International Journal of Intelligent Engineering and Systems

12

Authors: Shreenath Kannughatta Narasimhamurthy, Vasantha Kumara Mahadevachar, Ram Kumar Thayur Narasimhamurthy

Copy-move forgery detection (CMFD) is a well-known digital image forgery technique that is used in detecting forgery regions of images. The copy-move forgery in images occurs when a specific section of an image is attached to a new component of the same image to replicate the forged image parts as an original. The forgery appears to be realistic because even after being forged in the objective region, the image acquires all the basic qualities of the original image. Some post-processing functions, such as scaling or rotation invariant, noise handling, hybrid image manipulation, etc., limit the capability of CMFD models. To overcome these limitations, the features are extracted from the Hybrid feature extraction using modified speeded up robust feature (M-SURF) with accelerated KAZE (A-KAZE) feature description and detection algorithm. The main advantages of the SURF descriptor are: the SURF features are resistant to several post-processing attacks and use Laplacian of gaussian to differentiate background and foreground features clearly. The A-KAZE descriptor is robust in detecting scale-invariant of forged regions and poorly localized key points of objects. A-KAZE constructs the invariant scaling levels using nonlinear diffusion filtering. It smooths the image while preserving the edges and decreasing speckle noise. As a result, it captures the precise location of features during the feature extraction process. Following the feature extraction, the matched features are compared, and removed false matches by applying random sample consensus (RANSAC) technique. The performance evaluation of the proposed Hybrid MSURF with A-KAZE CMFD method is validated on MICC-F220, MICC-F2000, CoMoFoD datasets and surpasses the existing CMFD methods. The proposed CMFD achieved better results compared to existing methods such as when CMFD based on DHE-SURF features along with mDBSCAN clustering and CMFD based on SURF, BRISK features along with DBSCAN clustering measured in terms of Precision of 99.93%, Accuracy of 99.46%, F1-score of 98.21%, true positive rate (TPR) of 99.93%, and false positive rate (FPR) of 4.67%.

39

Authors: Amir Yasseen Mahdi, Siti Sophiayati Yuhaniz

This paper proposes an improved binary sparrow search algorithm (IBSSA) as a search strategy within the feature selection (FS) methods. Its main objective is to use clinical texts to improve COVID-19 patient categorization. The constant need for an efficient FS system and the favorable outcomes of swarming behavior in numerous optimization situations drove our efforts to develop a novel FS strategy. Additionally, clinical text data are frequently highly dimensional and contain uninformative features, which have a major impact on the classifier's accuracy, making FS a key machine-learning step in data pre-processing to reduce data dimensionality. The bi-stage FS approach is used in this work to elect the features. At the initial stage, we employed a term weighting scheme (TWS) that assigned a weighted score to each feature by measuring the significance of the features obtained from the pre-processing model using a new weight calculation method called root term frequency-core-inverse exponential frequency (RTF-C-IEF). Next, finding the most relevant and almost optimal feature subset for COVID-19 illness diagnosis is done in the second stage using a freshly developed methodology that was inspired by the way sparrow’s behavior. The suggested modification method for the sparrow’s algorithm is composed of several stages of advancement. The main objectives are to promote the exploration of the search space and increase the algorithm's variability. In order to evaluate the proposed model, various classifiers were employed on two datasets, each of which had 1446 and 3053 cases, respectively. The experimental and statistical results demonstrate that the proposed IBSSA is significantly superior compared to other comparative optimization algorithms, and it successfully upgrades the shortcomings of the original SSA. Moreover, the IBSSA has the highest accurate performance when compared to other rivals by the SVM classifier, Where, average removed features are 77.99% and 83.5%, with improvement percentages by F1-scores: 84.95% and 95.94 % for both datasets, respectively.

69

Authors: F. Al-dolaimy, Rabei Raad Ali, Noor Nabeel, Wisam Subhi Al-Dayyeni, Fatima Hashim Abbas, Hussein Muhi Hariz, Salama A Mostafa, Mohammed Ahmed Jubair

Vehicular ad hoc networks (VANETs) recently covered a wide range of intelligent transportation systems (ITSs) and applications. VANETs consist of convinced unique characteristics such as dynamic movement and high speed. Due to these characteristics, link failures occur, and delay and routing overhead is greatly increased, directly affecting the effectiveness, Quality of Service (QoS), and stability of VANETs. To achieve an efficient and reliable network performance, this paper proposes QoS Aware Hybrid Optimization for Improving Path Selection (HOIPS-VANETs) in VANETs. This hybrid optimization combines the Improved ant colony optimization (ACO) and Effective Whale Optimization Algorithm (EWOA). The EWOA algorithm is used for initial optimal path selection, and ACO is used to find the best optimal solution to achieve effective communication in VANETs. This optimization technique is also applied to low-density, medium-density, and high-density scenarios, as it is compared with the earlier methods. During the experimentation process, our findings reveal no noticeable change in the performance of the earlier methods when applied to the low-density and medium-density scenarios. Still, the performance is gradually reduced when applied to the high-density scenario. On the other hand, the performance of the proposed approach is better for all three tested scenarios and demonstrates effective communication for the VANETs.

94

Authors: Mohammed Kdair Abd, Luay G. Ibrahim, Firas M. F. Flaih

Incorporating renewable energy storage systems in power grids has presented significant challenges in maintaining a stable power generation structure and load frequency within interconnected grids. A promising solution to support the reliable and safe operation of the power system is the use of pumped storage units because of their excellent adjustment characteristics, including fast power response and convenient start and stop features. To optimize the performance of the pumped storage system, this article proposes the implementation of a linear-adaptive disturbance rejection controller (LADRC) in conjunction with the segmentation particle swarm optimization (Se-PSO) algorithm for parameter tuning. This article establishes a precise hybrid load frequency control (LFC) model for a two-area interconnected power grid incorporating pumped storage power plants (PSPP). It determines the optimal LADRC parameters for varying load disturbances in different areas of the power grid. The proposed control method systematically simulated pumping and power generation conditions in a PSPP. Simulation results indicate that the proposed method enhances dynamic stability by approximately 85%, offering greater anti-disturbance ability and adaptability than the traditional proportional-integral (PI) controller.

115

Authors: Arundhati Nelli, Rashmi Jogdand

Maximizing resource utilization by minimizing service level agreement (SLA) violation is considered to be extremely difficult process for scheduling of realtime workload on multi-cloud platform. In particular the proposed work considers both energy and performance as SLA minimization constraint for scheduling of workload in multi-cloud platform. The scheduling optimization of SLA minimization is done as single objective function using dragonfly soft computing approach. Further, the scheduling optimization of min-max is done as multi-objective for maximizing resource utilization with minimal SLA violation using dragonfly soft computing approach. The results show that the min-max workload scheduling (MMWS) has improved the resource utilization by 16.02%, 7.92%, and 2.52% respectively for montage workload and 9.92%, 4.07%, and 1.11% for SIPHT workload in comparison to the existing workload-scheduling adaptive-dragonfly algorithm (WS-ADA), reliability multi cloud-scheduler (REL-MC) and service-level agreement-based workload-scheduling (SLA-WS) method respectively. The results show that the MMWS has reduced the SLA violation rate by 81.92%, 69.23%, and 30.11% respectively for montage workload and 84.57%, 61.32%, and 12.62% for SIPHT workload in comparison to the existing WS-ADA, REL-MC and SLA-WS methods respectively. Hence, from the experiment outcomes it is shown that the proposed MMWS technique reduces SLA violation with improved resource efficiency in comparison with WS-ADA, REL-MC and SLA-WS.

137

Authors: Lakhvinder Singh, Dalip

An intelligent information technology-based system offering practical health advice can improve the superiority of lifespan for individuals with chronic illnesses. This study develops a BFCNN (Bees featured CNN) model for the classification or prediction of heart, liver, and kidney diseases. The main objective of the proposed classification model is to accurately diagnose a patient's disease. The research model aims to construct a highly efficient deep-learning system for patients with heart, liver, and kidney diseases. A novel approach for detecting multiple diseases in medical datasets using a BEE-featured CNN model. The BEE-featured CNN model is a modified version of the traditional CNN model that incorporates the behavior of bee swarms to enhance the accuracy of the model. The bee-inspired algorithm used in the model involves the use of multiple artificial bees that simulate the behavior of real bees to search for optimal solutions. The proposed model was trained and tested using a dataset consisting of medical datasets of patients with different diseases. The images were collected from various sources and labeled with the corresponding disease type. The dataset was split into training and validation sets, with 80% of the datasets used for training and 20% for validation. The results of the experiments showed that the proposed BEE-featured CNN model outperformed the traditional CNN model in detecting multiple diseases in medical datasets. The results of the prediction model demonstrate outstanding performance, with an accuracy of 96.7%, specificity of 99%, and Sensitivity of 97%. The research system demonstrates promising results in classification accuracy and holds potential for usage in medical and e-health applications.

