validated in practical experiments with an R² value of 0.997 for the part weight. The weight of the parts produced with the proposed machine settings deviates on average by 0.24 % from the target. [ABSTRACT FROM AUTHOR]

The reasons for pollution and the limitations of fossil fuels have driven the use of electric vehicles. Electric vehicles are synonymous with electric batteries that require periodic charging. There has been an increase in research on battery charging for electric vehicles. There is also the use of renewable energy resources such as photovoltaics for battery charging stations. This paper focuses on prototyping battery charging stations from solar photovoltaic energy using ANN-based MPPT controllers within the framework of optimizing solar energy harvesting and comparing them with conventional MPPT methods. Test results showed that the prototype of the proposed charging system took 33.3 minutes to charge a 7Ah battery. Compared to conventional MPPT methods, the charging time of MPTT methods with Neural Network is 6.01% faster. [ABSTRACT FROM AUTHOR]

Vertical farming (VF) allows for more plants to be cultivated in a particular land area, maximizing land-use efficiency. VF employs multileveled shelves enabling crops to be planted at each level. This paper aims to develop an optimization model that could generate a required setup of a vertical shelf that would meet the sunlight requirements of strawberries. Firstly, the simulation model was developed using Rhino (Rhinoceros) software with Grasshopper plugin, which can predict the sunlight availability at each VF shelf level. The model is validated by the experiment in Chiang Mai province, Thailand, at an elevation higher than 310 m above mean sea level. The experiment set up of the VF shelf with three levels was constructed, and the solar radiation values were recorded and validated against the radiation predicted by the model. The validated model is used to simulate the sunlight radiation on each level of the shelf for different shelf designs with varying vertical spacings between each level. These sets of predicted radiation values and vertical spacings are used to train the ANN (Artificial Neural Network) model. The ANN simulation model was developed using MATLAB software with Simulink and Neural Net Fitting App. The trained ANN model has an MSE (Mean Squared Error) of 0.00019621 and an R (the correlation coefficient) value of 0.9992. As a result, the optimization model could successfully give out the vertical spacing between each level of the VF shelf with an input of desired sunlight availability on each level. For the PPFD (Photosynthetic Photon Flux Density) of 600 µmol/m2/s on both levels, the spacing should be 51.34 cm from the top to the middle (y1) and 48.69 cm from the middle to the bottom (y2). To receive 600 and 500 µmol/m2/s on the middle and bottom level respectively, the levels should be placed 66.77 cm (y1) and 30.89 cm (y2) apart. [ABSTRACT FROM AUTHOR]

This paper proposes an ANN based adaptive Auto Reclosing AR which integrated of an operation of distance measuring relay in HV Over Head Transmission Line OHTL. This work is brought to enhance stability synchronism, and reliability during the transient faults of OHTL. Thus, it leads to increase the continuity of the electric power system, especially of hydropower plants are connected to utility grid. This work are used Matlab\Simulink environment to design and simulate a single-phase auto-reclosing, which depends on the Artificial Neural Networks ANN integrated with a distance relay DR. In addition, the simulation model consists of Boolean algebra circuits, and timers are utilised to build Auto-reclosure of a circuit breaker in each phase. The ANN is here trained to recognize angle values of actual impedances (in the positive sequence only) in which it can classify which phase exactly is line to ground fault. The simulation results show the ability of this combination to distinguish among various fault types and giving multi pulses-shot of tripping and reclosing for the specific circuit breaker with a successful mechanism. In this paper, a realistic system with real parameters was represented by the important transmission line (Mosul Dam-East Mosul), which currently is only the main and alone line that connects the northern Iraq network with a national grid. [ABSTRACT FROM AUTHOR]

There has been a lot of research in recent years into employing optimization approaches to improve artificial intelligence (AI). In this research review paper, we have compared and contrast some of the most usual optimization algorithms, such as Backtracking searching method (BSA), the genetic algorithm (GA), particle swarm optimization (PSO), an artificial bee colony (ABC), and the genetic algorithm, which are all artificial neural networks (ANNs)-based algorithms. The number of recently developed optimization techniques, such as the lightning search algorithm (LSA) and the whale optimization algorithm (WOA) are also compared. All the techniques are categorized based on randomly generated populations. To produce the best possible results, the processing parameters are set within a certain range according to their knowledge. This review paper emphasis on applying optimization techniques to improve the accuracy of simply adjusting the parameters of the neural network. This review paper also presents some results for enhancing neural network performance using various optimization techniques like PSO, GA, and ABC optimization methods to get optimal processing parameters, such as the number of hidden layers, neurons and learning rate etc. The findings of this research review paper will aid in the improvement in the quality of plastic injection molded parts. [ABSTRACT FROM AUTHOR]