164

Authors: Robert Marco, Sharifah Sakinah Syed Ahmad, Sabrina Ahmad

One of the main reasons that hinders making software effort estimation remains a most of the unresolved problem due to the heterogeneous nature of software data with complex structures. In processing nonlinear data, the long short-term memory (LSTM) model is often used for the purpose of solving discriminative and generative problems. Taking into account the fact that the LSTM network have low computational efficiency, due to the need to set a large number of hyperparameters. However, training deep learning models requires expensive work in terms of specifying hyperparameter configurations in the model. The grid search (GS) optimization method is used to find the best hyperparameter values for deep learning networks, which have many different hyperparameters that affect how well the network architecture works. In this paper, we proposed the grid search method as a quick step toward optimizing the parameters of the LSTM model. An empirical study was conducted using five datasets. From our results, we have seen that the LSTM-grid search model consistently performs better across datasets in mean absolute error (MAE) and root mean squared error (RMSE) compare to existing work on using machine learning approach for software effort estimation. The main advantage of training the LSTM show that the grid search finds hyperparameters which results in faster convergent, the possibility of generalization, and better coefficient of determination in estimating software effort.

191

Authors: Asmaa Ftaimi, Tomader Mazri

Recently, much effort and interest have been devoted to developing accurate and reliable blockage prediction systems. The latest research tends to apply artificial intelligence to enhance blockage prediction systems' accuracy, especially when dealing with non-line-of-sight radio signal propagation. However, AI models can carry inherent vulnerabilities that might damage the relevant system. Therefore, this paper aims to examine the effect of adversarial examples on AI-based blockage prediction systems. By performing adversarial example attacks on the AI-based blockage prediction model, we have shown the drastic impact on root mean square error that has arisen from 0.26 to 0.82 after introducing adversarial examples with 0.08 magnitude. Additionally, we have proposed a defensive approach based on encryption techniques to proactively prevent further compromise of the targeted system. The proposed method not only enhances data privacy but also help to prevent adversarial attacks that exploit transmitted data.

215

Authors: Shruti. R. Danve, Manoj S. Nagmode, Shankar B. Deosarkar

The extreme growth in smart mobile users and the demand for higher data rates have paved the way for developing 5G networks. Additionally, increased wireless data traffic implies more power consumption in network equipment. Mobile network service providers are concerned about the energy efficiency (EE) of cellular networks. In this era of advanced wireless communication, Massive multiple input multiple output (MMIMO) technology satisfies the demand for a hyper-connected society. To deal with the problem of increased hardware power consumption and cochannel interference in MMIMO, a quad methodology is proposed to evaluate the performance of a cellular network. This quad package comprises linear precoding (LP), power consumption reduction (PCR), User selection (US), and antenna selection (AS). The proposed quad methodology is used for enhancing the energy efficiency and capacity of a cellular network. Two LP techniques, zero forcing (ZF) and minimum mean squared error (MMSE) are compared to evaluate the EE of MMIMO with antenna selection. It is observed that EE for MMIMO with AS using ZF is three times greater than MMSE. MMIMO system with the suggested quad methodology provides 45% and 20% improvement in capacity and energy efficiency respectively over only the norm US.

239

Authors: Saif Mohammed Al-azzawi, Mohanad A. Deif, Hani Attar, Ayman Amer, Ahmed A. A. Solyman

Due to global lockdown policies implemented against COVID-19, there has been an impact on electricity consumption. Several countries have emphasized the significance of ensuring electricity supply security during the pandemic to maintain the livelihood of people. Accurate forecasting of electricity demand plays a crucial role in ensuring energy security across all nations; accordingly to achieve this objective, this study employs metaheuristics optimization algorithms to enhance the prediction model's operation, such as Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Random Forest (RF), at an optimized level to minimize errors. Two metaheuristics optimization methods, Particle Swarm Optimization (PSO) and Genetic Algorithms (GA), are utilized. The suggested prediction models are trained using daily power usage data from three US urban regions. In terms of prediction accuracy, the findings show that KNN with PSO surpasses the other models. The COVID-19 pandemic reduced power usage by 20% relative to pre-pandemic levels.

264

Authors: Purba Daru Kusuma, Faisal Candrasyah Hasibuan

This paper introduces a novel swarm-based metaheuristic called swarm magnetic optimizer (SMO). SMO imitates the behaviour of two magnets close to each other: pushing toward or pulling away from each other. This push-pull mechanism is then adopted to become a novel search in SMO. SMO is set up as a swarm of magnets that move autonomously. In the early iteration, the pull strategy is dominant. Meanwhile, it declines during the iteration and is replaced by a push strategy. The determination between these two strategies is calculated stochastically. SMO consists of three sequential phases in every iteration where the corresponding magnet and its reference run a push or pull search in each phase. The reference used in each phase is the global best magnet, a randomly selected magnet, and a randomly generated magnet. This paper evaluates SMO through simulation to find the optimal solution for 23 functions. The performance of SMO is compared with five latest metaheuristics: mixed leader-based optimizer (MLBO), guided pelican algorithm (GPA), zebra optimization algorithm (ZOA), coati optimization algorithm (COA), and clouded leopard optimizer (CLO). The result shows that SMO is better than MLBO, GPA, ZOA, COA, and CLO in consecutively 22, 15, 11, 13, and 14 functions. The superiority of SMO, especially, is in solving high dimension functions. Meanwhile, in the fixed-dimension multimodal functions, SMO is still superior to MLBO but less superior to GPA, ZOA, COA, and CLO.

289

Authors: Kunjan D. Shinde, Vijaya C

The advancement in very large scale integration (VLSI) technology and field programmable gate array’s (FPGA) parallel constructive nature for digital circuits has made the implementation of finite impulse response (FIR) filters increasingly relevant in real-time. In FIR filters the computational complexity increases with the length of the filter; numerous techniques have been developed to design viable architectures for realizing FIR filters. The multiple input multiple output (MIMO) based parallel FIR filter architecture often requires a large area for realization as the level of parallelism and filter order increases; this study presents an architecturally enhanced, novel parallel architecture for FIR filter design which eliminates the resource dependencies on the level of parallelism. The proposed architecture addresses the issue by limiting the number of multipliers required irrespective of the change in the level of parallelism and further modifies the data flow to improve the iteration period in the proposed design. The level of parallelism (L) is fixed to 8 in the presented study as the value is neither too low nor high; it is important to note that the increase in the level of parallelism will increase the number of samples generated at a given time. The presented work is demonstrated using tap 16 FIR filter implemented on FPGA VIRTEX 4 xc4vsx35-10ff668 and VIRTEX 5 XC5VSX95T-1FF1136 platforms, and filter-order 16 FIR filter is implemented on ARTIX 7 xc7a200t-2fbg676. The functionality is verified using Xilinx ISE 14.7 and Matlab 2018 environments. The post-synthesis results and comparative study show the design validation; the proposed architecture outperforms the benchmark and conventional MIMO-based parallel FIR filter architectures with an improvement in total delay by 70% w.r.t. [2] and 93% w.r.t. [3] and 2% w.r.t. [4] as compared in case studies 1 – 3. The reduction in area metric is observed as dipping in the slice requirement by 31%, 47%, and 110% and dipping in the LUT requirement by 58%, 66%, and 65%, further, the delay metric is improved by 49%, 80%, and 67% when compared with MIMO-based parallel architecture of [1] on the same platforms suggested in [2-4] respectively.

318

Authors: Rafael M. J. O. Silva, Attilio Sbrana, Paulo A. L. de Castro, Nei Y. Soma

Deep learning models, despite their potential, often function as “black boxes”, posing significant challenges to interpretability, particularly in sensitive fields such as healthcare and finance. Addressing this issue, we introduce a novel, human-understandable metric aimed at enhancing the interpretability of local interpretable model-agnostic explanations (LIME). Distinct from previous methodologies, this metric is designed to assess the shift in classification probability upon the removal of features (words), thereby providing a unique insight into interpretability. We deploy a convolutional neural network (CNN) for sentiment analysis, interpret predictions utilizing LIME, and evaluate these explanations using three distinct metrics: our proposed metric, a conventional model-based metric, and human evaluations. Through rigorous validation, our metric demonstrated high recall performance, a key indicator of relevant instance retrieval. Results showed worst-case and best-case recalls of 80.29% and 98.19% respectively, against a logistic regression metric for “good” and “excellent” classifications. Comparisons with human evaluations using single-word explanations revealed worst-case and best-case recalls of 82.03% and 94.37%, respectively. These high recall values highlight our metric's effectiveness in aligning with both human judgments and model-based metrics, emphasizing its capacity to capture essential explainability aspects. Furthermore, our study also outlines certain LIME limitations, setting the stage for future interpretability-focused AI research.