Artificial intelligence is the simulation of human intelligence and seeks to mimic human knowledge acquisition by using planning, reasoning, and research. Sometimes AI faces challenges due to a large set of rules, and growing demand for learning and searches optimization. Computational intelligence is a form of computing modeled on the methods by which humans learn. Artificial neural networks system learns by being given different image examples, and contain many layers of learning, some of them more complex than others. This paper presents a comprehensive survey on the various artificial neural network open source software for implementing computational intelligence as they seem to be effective in simulation, prediction, building, and studying artificial neural networks. This paper helps the researchers to find which software they should prefer by knowing their requirements. [ABSTRACT FROM AUTHOR]

The constant change in the components and variables of the economic environment is one of the most important phenomena that characterize this environment at present, and these changes have made the goal of survival in the market a priority due to the prevalence of financial distress in companies. The main purpose of this paper is to test the impact and effectiveness of the model of artificial neural networks and Naive Bayes technology and to see how accurate they are and their ability to classify companies. In this paper, the information of the companies listed on the Iraq stock exchange for 2017, which represents 36 companies with high financial distress, 20 companies with medium financial distress, in addition to 43 non-distressed companies for a group of 99 companies. According to the experimental results, the model of artificial neural networks gives the highest rating accuracy of 96.97% and an error rate compared to Naive Bayes, which gave a rating accuracy of 74.74%. [ABSTRACT FROM AUTHOR]

Associative Hopfield Memory is a very useful form of Artificial Neural Networks. This paper outlinesan optimization relaxation approach for signature verification based on the Hopfield neural network (HNN). The standard sample signature of the customer is cross matched with the one supplied on the check. The difference percentage is obtained by calculating the different pixels in both the images. The network topology is built so that each pixel in the difference image is a neuron in the network. Each neuron is categorized by its status, which in turn signifies that if the particular pixel is changed. The network converges to unwavering condition based on theenergy function which is derived in experiments. The Hopfield‟s model allows each node to take on two binary state values (changed/unchanged) for each pixel. The performance of the proposed technique is evaluated by applying it in various binary and gray scale images. This paper contributes to finding an automated scheme for verification of authentic signature on bank check. The derived energy function allows a trade-off between the influence of its neighborhood and its own criterion. This device is able to recall as well as complete partially specified inputs. The network is trained via a storage prescription that forces stable states to correspond to (local)minima of a network "energy" function. [ABSTRACT FROM AUTHOR]

Multiresolution analysis has wide-ranging applications in signal processing such as fault detection and isolation in sensors. Wavelet transforms can be used for feature extraction and compressing the signals without losing their characteristics. Coefficients of wavelet transform of a signal can be represented in the form of scalogram images. Neural networks helped researchers in replacing human interventions which can also be trained for specific tasks. Researchers have also demonstrated the high level of accuracy and usefulness of Deep Neural Networks(DNNs) in image classification. DNNs pre-trained with millions of images can be used for feature extraction or can be modified and fine-tuned for specific tasks. Scalograms contain the fault signatures of the transformed signal, which can then be classified with the help of a Deep Neural Network(DNN). Though DNNs are efficient in image classification it requires a lot of computing resources for training. To overcome this disadvantage we used transfer learning i.e. modifying and fine-tuning a pre-trained DNN. In this paper, we proposed a model for fault detection and isolation in sensors using transfer learning of DNNs and scalograms. A publicly available temperature sensor dataset was used for fault injection and classification purposes. The proposed model successfully classified sensor stuck faults with 95% accuracy by using scalograms as input to the fine-tuned DNN. [ABSTRACT FROM AUTHOR]

This research employs a novel approach to headcounting by comparing the performance of the ANN algorithm with that of the SVM algorithm. To determine how many individuals are moving under an overhead monitoring system, researchers have developed a revolutionary paper counting approach based on head recognition and tracking. Substances and Techniques: Twenty samples in all were taken from an electronic resource. We maintained a significance level of 0.05, a confidence interval of 95% (95% CI), an enrollment ratio of 1, and a significance level of 80% (G power). The precision value is calculated using the SPSS software. Result: The results of two separate t-tests show that the ANN algorithm outperformed the SVM method, although the difference is not statistically significant (p=0.721; p>0.05). In summary, it seems that the ANN algorithm outperforms the SVM method on the analysed datasets. [ABSTRACT FROM AUTHOR]

Using Regression Neural Networks and Probabilistic Neural Networks, implement a comparative overall performance evaluation of estimating food volume and monitoring meal intake for hospitalised affected persons using neural networks. This evaluation will be done using neural networks. Both the Materials and the Methods: Regression neural network and probabilistic neural network, with the same sample size (N=10) for each of the two groups, for the purpose of determining the accuracy of intake food estimation. Results show that the accuracy rate of the Regression neural network is 91.5 percent based on the obtained result, which is higher than the accuracy rate of the probabilistic neural network, which is 89.3 percent. According to the results of an independent sample T test, the statistical significance difference that was observed between two groups was P=0.00 (P0.05), 2-tailed. It has been determined through the research presented in this paper that the Probabilistic Neural Network is less accurate than the Regression Neural Network. [ABSTRACT FROM AUTHOR]

Physical disabilities cause significant distress in people's daily routine, especially for leg amputees as their movement is restricted. The civilized society has the responsibility to provide them with better prosthetics so they can live a normal life. Previously, prosthetic legs were developed for lower leg amputees. Today's electronic prosthesis provides a wide range of mobility compared to a mechanical prosthesis. Powered prosthetics are quite expensive and hence not widely used. We have proposed a low-cost control system above-knee powered prosthetics leg that is reliable, and the complexity of the model is less. The prostheses works by taking input from a sensor placed in the model and six FSR402 sensor and an IMU sensor that record force muscle activity of the thigh during variation of leg movement. The value of motion of the joint knee angle is determined by sensors. In this paper, we are discussing and comparing potential deep learning control system namely artificial neural network (ANN) as baseline and recurrent neural network-long short term memory (RNN-LSTM) to predict human gait cycle. A multi-layer perceptron (MLP) method is implemented. In general, network architecture consists of input layer which is represents by six inputs of the FSR402 sensor and an IMU sensor, fully connected layers, and an output layer as joint knee angle. As the result model of ANN as baseline has metric value MAE and R squared are 6.21 and 68.7% respectively whereas model of RNN-LSTM has metric value MAE and R squared are 4.99 and 96.9% respectively. The result indicates the RNN-LSTM model is most likely to perform fewer errors based on the dataset. RNN-LSTM outperforms the task with a greater value compared to the ANN model as the baseline. This present study suggests that deep learning with the recurrent neural network-long short term memory (RNN-LSTM) is suitable for the prediction of human gait cycle movement from the FSR402 and an IMU signal in the overall walking task. [ABSTRACT FROM AUTHOR]

In this study, we compare the effect of different hyperparameters of the deep neural network while solving partial differential equations. We consider feed-forward neural networks, and the hyperparameters such as the number of hidden layers, number of neurons, activation function, optimization methods, and learning rate are studied in this paper. Numerical results for the effect of all the hyperparameters while solving fourth-order partial differential equations are mentioned in this study. [ABSTRACT FROM AUTHOR]

Stepped, sinusoidal, and multi-level inverters are among the numerous types of inverters depending on the output voltage waveform. Inverters are most commonly used in commercial and industrial applications. Low output voltage harmonics are a distinguishing feature of the multilevel inverter. Inverters with several levels come in a range of designs and sizes. Because they require fewer components, cascaded multilevel inverters, for example, have attracted a lot of interest in the power industry. As the number of levels develops, the consonant substance of the yield voltage waveform diminishes. This paper illustrates a Cascaded H-Bridge fifteen-level inverter possible with the fewest switches. As the level increments, the synthesized yield waveform has more steps, coming full circle in a staircase wave that approaches the specified yield voltage waveform. By employing a limited number of switches, switching and power losses can be avoided. A fifteen-level inverter generally requires 28 switches, however the suggested inverter just requires 12. So that the inverter's cost can be kept as low as possible. The artificial neural network is trained to perceive the open circuit switches in the multilevel inverter and evaluated using data sets in order to eliminate problems that occur in inverters. Thus, using the suggested fault diagnosis method, a better understanding of fault behaviour, diagnostics, and detections of a multilevel inverter can be acquired with the help of ANN based fault detection using MATLAB Simulink and the results are analyzed. [ABSTRACT FROM AUTHOR]