348

Authors: Abhishek Singhal, Devendra Kumar Sharma

Voice signals are the most convenient and essential method to establish the communication between human beings. The linguistic characteristics of the human being can be recognized with the help of voice signal analysis such as the ascent of language, health status, emotions, age, gender, etc. Gender is an unprecedented peculiarity of the speakers. This peculiarity can be recognized through the analysis of the voice signals. This paper presents the precision value of gender, i.e., transgender, female and male. The live voice samples for females and males are recorded as well as the live voice samples for the transgender in India are recorded first time because the voice samples of the transgenders are not available at the authenticated sources. The error factor and accuracy is also calculated for the proposed model, which is utilized for the gender identification of the speakers by analyzing the voice signals. In the proposed work, bidirectional long short-term memory (BiLSTM) architecture with recurrent neural network (RNN) model is used as a classification algorithm with mel frequency cepstral coefficients (MFCCs). The MFCC is utilized as a trait of the signals used to identify the gender of the human beings. The outcomes of the proposed model are not dependent on the content of the articulation and the dialect of the speakers. The precision values for female, male and transgender, the accuracy and error rate of the proposed model, depend on applied quantities of voice signals in testing and training datasets. The average precision value is 93.91%, 94.70%, 97.43% and 95.35%for the male, transgender, female and proposed model, respectively. It is found that the female precision value is the highest, and the male has the lowest. Further, the average error rate of the proposed model is 0.045. The gender identification accuracy is 95.52%.

372

Authors: Robert Marco, Sharifah Sakinah Syed Ahmad, Sabrina Ahmad

The uncertainty inherent in the predictive environment of software effort estimation, is due to the limited availability of data information and the expensive costs associated with data collection. Consequently, there is difficulty in making accurate predictions with insufficient software effort estimation data, due to the limited information in the data. The probabilistic deep generative model conditional variational autoencoder (CVAE) is capable of producing synthetic data and modeling complex data distribution characteristics that is similarity to the real Unfortunately, because the method was previously developed for the classification task, the application of the technique to the regression task has received little attention. This study aims to construct a CVAE model combined with the relevance function contained in the re-sampling method. Relevance function method aims to create a label on the target. Statistical tests, such as the Levene test, t-test, and Kolmogrove Smirnov test, are utilized to compare the results to the real data. This study considers seven popular algorithms for comparison, such as synthetic minority oversampling technique regression (SMOTER), generative adversarial networks (GAN), conditional tabular generative adversarial networks (CTGAN), gaussian-copula (GC), copula GAN (CGAN), and variational autoencoder (VAE). CVAE approach has the best augmentation performance quality with accuracy values (MAE; RMSE; RAE; R2) compared to existing models in each china dataset (923.330; 1470.536; 0.436; 0.848) and desharnais dataset (9.540; 11.694; 0.041; 0.978), respectively. The proposed model generates synthetic data which is similarity to real data in the software effort estimation context. In addition, synthetic data improves the quality of performance on baseline machine learning in software effort estimation contexts rather than using real data.

398

Authors: Seema Sabharwal, Priti Singla

Sign language translation through deep learning is a popular topic among researchers nowadays. It opens the doors of communication for deaf and mute people by translating sign language gestures. The translation of input sign language gestures into text is called the sign language translation system (SLTS). In this paper, optimised machine learning-based SLTS for Indian sign language (ISL) has been proposed to facilitate deaf-mute persons. Further, this paper presents a simulation analysis of the impact of the number of convolution layers, size of stride function, epochs, and activation function on the accuracy of translation of ISL gestures. An optimised ISL translation system (ISLTS) for fingerspelled alphanumeric data of 36 classes using a convolution neural network (CNN) with a novel RADAM_NORM optimiser has been proposed. The proposed system has been implemented using two datasets- the first customised ISL alphanumeric dataset has been taken from Kaggle and the second dataset has been prepared by the author consisting of 36 classes and nearly 50K images. The accuracy of the proposed ISLTS on the first dataset is 99.446 % and on the second dataset is 97.889%.

422

Authors: Ibrahim A. Zedan, Mona M. Soliman, Khaled M. Elsayed, Hoda M. Onsi

SIFT-based techniques have achieved satisfying performance in detecting copy-move forgery (CMF). Typically, these techniques find the matched regions of an image and re-examine them using a variety of methods to determine whether CMF has occurred or not. However, these techniques have some shortcomings related to how they handle false matches, which usually occur due to image continuity or self-similarity. First, a spatial distance threshold or segmentation-based methods are commonly utilized to handle image continuity. Second, several external methods along with manually created thresholds are utilized to handle image self-similarity. In this paper, we propose a new matching strategy that is resistant to false matches while reducing reliance on external methods and thus avoiding several thresholds. We model the keypoint as a whole region rather than a single point and employ the intersection over union measure to deal with image continuity. To reduce false matches caused by image self-similarity, we combine the cross-matching test with a modified distance ratio test. This combination takes into account the ability to detect multiple cloning. Moreover, we utilize a support vector machine to learn the threshold(s) needed to decide if CMF has occurred or not. The proposed methodology is evaluated over three challenging datasets: MICC-F600, Coverage, and MICC-F220. On MICC-F600 dataset, our proposed methodology outperforms other state-of-art techniques and achieves high precision of 99.38%, recall of 97.5%, and 98.42% of F1 score. Additionally, the comparative evaluation using Coverage, and MICC-F220 datasets proved the effectiveness of the proposed methodology to handle a variety of attacks.

450

Authors: Suadad Zaidan Khalaf, Mohamed Ibrahim Shujaa, Ahmed Bahaaulddin A. Alwahhab

Law enforcement and civilian protection are increasingly dependent on surveillance technologies. To prevent social, economic, and environmental harm, video surveillance situations like those in train stations, schools, and hospitals must automatically detect aggressive and suspicious behavior. Intelligent video surveillance systems now enable human interaction monitoring. Despite its security benefits, crowds and the camera's vision make it difficult to distinguish regular from abnormal activity. Therefore, there is a great deal of investigation into the many methods of identifying violent behavior. The detection approach presented in this research can be broken down into three distinct subfields. The classification methods employed led to the formation of these classes. Classifications include Gradient Descent (SGD), Adaptive boosting (ADA), Naive Base (NB), and Logistic Regression (LR) for detecting violent acts with machine learning. Methods for feature extraction and violence detection are derived using Principal Component Analysis (PCA). Dataset learning- based models for violence detection can also be evaluated using the AIRTLab dataset, a model developed to test the robustness of algorithms against false positives. In order to determine if the proposed model's resistance to false positives, accuracy (ACC) metrics were computed on the proposed model and used to set a benchmark for the AIRTLab dataset. According to the research literature, the proposed models are accurate of 98.31% with the logistic regression classifier having superior generalization skills.

472

Authors: Lusia Rakhmawati, Hapsari Peni Agustin Tjahyaningtijas, Wiyli Yustanti, Wiryanto Wiryanto

Due to extensive Internet usage, digital media can be quickly altered, copied, and distributed via information networks. As a result, digital information's illicit reproduction became a serious issue. As a means of copyright protection for digital media, digital watermarking has been demonstrated to be effective. Based on a discrete cosine transform (DCT), an improved robust watermarking technique considering image characteristics, a 1-D linear transformation, and an optimal embeddable AC coefficient is proposed in this paper. The host image is first divided into 8 x 8 non-overlapping block and block selection for DCT transformation is performed. The variance in each block is used to identify the embedding watermark location. A linear map is used to scramble the watermark image, which is embedded into the AC coefficients which has a maximum difference between embeddable and non-embeddable non-zero AC coefficients. Experiment results show that the suggested approach can withstand a range of a non-malicious attack with very satisfactory results for BER close to 0, and NC near to 1.

497

Authors: Noora Nazar, Qusay Al-Doori, Omar Alani

Low density parity check (LDPC) codes are highly regarded for their exceptional error correction capabilities, making them a preferred choice for error correction coding (ECC) in modern communication standards. These codes are implemented in specialized integrated circuits (ICs) that facilitate data transmission. However, the power requirements of complementary metal-oxide-semiconductor (CMOS) circuits, which operate across a wide frequency range, present a significant challenge for future communication systems. To address this challenge, this paper proposes a novel multi-frequency power reduction strategy for LDPC encoders. The strategy utilizes dynamic voltage and frequency scaling (DVFS), a well-established power reduction method in CMOS circuits. By combining DVFS with fuzzy logic control, the system optimizes the voltage supplied to the LDPC encoder, achieving substantial power reduction while maintaining coding efficiency, flexibility, and performance. The proposed approach dynamically adjusts the encoder's voltage levels based on real-time conditions, ensuring efficient power consumption across different frequencies. This innovative strategy aims to overcome the power challenges faced by future communication generations, which have been largely overlooked in previous research efforts. This paper aims to assess the effectiveness of a multi-frequency power reduction strategy for LDPC encoders. Through extensive analysis and evaluations, it demonstrates significant power reduction while maintaining LDPC coding efficiency. These findings contribute to the development of power-efficient LDPC encoder designs, meeting the evolving needs of communication systems. Results reveal the system's performance across different frequency ranges. In the low-frequency range (1 MHz to 100 MHz), a remarkable 90% power reduction is achieved. In the medium-frequency range (100 MHz to 1 GHz), the reduction is 50%, while in the high-frequency range (1 GHz to 5 GHz), it reaches 20% to 14.5%. Notably, the system operates most effectively in the low and medium frequency range, providing substantial power savings. These results highlight the potential of the multi-frequency power reduction strategy for LDPC encoders in addressing power challenges across different frequencies. The significant power reduction achieved demonstrates its effectiveness and adaptability to varying frequency demands. This contributes to the development of power-efficient LDPC encoder designs, optimizing performance and energy consumption in various communication systems.