The efficiency of the systems for the automated formation of artificial neural networks is beyond any disputes. Such systems make it possible to automatically generate ready-to-work models without participation of a domain expert while creating a model. The present paper considers a modification of the evolutionary algorithm for the automated formation of artificial neural networks of direct propagation, which helps build recurrent neural networks. The peculiarity of the algorithm lies in the network structure encoding in the form of a tree allowing the network structure to be represented more compactly, that well correlates with text analysis tasks, the dimensions of which vary from hundreds to thousands of features. To test the approach, the problems of classifying textual information were solved. The averaged test results as well as some of the obtained models are presented to demonstrate the approach. It is showed that the algorithm successfully builds models with a recurrent structure. This approach can be potentially improved by adding various recurrent links. [ABSTRACT FROM AUTHOR]

The electrocardiogram (ECG) shows the plot of the bio-potential produced by the movement of the heart and is utilized by doctors to foresee and treat different cardiovascular illnesses. The arrangement of electrocardiogram (ECG) signals assumes a significant function in the conclusions of heart illnesses. An exact ECG grouping is a difficult issue. Early and precise discovery of arrhythmia, Congestive Cardiovascular breakdown types is significant in distinguishing heart infections and picking suitab le treatment for a patient. Various classifiers are accessible for ECG orders. Among all classifiers, Convolution Neural Organizations (CNNs) like ALEXNET have become exceptionally famous and most broadly utilized for ECG grouping. This paper examined the issues engaged with ECG order and presents a definite studyof pre-handling strategies, ECG information bases, highlight extraction methods, CNN-based classifiers, and execution measures to address the referenced issues using the IoT-based mechanisms. [ABSTRACT FROM AUTHOR]

This paper investigates the most commonly used Deep Learning (DL) based techniques for estimating the wireless channel at the User Equipment (UE). The generalized Nakagami-m fading distribution is considered for deep learning network. The motive of this work is to provide an experimental investigation on the wireless channels using three different deep-feed forward neural networks, namely Artificial Neural Networks (ANN), Long Short Term Memory (LSTM), and Convolution Neural Networks (CNN). The analysis includes the preprocessing approach for handling complex received signals as one of the inputs and actual channel coefficients as output. Bit Error Rate (BER) performance analysis is performed and is compared with the classical LS estimator. It is demonstrated that DL channel estimation outperforms the Least Square (LS) based estimation, and simulation results support that DL-based channel estimation is very efficient. [ABSTRACT FROM AUTHOR]

Floods are a serious problem for communities located on the banks of rivers. They can destroy buildings, endanger lives, and cause economic losses. Hydropower plants can also affect the behavior of rivers and increase flood risk. In recent years, neural networks have proven to be a powerful tool for modeling various processes. In this paper, we will consider the development of a model of flood water passage, taking into account the impact of hydroelectric power plants, using neural networks. As a model object, we have chosen a settlement located in the northern part of Krasnoyarsk region at the confluence of two full-flowing rivers Angara and Yenisei. The model was developed using the python programming language. The recurrent network with long short-term memory LSTM using 64 neurons was chosen as the network architecture. The data array for modeling was obtained from Yenisei Basin Water Department and FGBU Sredne-Sibirskoye UGMS. The input parameters were selected empirically. The analytical forecasting method and the neural network model were compared. [ABSTRACT FROM AUTHOR]

Stress and emotion assessment play a pivotal role in healthcare, aiding in the early detection and prevention of various diseases. In this research paper, we propose a novel healthcare model that harnesses the power of artificial neural networks (ANN) and discrete wavelet transform (DWT) to assess stress and emotion through speech signal recognition. The proposed model extracts essential features from speech signals using DWT and employs ANN for the classification of stress and emotional states. Both ANN and DWT are established signal processing techniques that find applications in diverse domains. We present a comprehensive analysis of our healthcare model, which effectively analyzes speech signals to identify the stress and emotional states of individuals. The model's performance is rigorously evaluated using a publicly available dataset comprising speech signals recorded from subjects experiencing varying stress and emotional states. Experimental results demonstrate that the proposed model achieves high accuracy and surpasses existing state-of-the-art techniques. By employing recurrent neural networks (RNN), we visualize speech's phonetic characteristics, such as Mel-frequency cepstral coefficients (MFCCs), transforming them into sequences of phonemes or words. RNNs, designed to handle sequential data, form the basis of our speech recognition using neural networks, contributing to the model's effectiveness and accuracy. [ABSTRACT FROM AUTHOR]

Recently, researchers and clinicians have been searching for new technologies to slow down or stop COVID-19 pandemic. The utility of Data Mining (DM) algorithms to suggests new opportunities to combat the spread of the new Coronavirus. This paper suggests a comparative study on data mining approaches to predict COVID19. We used common classification algorithms like the Support Vector Machines, Random Forest, Logistic Regression, K-Nearest Neighbor and Artificial Neural Network with Python simulation to compare it in metrics accuracy, recall, precision and AUC; results showed that Random Forest model had a 98.43% accuracy – which is a higher accuracy than many other previous studies known COVID-19 data mining algorithms. [ABSTRACT FROM AUTHOR]

Heart disease forecasting is a key issue in the clinical data analysis field. Artificial Neural Network (ANN) is a machine learning tool that can help doctors diagnose heart diseases more accurately and quickly. However, the design of ANN is complicated because it needs to identify optimal weight values and suitable network structures. This paper aims to optimize the weights of ANN for forecasting the existence of heart disease among humans by using Artificial Bee Colony (ABC) algorithm to train ANN and select optimal network weights. The experimentations are carried out on the UC Irvine Machine Learning Repository (UCI) heart disease datasets and tested by the MATLAB software machine learning tools. The accuracy, specificity, sensitivity (recall), precision, and F-score of the ANN trained by the ABC model (ABC-ANN) and ANN trained by ABC with backpropagation (ABC-BpNN) model are investigated. Between ABC-ANN and ABC-BpNN, it was shown that the ABC-BpNN model has a better predicting ability and could reach considerably higher accuracy with 90% and 90.9%, 88.9%, 90.8% and 90.9% for precision, specificity, recall, and F-score, respectively. The findings showed significant enhancements compared with previous studies that have utilized the same dataset. [ABSTRACT FROM AUTHOR]

Face recognition is a biometric based application software capable of verifying or identifying exclusively a person by analyzing and comparing person's facial contours using patterns. The main applications of face recognition system is found in security systems, law enforcements and other initiatives. The variation in illumination, effects immensely on recognizing faces or images. Even for the familiar faces, illumination variations lead to alteration in the shadow shapes, shift in locat ions, fluctuations in features and contract gradient variations. Under these conditions, larger impact is observed to recognize these face images. A face recognition system is developed under these varying illumination conditions in this paper. Various methods of illumination normalization like Self-Quotient Image (SQI), Histogram equalization (HE), Gross & Barovick's anisotropic smoothing (GB), Gross & Barovick's anisotropic smoothing (GB) and Multiscale Retinex (MSR) are comparatively analyzed. Illumination invariant face recognition is developed by first acquiring 40 illumination variants face s. Face detection is performed by deformable template with scaled color map and its negative. Further face alignment is estimated by supervised learning by estimating the locations of a set of facial landmarks. Later image normalization is performed using Tan Trigg's Normalization Technique (TTVT) which includes Gamma correction, MaskingDifference of Gaussian (DOG) and contrast stretching methods [1]. The resultant normalized face image is classified using Convolutional Neural Networks. The evaluation and performance analysis are discussed to determine the applicability and suitability for various applications. [ABSTRACT FROM AUTHOR]

Currently, many companies are competing to produce robots that can help humans. Robots can work with high accuracy, efficiency, and low power in the production of goods. Automated Guided Vehicle is one form of robot that is useful in industry. Automated Cars are another form of functional industrial robot that transports with programmable paths and directions automatically. In this paper, we will discuss the navigation system in Automatic Cars based on previously mapped paths. The robot wheel serves to adjust the direction and position of the robot. The automatic movement can be carried out using computer vision. Warehouses that are not neatly arranged cause the navigation system to be less than optimal. The types of controllers used include classical controllers, fuzzy logic, artificial neural networks, and H∞. The purpose of this research is to produce a precise car front view monitoring system using image processing with high enough quality for automatic car navigation. The benefit for science is the development of automatic cars, which so far have not been widely carried out in Indonesia. The results of this study can be useful for the development of methods to build an automatic car system. The benefit for the electronics industry is the availability of control equipment designs to be able to control cars automatically. The effect on society is in the form of future vehicles that can move automatically without a driver. Based on the computer vision system test for Raspberry Pi-based car navigation, it can be ascertained that the car works well and has a small physique, so it looks neat and concise. The movement of the car from the first coordinate point to the next coordinate point is very precise and precise. By installing the L298N motor driver module, it is easier to set the program in python, so that the results sent to the DC motor can be used directly. Using a DC motor that has an excellent performance in terms of torque and speed. [ABSTRACT FROM AUTHOR]