523

Authors: Dhiza Wahyu Firmansyah, Riyanarto Sarno

Breath acetone concentrations were found to be correlated with blood ketone levels. Based on this evidence, predicting blood ketone levels using breath analysis and machine learning (ML) becomes possible. Nevertheless, a good ML model requires a large amount of training data. Under certain conditions, it is difficult to collect large amounts of data such as during the Covid-19 pandemic. To overcome this problem, we propose an augmentation technique to extend the number of training datasets using two step synthetic minority oversampling (SMOTE). The first step was to increase the amount of training data by combining it with synthetic data, while the second step was to balance the data at each ketone level. The strategy for using SMOTE with regression was further explained since this study aims to predict ketone levels with numerical output values and SMOTE is typically used in classification cases. The proposed method was evaluated by entering the data into several ML methods such as deep neural network regression (DNN-R), linear regression (ML-R), ransac regression (RC-R), K-nearest neighbour regression (KNN-R), decision tree regression (DT-R), random forest regression (RF-R), Ada boost regression (AD-R), Gradient boost regression (GB-R) and XG-boost regression (XGB-R). Based on the test results, when compared without the proposed method, an increase in accuracy was obtained on DNN-R, ML-R, RC-R, KNN-R, DT-R, RF-R, AB-R, GB-R, and XGB-R by 0.958%, 9.51%, 35.74%, 18.133%, 8.236%, 11.348, 9.47%, 5.093%, and 11.264% respectively.

548

Authors: Wissal Farsal, Mohammed Ramdani, Samir Anter

The graded multi-label classification (GMLC) is an extension of multi-label classification. Whilst a multi-label classifier is limited to predicting the set of relevant labels, a graded multi-label classifier predicts the degree of relevance of a set of given labels. A key challenge of this learning problem consists of modelling the dependencies among the labels to improve the predictive accuracy. The algorithm adaptation-based solutions, which modify the algorithms directly to handle GMLC were proven effective in modeling these dependencies in comparison to the transformation-based models which reduce the graded multi-label datasets into a set of multi-class or binary datasets. In this paper, we propose an adaptation of random forest algorithm (GML_DT) with an adapted CART (classification and regression trees). The adapted algorithm is based on a modified formula for the Gini Index which fully models the label dependencies. The performance of the new model was tested on a 101 benchmark datasets and compared against the most influential methods for GMLC. The evaluation metrics considered in the experimental study are proper to the graded multi-label setting, i.e., hamming loss and vertical 0-1 loss. The experimental results show that the proposed model outperforms the considered models for the graded multi-label performance metrics. The increase in performance and overall accuracy is quantified by a decrease in the hamming loss of more than 13% and a decrease in vertical 0-1 loss that exceeds 10% when compared to the state-of-the-art models.

568

Authors: Christy Atika Sari, Muhammad Hafizh Dzaki, Eko Hari Rachmawanto, Rabea Raad Ali, Mohamed Doheir

The exchange of confidential information should be done in a secure environment. Therefore, security is needed if the exchange of information is carried out using Internet media. The way to secure the information can be done using cryptography and steganography techniques. and vigenere cipher had been applied as cryptography techniques and the LSB method as steganography techniques. This research has also compared the effect of the Fibonacci sequence in the encryption process. In this study, six images were used, each with a different extension, size, and type of images. All these images will be inserted with the secret message resulting from the encryption that applies the Fibonacci sequence or without applying the Fibonacci sequence. Fibonacci series is used to generate a random key, which is used for encrypting the message in playfair encryption algorithm. The level of effectiveness of the Fibonacci sequence has been determined and measured using MSE, PSNR, Entropy, UACI, and NPCR. The highest PSNR was in a 512x512 pixel grayscale image with a result of 7.9632, the best PSNR obtained was more than 85 dB, while the best UACI obtained was more than 49% using Fibonacci and 34% without implementing Fibonacci. The NPCR obtained is both above 99%, it is just that the image that has been implemented by Fibonacci produces a slightly higher UACI.

591

Authors: Balaji Sreenivasulu, Anjaneyulu Pasala, Gaikwad Vasanth

Early discovery and precise classification of plant-lead diseases are used to handle their spread and enhance the overall yield and product quality. The deep learning approach attains significant achievements in image classification and recognition. Since the classification using deep learning mainly depend on large-scale dataset for preventing the overfitting issue. Image augmentation is required to be developed to eliminate the risk of overfitting during the classification. In this research, the deep learning model based augmentation namely uncertainty based progressive conditional generative adversarial network (UPC-GAN) is developed for improving the plant leaf disease classification. The UPCGAN is used to map the images from one domain to another domain in a paired manner and estimate the uncertainty with the created images. Moreover, UPCGAN performs pixel wise residual distribution using the independent distributed zero mean generalized Gaussian distribution (GGD). The progressive learning of UPC-GAN increases the differences in the augmented synthetic images for improving the classification using DenseNet121. The dataset used to evaluate the proposed UPCGAN-DenseNet121 method is PlantVillage dataset. The performance of UPCGAN-DenseNet121 is analysed using accuracy, precision, recall and F1-score. Existing research such as deep convolutional GAN (DCGAN)-GoogleNet, conditional GAN (CGAN)-DenseNet121 and Fast wide and deep feature extraction block (WDBlock) based GAN namely FWDGAN are used to evaluate the UPCGAN-DenseNet121. The accuracy of UPCGAN-DenseNet121 for 10 classes is 98.2%, which is high when compared to the DCGAN-GoogleNet, CGAN-DenseNet121 and FWDGAN.

611

Authors: Fatima Abdul Sattar AL-Taie, Ahmed Sabah Al-Araji

In this research, a predictive voltage neural controller and remote monitoring for the nonlinear proton exchange membrane fuel cell (PEMFC) system are implemented in real-time. This paper's major purpose is to precisely and rapidly determine the appropriate hydrogen partial pressure (PH2) control action to improve the fuel cell's nonlinear performance under varying load currents and to enable remote monitoring of the nonlinear PEMFC system based on the internet of things (IoT). The laboratory, virtual instrument engineering workbench (LabVIEW) package is used to demonstrate the real-time performance of the proposed predictive voltage controller applied to the 150-watt PROTIUM PEMFC, which will be used to generate the appropriate amount of hydrogen partial pressure control action that will enter the fuel cell for stabilizing the desired output voltage. The message queuing telemetry transport (MQTT) protocol and a Raspberry Pi 4 acting as a local server are the building blocks upon which the monitoring component of the proposed system is implemented in order to monitor the desired output voltage, the fuel cell output voltage, and PH2. The Raspberry Pi collects the necessary fuel cell data and sends it to the node-red dashboard for monitoring. According to the simulation and the experimental results obtained using the proposed predictive controller on PROTIUM PEMFC, the proposed controller can generate an accurate, prompt, and timely reaction to the hydrogen partial pressure control action to reduce the tracking voltage error and to get rid of the fuel cell output voltage oscillation. The proposed experimental work was compared to the simulation findings to confirm its effectiveness in terms of effectively tracking the desired output voltage, providing a fast response, and achieving the optimal partial pressure of hydrogen. However, in the simulation findings, a voltage error of 0.01 volts was observed without any oscillation. On the other hand, the experimental results indicate a slightly higher voltage error of approximately 0.1 volts, accompanied by oscillations of around ± 0.1 volts.

642

Authors: Yamamh Alaa Alamari, Ahmed Fanfakh, Esraa Hadi

Currently, there are around 4.95 billion people who use the internet, which creates a large audience with an increasing demand for activities that can be done online. Some examples of these activities include social networking, information sharing, and online shopping. Therefore, there is an urgent need for improved levels of secrecy and privacy. It has recently come to light that one of the most serious challenges to the mainstream adoption of business apps is that of online fraud. As a direct consequence of this, issues pertaining to authentic cation, authorization, and identification have emerged as critical considerations in today's open and accessible society. The act of recognizing an entity, whether it be a human, a computer, or a software program, is referred to as the identification process. Authentication and authorization are complementary processes that are used in security systems to decide which persons are permitted to access information resources over a network. Authentication is performed by the user's device, while authorization is performed by the security system. However, a variety of potential solutions have been suggested, one of which is the use of parallelism in order to boost the efficiency of authentication methods. This work proposes a new message authentication algorithm that is applied in parallel over a multicore processor. The proposed message authentication algorithm uses two PNGRs and two substitution boxes to encrypt and authenticate the plain message. The encryption process has a one-round operation, making it a fast technique to encrypt and decrypt blocks of messages. In comparison to the existing method, the proposed authentication method outperforms the parallel speck-based authentication method by an average of 3.27 times faster when executed over a multicore CPU. The average speedup compared to the sequential version of the proposed algorithm and its parallel implementation is 2.99. The proposed method passes the most difficult randomness test, and the obtained MAC values are tested further to meet other security measurements.