Facial expression is an important basis for judging human emotions. The driver's mood can be accurately determined by recognizing their facial expression. Because the facial expression data set based on ordinary life scenes cannot accurately show the relationship between the driver's facial expression and his psychological state in the real driving environment. This paper uses 1000 facial expression images of drivers based in the real driving environment. These images include 6 different facial expressions to express 6 different emotions. This paper uses Edge Impulse to train several neural network models capable of recognizing these 6 different facial expressions and to realize the purpose of judging the driver's psychological state. From the experiment results, the MobileNetV2 model used has the highest accuracy rate of 99.1% on training data and 97.4% on testing data. This shows that the lightweight neural network model MobileNetV2 can accurately recognize the driver's facial expressions in the real driving environment. [ABSTRACT FROM AUTHOR]

In this paper, we aim to provide a method based on deep learning to help clinicians diagnose lumbar spine MRI for stenosis more efficiently and with instruments that produce noisy images to make maximal usage of medical resources. To achieve the same, this paper evaluates the effect of various MRI noises, such as Gaussian, salt and pepper, speckled, etc. in boundary delineation of stenosis in MRIs of the lumbar spine using recently introduced semantic segmentation algorithm based on deep neural networks. A U-Net model was trained for segmentation and its performancewas tested on various MRI images with added noise. The proposed methodology adopted in this paper is successful in producing a considerably good performance with higher scores, which helps in identifying Lumbar Spinal Stenosis in a more efficient manner. [ABSTRACT FROM AUTHOR]

The soil under any foundation should be capable enough to support the transmitted loads so that the structure– foundation–soil system is stable and safe against any settlement. Any change in the unconfined compressive strength is one of the principal causes of settlement. In this paper, an approach to forecast the unconfined compressive strength of gypseous soils based on basic soil properties was developed using the Artificial Neural Networks technique. An equation was developed to estimate the unconfined compressive strength using a backpropagation algorithm to train multi-layer perceptron, in which good agreements were achieved. The model inputs were; the depth, gypsum content, liquid limit, plastic limit, plasticity index, passing sieve #200, dry unit weight, water content and initial voids ratio. The output was the forecasted unconfined compressive strength. A parametric study was conducted to investigate the generalization and robustness of the model. The findings indicated that the model was reliable within the range of utilized data. The sensitivity analysis of the model revealed that the water content had the highest relative effect, followed by passing sieve #200, dry unit weight, liquid limit, plastic limit and gypsum content, while depth, initial voids ratio and plasticity index have lower relative effects. [ABSTRACT FROM AUTHOR]

This paper presents a complex prediction of travel times of city transport of Sofia (Bulgaria) using Artificial Neural Networks for the needs of On-demand public transport scheme proposed by INNOAIR project. The model is tested on a particular use-case data. For validation the root-mean-square error measured in seconds for each stop is used, calculated on the forecast for the test set (last week of observations). [ABSTRACT FROM AUTHOR]

The geomagnetic disturbance is one of the effects of space weather which is expressed by the Dst index. The Dst index is an indication of geomagnetic field activity at middle and low latitudes. In this paper, predictions of the Dst index are carried out using an artificial neural network using the NARX method (Nonlinear Autoregressive Network with exogenous input) with the input parameters consisting of the width angle of the coronal mass ejection, the position center angle of the coronal mass ejection, the coronal mass ejection velocity, the coronal hole area, solar wind velocity, solar wind density, and interplanetary magnetic field. The data used for learning, validation, and testing with the 7 exogenous input parameters above are from 1997 to 2019. The Dst index obtained is quite good with a regression between output and target of more than 98%. This model can predict the hourly Dst index one to five days ahead. From the 8 events that occurred in 2020, we found the average RMSE error is 11.41 nT. The Dst index prediction can be used as a guide in forecasting geomagnetic disturbances in providing space weather information. [ABSTRACT FROM AUTHOR]

Healthcare systems deal with a bulk of digital data everyday. Most of the healthcare organizations are focusing on using deep learning for supporting better patient care. Deep learning is nothing but the subset of Artificial Intelligence. The concept of deep learning arises from artificial neural networks. Deep learning performs very well on unstructured data. This paper aims at reviewing deep learning along with its various techniques and also its applications in medical healthcare system. [ABSTRACT FROM AUTHOR]