669

Authors: Al_hussein M. Alturfi, Sonia Goyal, Amrit Kaur

Artificial bee colony (ABC) is a very popular algorithm and has been widely used in the literature. The algorithm has some inherent drawbacks, including slow convergence, and poor exploration versus exploitation, among others. In order to deal with these problems, we are hybridizing the ABC algorithm with the JAYA algorithm. The proposed algorithm has the added properties of both ABC and JAYA and has been named as JABC algorithm. The key idea is to add prospective equations of JAYA into ABC for better exploitation, and adding new mutation operators for enhanced exploration and better convergence. The proposed algorithm is tested on CEC 2005 benchmark problems and real world synthesis of linear antenna array (LAA). SLL reduction in five different LAA’s is done, for position, amplitude, and phase optimization. We use 10-element, 24-element, 28 element and 40-element LLA for optimization. For performance evaluation, the algorithm is tested with respect to basic ABC, JAYA, spider monkey optimization (SMO), Moth flame optimization (MFO), Chameleon Swarm Algorithm (CSA) and other algorithms. Experimental results on both benchmarks and LAA problems, show that JABC is significantly better as compared to others. Also, statistical analysis using Wilcoxon’s test and Friedman tests shows the superior performance of JABC algorithm with respect to other algorithms.

19

Authors: Ashraf Ali SK, Prasad M.V.D, Hima Bindu G, P. Praveen Kumar P, Anil Kumar D, Kishore P.V.V

Sign language recognition in real time has been leveraged by the continuously varying hand movements in both shape and orientation across Spatio-Temporal dimensions. This is accomplished by either independent view or shared view feature learning. However, information movement between views is neither total nor restricted to achieve view insensitivity during testing where all views are needed. The objective is to perform pose based multi view sign language recognition by applying triplet loss on a pair of specific and shared view features. Shared view features are obtained using view compatibility matrix which maps within class between view features and between class within view features. This mapping helps in increasing information flow between views from the same class and restricting it between classes thereby making highly discriminative feature representation for all views. Subsequently, metric learning enables to build a view invariant feature embedding by stacking view specific and shared features from different layers for training deep models. In the end, a blended view feature representation is obtained per class. Experiments were designed on our multi view skeletal sign language video dataset and three benchmark action datasets. The results of the experimentation have shown that the performance of the classifier has improved by 8% over the linear view combiners such as Laplacian eigenmaps. Further, the proposed model is useful in constructing a view invariant feature for recognition of multi view sign language.

48

Authors: Md. Irshad Hussain B., Mohamed Rafi

Now-a-days biometrics is used to recognize an individual, and is thought of a more secure way to authorize an individual. Face is the most important factor to recognize an individual. In order to enforce security features, many verification systems use face recognition to identify a person, such as - automobile security, security of home, smartphones owner recognition, attendance system, identification of criminals, etc. Face biometric works in two phases –face detection and face recognition. In this paper we are proposing a hybrid model that works in two phases, face detection is the first phase which includes two stages, in first stage it detects the face from the acquired image and then in the second stage acquired image is subjected to intensity equalization. The second phase is of Face recognition, where finally the individual is recognized as authenticated user or not. We are using Viola-Jones algorithm to detect the face, histogram stretching for intensity equalization and principal component analysis (PCA) algorithm for image recognition. The effectiveness of the suggested facial recognition system is shown through experimental findings using the face dataset constituting of 100 subjects using VTU-BEC-DB multimodal biometrics database with constraints of pose and illumination in the face images, without the intermediate step of intensity equalisation, we had a recognition rate of 93.60%. However, when the intermediate step of intensity equalisation was included, a novelty was added to the hybrid algorithm, and we achieved a recognition rate of 98.20%, which is higher recognition rate than the state-of-the-art methods.

75

Authors: D. Shankar, G. Victo Sudha George, N kanya

Security is considered one of the major challenges in the network due to the increase in network services. Many systems linked to the network play a major role in business and other applications that provide network services. Therefore, it is necessary to identify effective ways to secure the system. One of the significant information security technologies is the intrusion detection system (IDS) that utilizes various deep learning (DL) and machine learning (ML) algorithms to identify network issues. As a result, a novel DL technique is used to classify and detect attacks in the network. To perform this approach, pre-processing is initially done for the collected data using Z-score normalization (ZsN). Then, feature selection is performed using the hybrid pattern search whale optimization algorithm (HPS-WOA). Followed by, network attacks are detected and classified using the optimum bi-directional long short-term memory with the gated recurrent unit (OptiBiNet_GRU). Finally, the hyper parameters such as loss and weight of the proposed DL model are optimized using the pelican optimization algorithm (POA). The effectiveness of the proposed method is validated using two datasets on various performance metrics with some existing baselines. The results show that the proposed method outperforms other methods in terms of accuracy of 99%, precision of 95%, f1-score of 97%, specificity of 99%, and recall of 98% for UNSW-NB15 dataset, whereas accuracy of 99%, precision of 95%, f1-score by 99%, specificity of 99%, and recall of 99% for CICIDS 2017 dataset, respectively.

105

Authors: Gopalsamy Vanitha, Palaniswamy Amudha, Subramaniam Pillai Sivakumari

WSN has been widely used in many sensitive applications and it also has novel possibilities for laying the groundwork for using ubiquitous and pervasive computing, but it has also presented a number of issues and challenges, such as a dynamic network topology and a congestion problem that hinders not only network bandwidth utilisation but also performance. Proficient rate control and fair bandwidth allocation (PRC-FBA) was one of the schemes in the literature to solve issues of WSN by combining the ideas of traffic class priority and bandwidth fairness. However, because of the nature of WSN, the energy of nodes near the sink node is diminished when packets move from lowly congested nodes to highly congested nodes. This paper proposes a proficient rate control with data aggregation and fair bandwidth allocation (PRCDA-FBA) to address this problem by using an effective data aggregation approach for reducing the number of transmissions. In the proposed method, fair bandwidth allocation is simplified by an artificial intelligence-based bandwidth prediction method. Thus, PRCDA-FBA increases the network's durability. Despite having lower bandwidth utilizations, energy-critical sensor nodes require careful power management to avoid being eavesdropped upon. Along with data aggregation and fair bandwidth allocation, the effects of overhearing packets by energy-critical nodes are mitigated through network-wide route adjustments based on the energy level of nodes. Thus, in the proposed method, data aggregation is scheduled based on the availability of bandwidth, energy, queue size and packet priority. The proposed method is named as energy-aware proficient rate control with data aggregation and fair bandwidth allocation (EPRCDA-FBA). The proposed algorithms have been deployed on the Network Simulator 2.35 platform, and a comparative analysis has been performed using several metrics, including throughput, packet loss, End-to-End (E2E) delay and energy utilization. The EPRCDA-FBA method archives highest throughput which is 9.17%, 5.48%, 4.68% and 2.45% higher than congestion control strategies like discrete-time sliding mode congestion controller (DSMC), weighted priority based fair queue gradient rate control (WPFQGRC), PRC-FBA and rate adjustment-based congestion control (RACC).

126

Authors: Zuhair A. Ahmed, Safanah M. Raafat

Accurate localization and mapping are essential for autonomous navigation systems. The simultaneous localization and mapping (SLAM) algorithm, gmapping, is widely used for creating two-dimensional occupancy grid maps (2D-OGMs) due to its low cost and effectiveness in indoor environments. According to the SLAM interaction between localization and mapping terms, the evaluation of map accuracy implicitly means the evaluation of localization. This study focuses on improving the accuracy of the estimated 2D-OGMs by fine-tuning the 32 initialization parameters of gmapping using a turtlebot3-burger robot. An improved experimental procedure was developed, incorporating image registration and similarity measurement to evaluate the map accuracy. The results show a substantial improvement in map accuracy, from 78.84% to 94.18%. The study highlights the importance of fine-tuning the SLAM algorithm for improved map accuracy and provides valuable insights for developing autonomous navigation systems. The key contribution of this paper lies in the systematic classification, fine-tuning, and evaluation of the Gmapping initialization parameters.