Metaheuristic Algorithms have gained remarkable attention in the past years due to their capability of optimization. These algorithms tend to provide near-optimal solutions to complex computational problems. Nature-inspired metaheuristic algorithms take inspiration of their approach from natural processes. In this paper, we have studied 3 nature-inspired metaheuristic algorithms: Genetic Algorithm, Particle Swarm Optimization and Ant Colony Optimization. Artificial Neural Networks are adaptive models that have the capability to learn from the past mistakes and keep improving on a continuous basis. These mimic the biological neuron network in a human brain. However, there is always a need to optimize these neural networks. Many tasks such as parameter selection, data training etc. are major challenges in ANNs. In this study, we see how numerous researchers have applied metaheuristic techniques to ANNs in order to tackle these problems. [ABSTRACT FROM AUTHOR]

Various methods have been proposed for the automatic speech recognition system. The main emphasis has been given to the signal or voice classification of individuals. In this paper, we are considering the four prominent neural network architectures namely, Multilayer Feed-Forward Neural Network, Cascade Backpropagation Network, Elman Backpropagation Network and Adaptive Linear Neural Network for the performance evaluation on collected sound samples of ten different people. These networks have been trained using mini-batch stochastic gradient descent learning with batch normalization. The feed-forward neural network model has been extended up to the four hidden layers to perform the desired classification of sound samples. The collected samples of the analog form of sound signals are converted into the digital form using the digital signal processing technique. The time & frequency patches of these digital signals have been considered as the training pattern samples. The test pattern sets have also been constructed in the same manner and the performances of these neural network architectures have been measured. The simulation results are exhibiting the better performance of Cascade Neural Network over the other architectures. [ABSTRACT FROM AUTHOR]

Genetic algorithms (GA) are one of the most popular choices for solving optimization problems under uncertainties, imprecision, partial truth inexactness in which the solution space is large. It is inspired by population genetics and natural selection. The goal is to optimize an objective function to its global maxima or minima under a set of imposed constraints. The usefulness of genetic algorithms lies in the fact that it can reduce the state-imposed complexity of such an optimization problem to many fold and thereby improving the search time to find the optimized solution in the solution space. Artificial neural networks and its many variants like CNN, RNN, R-CNN etc. has been instrumental in solving various supervised learning problems. However, it is generally accepted that parameterized training of a neural network is a difficult and time-consuming task due to its inherent hit and trial approach. In this paper we tend to use a genetic algorithm technique to optimize an Artificial Neural Network (ANN). The intended outcome is to optimize the learning rate, optimizer and weights of an ANN so minimize the cost function. We will run this experiment in an artificial neural network to create a prediction model to predict the cooling and heating load in a building. [ABSTRACT FROM AUTHOR]

This paper describes an novel classification approach of EEG signals for the detection of epileptic seizures using K-Nearest Neighbour and Artificial Neural Networks. : The total of 1372 samples are collected from the UCI repository which contain 15 attributes. The epilepsy detection is carried out with the help of two groups where group 1 is K Nearest Neighbour and group 2 is Artificial Neural Networks. K Nearest Neighbour Achieved 97% respectively compared to 88% by Artificial Neural Networks. The obtained significant value is (p<0.05). We can conclude that K-Nearest Neighbor has significantly greater accuracy when compared with the Artificial Neural Networks. [ABSTRACT FROM AUTHOR]

Ayurvedic medicines plays significant role in maintaining the biological state of human beings. The major components of ayurvedic medicines are plant leaves, roots, stem, bark, leaf stalk, flowers etc. Ayurvedic plants cure several disorders such as lung diseases, gastro-intestinal disease, cardiac disorder, skin disorder, excretory diseases etc. So knowledge about these plants sustained by generations must be conserved. Also appropriate classification and identification of ayurvedic plants is essential for researchers, botanist, cultivator and doctors. Manual identification will be time consuming, imprecise and inaccurate. This paper proposes an automatic recognition system for the identification of different types of ayurvedic plant images using deep Convolutional Neural Network (CNN) and Artificial Neural Network (ANN) based on transfer learning method. For the performance evaluation the dataset is generated by capturing images using mobile device from the environment and from Google network. The dataset consist of 424 plant images of 21 types of ayurvedic plants. A Graphical User Interface (GUI) is also developed to display the informations of the detected ayurvedic plant. [ABSTRACT FROM AUTHOR]