151

Authors: Ujjwala S Rawandale, Sanjay R. Ganorkar, Mahesh T. Kolte

In recent days, one of the most serious health problems in humans is hearing loss. For helping the hearing impairment, the Hearing Aid (HA) system is used, but it may lack in various aspects such as delay, matching error, power consumption, low complexity and unwanted noise. Many techniques were introduced to overcome these issues, but they still have problems. Therefore, this research uses a novel Aquila-based Adaptive Filter (AbAF) for improving the HA system with the optimized parameters of filter. Here, the incorporation of the Aquila optimizer along with the adaptive filter is considered as the key novelty of this present work. The main motive of this proposed model is to filter the noisy features more possible range for improve the HA performance. Hence, the Aquila agent removes the unwanted noise by updating the adaptive processor's filter coefficient and optimizing the filter parameters such as matching error, delay, complexity, and bandwidth and power consumption. Moreover, the suggested model is designed and executed in the MATLAB platform. Finally, the performance parameters were estimated and compared with the past studies to validate the performance improvement. The presented model scored the reduced Matching error as 0.3% and reduced power consumption as 10%, which quite lower than the compared models.

175

Authors: Ahmed R. Kadhim, Furkan Rabee

Cloud computing has become an essential technology in many industries due to its scalability and cost-effectiveness. The dynamic nature of cloud computing, including elasticity, on-demand provisioning, diverse resource types, and varied pricing model, presents a significant challenge in scheduling tasks for cloud-based systems, especially when considering user quality of service (QoS) constraints such as deadlines and budgets. Therefore, in order to optimize the performance of the cloud systems and end-user satisfaction, an efficient budget and deadline-aware scheduling model is necessary. Game theory provides a framework for modeling and analyzing the strategic interactions between self-interested entities, which makes it an ideal tool for task scheduling in cloud computing. Additionally, the versatility of game models enables the analysis of various cloud computing architectures. This paper proposes a dynamic Stackelberg (leader-follower) game model for modeling the interactions between tasks, scheduler, and cloud resources to find an equilibrium for the game under both budget and deadline constraints. The proposed dynamic task scheduling based on Stackelberg game (DTSSG) model is assisted by the pricing model and satisfaction factors to select the optimal virtual machine for processing the user task. To achieve high average resource utilization, the utilization factor of the cloud resources is considered in the proposed work. Experimental results show that the Stackelberg model equilibrium has been very effective in scheduling the user tasks across the data center resources by selecting the optimal virtual machines. The results demonstrate improved execution efficiency in terms of decreased makespan by 30%, reduced number of deadline violations by 52%, decreased total gain cost by 27.13% and increased provider profit by 19.15 % on average as compared to existing deadline budget scheduling (DBS), genetic algorithm, and MAX–MIN methods. Also, the results show the effectiveness of the proposed work in terms of increased throughput by 59.4 % and decreased makespan by 27.95% using Google cloud jobs dataset (GoCJ) as compared to existing gradient-based optimization (GBO), multi-verse optimizer (MVO), enhanced multi-verse optimizer (EMVO).

201

Authors: Abbas M. Ali Al-muqarm, Naseer Ali Hussien

The unpredictable and huge data generation nowadays by smart devices from IoT and mobile crowd sensing (MCS) applications like sensors, and smartphones requires processing power and rapid responses in a specific time frame (deadline), and user cost, i.e., budget. Cloud provides these capabilities to serve organizations and customers, but when using cloud appear some limitations, the most important is task scheduling. Nevertheless, the most of earlier scheduling algorithms have focused on only one evaluation parameter such as makespan or cost and so on, as the scheduling process goal. Optimizing a single metric may not ensure an improvement in the cloud performance in general, so the need may arise for algorithms that concentrate more broadly on enhancing more than one parameter. Therefore, this paper presents a deadline-budget multi-objective dynamic scheduling scheme (DB-MODS) to execute user tasks on VMs within QoS limits in order to finish the task within the deadline and budget if possible. DB-MODS using K-Means based on task length and deadline for clustering incoming tasks into groups and categories VMs based on capacity thresholds. In addition, the task belonging to every cluster is assigned to a suitable VM in VMs groups. based on objective functions, which can be one for user request depending on (deadline and budget) and the second is for system depending on (execution time and cost). This paper used CloudSim plus to simulate and evaluate our approach. The DB-MODS approach performance was compared to scheduling algorithms from the previous literature. Both random and Google cloud jobs (GOCJ) workloads are used to evaluate the efficiency of DB-MODS. The results show that DB-MODS outperforms the other algorithms by minimizing makespan by 41%, energy by 46%, and increasing success ratio by 35%, in comparison to existing load balancing and scheduling methods: EEVS, LAS, Greedy-R, DTS, Random, and Greedy-P in the first scenario. Minimizing makespan by 46%, cost by 44%, and maximize throughput by 80% when compared with MOCS, Max-Min, HABC-LJF, FCFS, MOPSO, Q-learning, MOABCQ-FCFS, and MOABCQ-LJF in a second scenario, and minimizing makespan by 46%, and increase throughput by 39% when compared with HESGA, GA, and ACO in a third scenario in average. In addition to the possibility of applying and employing DB-MODS for scheduling deadline-critical tasks in a heterogeneous cloud system.

231

Authors: Mohamed Cherradi, Anass El Haddadi

The evolution of the huge amount of heterogeneous data sources introduces the emergence of several fresh new concepts. Among one of the most well-known concepts emerging as a recent and trending topic in big data is the data lake. Which represents a central repository that stores heterogeneous data sources in its native format without any predefined schema. In the absence of any enforced schema, effective metadata management based on metadata models remains an active research topic to address the data lake issues knowing by a data swamp. However, the examination of existing metadata models shows that none of them proposes a complete model. In this article, we propose a generic and extensible metadata model that supports high flexibility and better scalability in the integration of metadata. With these capabilities, EMEMODL enables comprehensive metadata management for data lakes. We show EMEMODL’s feasibility through a prototypical implementation based on TPC-H datasets of different sizes to prove the scaling feature, using MySQL and Neo4J. The findings of these experiments revealed encouraging results for big data queries that give us a graph database, which shows an efficient ability to manage and process large amounts of data compared to a relational database in terms of retrieval time queries and resource consumption, including cpu and memory usage.

258

Authors: Feby Artwodini Muqtadiroh, Tsuyoshi Usagawa, Riris Diana Rachmayanti, Supeno Mardi Susiki Nugroho, Eko Mulyanto Yuniarno, Mauridhi Hery Purnomo

Classification of uncertain conditions requires computational modeling to obtain exact non-vague results for making the right decision, such as opening and closing school cases during a pandemic. We cannot rely solely on normative and textual government regulations because of numerous constraints and uncertainty in implementation. Unsupervised classification techniques can deal with such issues without needing prior references that contain definitive hesitancy. This motivates us to use a fuzzy system based on knowledge-based composition rules for complex problems such as the dynamics of COVID-19 because of its ability to adapt to changes and uncertainties. Therefore, we construct rules based on knowledge about COVID-19 to the issue of opening/closing schools using three fuzzy approaches: conventional fuzzy, intuitionistic fuzzy system (IFS), and fuzzy c-means (FCM). We can demonstrate a correlation between the number of school openings and the COVID-19 dynamics by utilizing the fuzzy approach to reduce the degree of hesitance. Experiments on available public time-series datasets demonstrate that the IFS is more efficient in forming rigidly distinct two classes. The results indicate that the accuracy of IFS is 99.47%, FCM is 91.28, and conventional FS is 84.33%, including the IFS silhouette score, which is higher than the others, at 0.91 or closer to 1, indicating excellent classification results. IFS is less superior in running time, while FCM is the fastest. This is because there are multiple stages in the IFS by considering non-membership functions.

280

Authors: Mohammad Osman, Kaleem Fatima, P. Naveen Kumar

Security and privacy are the major concerns in the internet of things (IoT) which are uncertain and unpredictable. Trust aware routing is one of the recent and effective strategies which ensure better resilience for IoT nodes from different security threats. Towards such concern, this paper proposes a new strategy called independent onlooker withstanding trust aware routing (IOWTAR) for IoT. IOWTAR introduced a new compound trust metric by combining three individual metrics namely independent trust, onlooker trust, and withstanding trust (a combination of resilient trust and immovability trust). Independent trust and onlooker trust are assessed based on direct and indirect experiences of nodes about their neighbor nodes. Withstanding trust is assessed based on the stability and resilience of nodes towards dynamic topological changes and network failures respectively. Further, this work adapted the Firefly algorithm (FFA) to optimize the weights of individual trusts and establishes a secure path. Simulation experiments carried out over the proposed method had shown its superiority in terms of packet delivery ratio, delay, and throughput. The proposed method has gained an average improvement in the throughput is of 23.71%, 20.18%, 17.27%, and 2.88% from PSO, GSA, WOA, and CBBMOR-TSM-IOT methods respectively.

302

Authors: Nadia A. Khalaf, Hazim G. Daway, Baida M Ahmed

Improvements in aerial images have an important role, given the increase in dust storms due to global warming and climatic changes. Apart from improving aerial images, image enhancement is important in many applications, such as tracking and monitoring studies during wars and for the environment. However, aerial images often suffer from a lack of contrast due to dust and pollutants. In this study, we propose a perceptual dark channel prior (PDCP). In this method, the light value (Y) is improved by using stretch and perceptual color space, and then the chromatic compounds (CbCr) are improved by dark channel prior (DCP). The quality measures lightness order error (LOE), entropy, and color-fullness metrics (CM) and then compares these with the improved methods (DCPMV, YCBCR, MLA, SEA, DCP, and IEIF) to determine the method efficiency. Results show that the proposed PDCP effectively enhances the aerial images better than other methods, because owning it the average are LOE (194.096), EN (7.361), and CM (39.403).

322

Authors: Ernando Rizki Dalimunthe, Erwin Susanto, Kharisma Bani Adam

The most frequently used maximum power point tracking (MPPT) technique is the perturb & observe (P&O) algorithm as a power tracking tool in PV system. The P&O algorithm is easy to compute and implement, but the algorithm is prone to oscillations at the maximum power point. Hence, the power becomes inaccurate due to a lot of power loss. This study utilizes the deep q- network (DQN) algorithm to improve P&O performance by correcting the output value algorithm using various step sizes by DQN. The proposed method increased the tracking speed when receiving the same value at different times by 33.3%–50%, and the oscillation rate was successfully reduced by 73.99%–83.5%. The advantages of increasing tracking speed and decreasing oscillation rate are accompanied by tracked power with averages of 95%, which is better than the P&O and DQN algorithms. It shows that the proposed method can work optimally regarding efficiency and oscillation rate and be the fastest in tracking maximum power from previous related works.

345

Authors: Usman Ependi, Adian Fatchur Rochim, Adi Wibowo

This research presents a novel hybrid sampling technique, implemented at the data level, to effectively address imbalanced and noisy data in classification processes. The proposed technique expertly combines two established methods, namely, the random over sampling (ROS) and neighbourhood cleaning rule (NCL) approaches, to tackle imbalance and noise issues, respectively. The study carried out an empirical evaluation of the proposed approach using crowdsourced text data that primarily emphasized the triple bottom line (TBL) dimension of a smart social, economic, and environmental city. The study used the long short-term memory (LSTM), convolutional neural networks (CNN), and CNN-LSTM classification models to validate the efficacy of the proposed hybrid sampling technique and compare its performance with other existing approaches, including ROS oversampling, NCL undersampling, synthetic minority over sampling & tomek links (SMOTE-Tomek), and synthetic minority over-sampling and edited nearest neighbours (SMOTE-ENN) hybrid sampling. The results are impressive, with the ROS-NCL hybrid sampling technique achieving high accuracy rates across all three classification models, at 97.71%, 98.01%, and 98.11%, respectively. This approach provides a robust and effective solution for handling impure data and holds great promise in identifying complex data patterns in real-world classification problems.

378

Authors: P. Muthukumar, M. V. Ramesh, Ponnam Venkata Kishore Babu, P. Rohinikumar, S. V. Satyanarayana

Most electrical distribution systems (EDS) are radially structured so that primary and secondary protection control devices can work together well. Because they have a high R/X ratio in their design, they also have a bad voltage profile, high distribution losses, and less stability margins. As a result, integrating reactive power compensation devices like the distribution-static VAr compensator (D-SVC) can solve these issues. Yet, the situation can get worse if D-SVCs are placed and rated incorrectly in radial distribution networks (RDNs). Determining the optimum locations and their ratings simultaneously requires an updated version of the Firefly algorithm (FA) with adaptive parameters, which is introduced in this paper as the adaptive firefly algorithm (AFA). The multi-objective function that has been presented relates to improving loadability, voltage stability, and reducing active power loss. On the IEEE 69-bus, simulations are run for three different VAr compensation levels. In comparison to the base case, the losses are reduced by 34.04% and 33.14% with 50% and 75% VAr compensation, respectively. But for the optimal VAr compensation of 73.14 percent by AFA, the losses are reduced by 35.29 percent, which is higher than both under and over compensation cases. Similarly, the loadability margin is increased to 3.099 p.u. with optimal VAr compensation, but it is observed as only 2.833 p.u. and 2.939 p.u. with 50% and 75% VAr compensation, respectively. On the other hand, the findings produced with APF demonstrate its efficiency for resolving complex optimization issues and outperform those obtained with previous research. Also, the proposed D-SVCs allocation has improved RDN's overall performance, demonstrating how well it adapts to real-time applications.

405

Authors: Vivek Kadam, Surendra Bhosale

Inflated communication traffic and expanding radar applications result in frequency congestion and require extra spectrum resources. Dual-functioning multi input multi output radar benefits both sensing and communication operations via real-time cooperation. This decongests the radio frequency environment and allows a single hardware platform for both functionalities. Analog beamforming fails to provide multiplexing gain and digital beamforming is not preferred practically because of heavy power consumption. A reliable hybrid beamforming design is an inherent requirement due to the concerns related to limited RF chains, interference and the cost of fully-digital beamforming. In this paper, Enhanced-Social Ski Driver’s algorithm is presented to investigate the performance of hybrid beamforming in dual-functioning MIMO radar. This study investigates location-based direction sensing for Rayleigh and Rician fading channels. This amendment provides improved performance with a minimal bit error rate, maximum gain, and significant normalized array power distribution. The simulation results show 24% reduction in bit error rate, 15% improvement in gain and persistent array power than existing social ski driver’s algorithm.

428

Authors: Asif Abdullah, Omkar S Powar, Krishnan Chemmangat

Surface electromyography (sEMG) based silent speech interface (SSI) is an actively investigated topic among the broad area of human computer interaction studies which is currently dominated by acoustic sound based speech recognition research. This research is an attempt to help people who have an impaired vocal system if they are having no issues with their facial muscle functions. The basic idea is to reduce the total number of sEMG electrodes that has to be affixed on the face thereby reducing the invasiveness of the silent speech recognition module. This is achieved by incorporating a new detrended fluctuation analysis (DFA) based feature along with the already existing features associated with electromyographic signals. DFA is used for the first time in literature in the area of surface electromyography based silent speech recognition. The main idea is to incorporate the DFA feature along with the state-of-the-art features to improve the performance of a sEMG based SSI model so that an efficient channel reduced model can be realised. Different channel combinations were tried to analyse the impact of each channel in word recognition accuracy and the optimal channel combination was identified. As a result of this research work, a reduced channel setup with 5 electrodes was proposed in place of the conventional 7 channel data acquisition setup. This was achieved while maintaining an accuracy of 83.88 % and 92.92 % using the decision tree (DT) model and K-nearest neighbours (KNN) model respectively.

455

Authors: Md. Irshad Hussain B., Mohamed Rafi

In India, online digital transaction has become prominent means of financial transaction in everyone’s day-to-day life as per reserve bank of India's (RBI) reports. Presently, Indian banks protect users' online transactions with only username and password, CAPTCHA, and dynamic 6 or 8-digit one time passwords (OTPs), which are not very secure. According to literature, this straightforward method of authentication is easily vulnerable to numerous threats. The time offered to enter the OTP with numerical pattern is more than 100 seconds, which is in defiance of RBI guidelines and can be used by fraudsters to commit fraud. In this paper we are suggesting a novel 3-Level authentication which is in-line with RBI authentication guidelines along with the authenticator assurance standards of National Institute of Standards and Technology of the United States and the European regulation for electronic payment services, for securing the users online transactions aided by username and password, OTPs and biometrics. At Level-3 user is verified with biometrics or by OTP. For biometric authentication, we have chosen facial trait, a hybrid algorithm for face detection and recognition is designed, implemented and tested on a standard dataset of VTU-BEC-DB. Suitable OTP and time constraints were identified with the implementation of proposed algorithm being tested by conducting series of trials involving 200 participants with age group of 22 to 55 years. Based on online-survey, we are suggesting for the use of fall-back authentication technique powered by security question, in-case the authentication is only on the basis of OTP. We have achieved 98.20% recognition success rate for facial recognition. The significant time to enter dynamic 6 digit OTP and dynamic 8 character alpha-numeric OTP with time limits of 45 and 60 at Level 2 and 3 is 33.821.10 and 45.310.92, respectively.

482

Authors: Purba Daru Kusuma, Ashri Dinimaharawati

A new metaheuristic can be developed by constructing from scratch, modifying the existing metaheuristics, or hybridizing some metaheuristics. This work presents a new metaheuristic: extended stochastic coati optimizer (ESCO). ESCO is developed by expanding the shortcoming coati optimization algorithm (COA). ESCO expands the number of searches and references used in COA. ESCO also implements a stochastic process for each unit to choose the searches that will perform. It differs from COA, which splits the population into two fixed groups, each performing its strategy. ESCO implements three sequential phases in every iteration. Two options can be chosen in every phase. ESCO has three references in its guided search: the global best unit, a randomly selected unit, and a randomized unit within the search space. In this work, ESCO is challenged to solve 23 classic functions and benchmarked with five shortcoming metaheuristics: guided pelican algorithm (GPA), puzzle optimization algorithm (POA), average subtraction-based optimizer (ASBO), and coati optimization algorithm (COA). The result presents the superiority of ESCO among five shortcoming metaheuristics by outperforming the GPA, POA, GSO, ASBO, and COA in solving 13, 21, 23, 16, and 13 functions, respectively. Through investigation, the multiple search approach is more effective than the single search approach.