Driving while drowsy is one of the significant causes of road accidents, and detection of drowsiness of drivers is necessary to minimize such accidents. There are many methods proposed for drowsiness detection. In this paper, we present the details of a drowsiness detection system using deep learning techniques with ECG as its input. Continuous wavelet transforms (CWT) are applied on the input signal, segmented to 1-minute duration, and then it is converted into scalogram images to be fed as input to deep neural networks (DNN). We use pre-trained transfer learning models, AlexNet, ResNet-50, and VGG-19 for this work. The best results were obtained for Reset-50 with an accuracy of 88.31% without dropping the final layers, which is an improvement of 4.5% over the spectrogram-based implementation. This paper evaluates the effectiveness of wavelet-based features, and we note that the time-frequency domain features outperform. To check the robustness of the net-work, a second model is implemented by dropping the final layers, and the validation accuracy of AlexNet is 51.5%, ResNet-50 is 67.2%, and VGG-19 is 81.8%. [ABSTRACT FROM AUTHOR]

Space weather impact on the operation of technical systems is significant and may cause technogenic disasters. Timely forecast of space weather parameter variations allows one to decrease the risks of incidents in different spheres of life. To solve the problems of forecasting the energy volumes penetrating into the geosphere shells and initiating strong magnetic storms, the paper proposes an approach based on wavelet processing and neural networks. We presents the modeling results for AE index, which characterizes the energy volume penetrating into the magnetosphere during magnetic storms. The results of the constructed neuron network were compared with AE index initial data. The obtained neural network models can be used in the tasks of forecast of extreme impact on technical systems. [ABSTRACT FROM AUTHOR]

A popular approach in machine translation is neural machine translation. It predicts and translates word sequences using artificial neural networks, generally modelling full sentences in a single integrated model. It enables the translation of languages from one natural language into another with or without the assistance of human beings. Even though deep neural architecture-based machine translation produces accurate results when translating western languages, we are unable to straight away apply these algorithms to translate languages in India, primarily because good corpora are not readily available and our languages are generally morphologically bountiful in nature. This paper aims at providing a general overview on language translation, which consists of Neural Machine Translation and Natural Language Generation involving the translation of Indian languages. [ABSTRACT FROM AUTHOR]

In recent decades, the uncontrolled expansion and increasingly rapid urbanization of cities are leading to dramatic consequences, such as the fragmentation of ecological assets, the loss of biodiversity and, more generally, the deterioration of environmental quality. Therefore, it is becoming increasingly urgent to invest in the conservation and restoration of green areas, so that the urban system itself can in turn become a provider of ecosystem services. Such initiatives, besides generating positive environmental and social benefits, can lead to significant increases in the real estate values of the area involved in urban regeneration. The aim of this paper is therefore to characterise an innovative Artificial Intelligence (AI) model to assess the incidence of urban greening on property prices. Specifically, we propose an Artificial Neural Network (ANN) that includes among its inputs also proxy variables of environmental quality. The ANN, besides being little used to predict property prices, can set up non-linear relationships between inputs and outputs, and returns results that perform better than traditional forecasting models. The critical steps of the model concern the choice of input variables and the setting of the ANN. In the second part of the work, an application to a real case study will allow the model to be tested and demonstrate how measures to preserve green areas can have an impact on both the community and urban real estate. In conclusion, the aim is to show how AI models, also by integrating them with traditional forecasting models, can: on the one hand, provide valuers with a more rigorous set of information on property performance; on the other hand, enable to appreciate the impact of environmental externalities on real estate value, with consequent effects on the whole decision-making process. [ABSTRACT FROM AUTHOR]

In recent decades, the uncontrolled expansion and increasingly rapid urbanization of cities are leading to dramatic consequences, such as the fragmentation of ecological assets, the loss of biodiversity and, more generally, the deterioration of environmental quality. Therefore, it is becoming increasingly urgent to invest in the conservation and restoration of green areas, so that the urban system itself can in turn become a provider of ecosystem services. Such initiatives, besides generating positive environmental and social benefits, can lead to significant increases in the real estate values of the area involved in urban regeneration. The aim of this paper is therefore to characterise an innovative Artificial Intelligence (AI) model to assess the incidence of urban greening on property prices. Specifically, we propose an Artificial Neural Network (ANN) that includes among its inputs also proxy variables of environmental quality. The ANN, besides being little used to predict property prices, can set up non-linear relationships between inputs and outputs, and returns results that perform better than traditional forecasting models. The critical steps of the model concern the choice of input variables and the setting of the ANN. In the second part of the work, an application to a real case study will allow the model to be tested and demonstrate how measures to preserve green areas can have an impact on both the community and urban real estate. In conclusion, the aim is to show how AI models, also by integrating them with traditional forecasting models, can: on the one hand, provide valuers with a more rigorous set of information on property performance; on the other hand, enable to appreciate the impact of environmental externalities on real estate value, with consequent effects on the whole decision-making process. [ABSTRACT FROM AUTHOR]