509

Authors: Seema Swamy Gowda, Ambika Ramalingappa

Underwater wireless sensor networks (UWSNs) have a huge amount of sensors located underwater to collect data from the underwater scenario. UWSN is considered a promising method for monitoring and exploring an underwater scenario. Energy-efficient and reliable data broadcasting are considered challenging tasks, because of the limited energy source of sensors. To address this issue, an energy-efficient cluster head (CH) selection and multi-hop routing are developed in UWSN. The multi objective energy based improved jellyfish swarm optimization (MOEIJSO) is proposed to select optimal CHs from normal sensors. The input parameters of MOEIJSO are residual energy, neighbor node distance, sink distance, and CH balancing factor. Next, multi-hop routing is developed by using ant colony optimization (ACO) for delivering the data packets. Therefore, the proposed MOEIJSO-ACO method is used to improve energy efficiency while increasing data delivery. The proposed MOEIJSO-ACO method is evaluated by using the alive and dead nodes, total energy consumption, data received in base station (BS), and life expectancy. The existing researches such as metaheuristics-based clustering with a routing (MCR) for UWSN, modified low energy adaptive clustering hierarchy (MLEACH) and cooperative energy-efficient routing (CEER) approach are used to compare the MOEIJSO-ACO method. The alive nodes of the MOEIJSO-ACO are 399 for 800 rounds, which is high when compared to the MCR-UWSN.

532

Authors: Shalini Sharma, Syed Zeeshan Hussain

A mobile Ad Hoc network (MANET) is a dynamic wireless network developed using wireless nodes without using any infrastructures. The wireless nature of mobile networks causes several challenges in secured communication which makes it susceptible to malicious attacks. Therefore, security is a very essential factor in facilitating a secured message transmission among mobile nodes in the wireless medium. The energy and trust aware - multi objective african vultures optimization (ETA-MAVO) method is used in this paper to provide high security for the data packets, thereby preventing malicious attacks. Using the ETA-MAVO algorithm, the network nodes' energy efficiency and trust value are calculated. For choosing the best cluster heads (CHs), ETA-MAVO considers fitness functions, namely: trust value, energy ratio, communication cost, the total number of gateway hops, and network load. To discover malicious nodes in the network and minimize packet loss, the trust value of a node is calculated from its neighbours. Because of trust value, updating routing information is simpler and increases network throughput. The results obtained from the simulation show that the proposed ETA-MAVO achieved better performance with respect to parameters such as delay, energy consumption, throughput, and detection rate when compared with the existing cat slap single-play algorithm (C-SSA) and data-driven zone based routing protocol (DD-ZRP), angular based energy proficient trusted routing (ACEPTR) protocol and aquila optimizer (AO) and puzzle optimization algorithm (POA). The proposed method achieved a detection rate of 93% while the existing C-SSA, DD-ZRP, ACEPTR, AO and POA achieved a detection rate of 90%, 68%, 72%, 91% and 90% respectively.

552

Authors: R. Sridevi, V. Sinthu Janita Prakash

Clustering is one of the important functionalities of networking, which has a dominant impact on the performance in terms of Throughput, packet delivery ratio and communication delays. The challenge in performing the clustering process is more challenging for heterogeneous networks. The divergence between the member nodes is highly variable for a network that contains IoT nodes. As the result of vast availability of different IoT hardware, organizing them under one roof is a critical task which is very vital too. This work is intended to develop a new hierarchical based clustering to improve the performance and the lifetime of the IoT / Wireless sensor network nodes. Three legacy functional blocks such as idiosyncratic fuzzy energy estimator, cumulative fuzzy energy hierarchy builder and energy based cluster head exchanger are integrated in this work and submitted here as “Idiosyncratic Fuzzy Low-Energy Adaptive Clustering Hierarchy (IFLEACH)”. An industrial standard network simulator is used to evaluate the performance of IFLEACH in terms of throughput, packet delivery rate, latency, end-to-end delay and energy consumption by simulating the real-worlds IoT network environment replica. An achievement around 11 kbps throughput and 4 % security level are obtained by the proposed method in high density network.

579

Authors: Meeradevi, Monica R. Mundada

Agriculture is been majorly contributing to the country’s GDP. So, it is foremost important to improve the production quality and quantity to improve the agricultural economy. Using modern technologies like IoT, big data, blockchain, robotics, 5G technologies, etc., farming can be improved by replacing humans in agricultural operations. The proposed work focuses on forecasting of yield and prices of agro-products which will be useful for farmers to increase their productivity and which in-turn increase their economy. Data science techniques can be used to perform this by predicting which crop can be grown for given environmental features and what is the outcome before harvesting period. The proposed model uses computational data-driven approach for crop yield prediction and price forecasting of agro-products. The study uses hybrid approach for crop yield prediction which integrates LSTM and genetic algorithm based evolutionary algorithm known as enhanced long short-term memory (ELSTM). The weight assignment plays a major role in learning the behaviour of the network. The gradient value with respect to weight is calculated for each epoch, and weight is updated to compute the new weight of the network to minimize error. The model accurately predicts the crop yield for the dataset which comprises of features like soil data, rainfall, history of production, fertilizers etc., the accuracy of the model is measured using metric root mean square error (RMSE) and mean absolute error (MAE). In the proposed study, the computation model is used to enhance the knowledge about agricultural yield before the crop sowing period. The proposed model will help farmers and government agencies to improve production. The proposed ELSTM model is compared with other machine learning models such as naïve bayes, decision tree, which shows accuracy between 75% to 80% and the experimental result show the proposed ELSTM model is performing better with 85% accuracy. The price forecasting model is necessary to identify the commodity prices well in advance. The agro-product price forecasting is made using auto-regressive integrated moving average (ARIMA) which uses dataset collected from various agricultural produce market committee (APMC). Auto correlation function (ACF) and partial auto correlation function (PACF) time series plots have been used to assess the proposed model's performance in predicting prices of crops.

604

Authors: Nelli Sreevidya, Penamakuru Siva Jyothi, Gaddam Sumalatha, Shanthi Pannala

Breast cancer is considered one of the main reasons for the mortality of females around the world, therefore the recognition and classification of the initial stage of breast cancer are essential to help the patients to perform the required action. However, the precise classification of breast cancer is a difficult task in medical imaging because of the complexity of breast tissues. In this paper, Otsu's thresholding based segmentation (OTS) is used for segmenting the cancer regions from the images of mammographic image analysis society (MIAS) dataset. Subsequently, feature extraction using Alexnet is done for extracting the features. The feature selection using multi-objective improved ant colony optimization (MIACO) is proposed in this paper to select the optimal features from the overall features acquired using Alexnet. The MIACO is mainly selects the features based on correlation coefficient to choose the features with higher correlation. Further, the stacked autoencoder (SAE) is used to perform an effective breast cancer classification using the features selected by MIACO. The performance of MIACO-SAE method is analyzed by means of accuracy, precision, recall, characteristic stability index (CSI) and F-measure. The existing research namely threshold-based and trainable segmentation (TTS) with machine learning (ML), infinite feature selection with genetic algorithm (IFS-GA) and deep convolutional neural networks (DCNN) are used to evaluate the MIACO-SAE method. The classification accuracy of MIACO-SAE is 99.36%, which is high when compared to the TTS-ML, IFS-GA and DCNN.

625

Authors: Mahmoud Zaki Iskandarani

The objective is to investigate effect of slot probability and its effectiveness in vehicular communication. The work carries out MATLAB simulation covering three different levels of slot probability; low, medium, and high. The goal behind such simulation is to uncover the process that leads to high reduction in throughput. The normal explanation for the reduction is high levels of collisions and re-transmission of data through new attempts. However, this work shows that such process can be characterized and controlled as the work shows through sinusoidal regression, the oscillatory behavior caused by collisions, attempts, retransmission, as a function of slot probability. The simulation results, which covers 20 vehicles at various slot probability levels, shows that there is a change in the data traffic-throughput shape function. This is due to the effect of slot probability and reflects on the level of obtained oscillations. The work also discovered an oscillation pattern that infiltrate and affect communication channels at different levels as a function of collisions and data delivery attempts, which is related to slot probability and data traffic. Such oscillatory pattern should be investigated further and the conditions for such pattern should be eliminated, in order to achieve better utilization of communication channel and higher throughput, with lower attempts and less collisions. This is also related to both slot probability and number of available slots. The obtained results in this work showed that at the upper part of the mid slot probability range and all the way to through the high slot probability part, a noticeable oscillatory behaviour is observed, with almost constant values of average data traffic and throughput values. These effects are accompanied by large acceleration and accumulation of collisions. Thus such new approach, and in conjunction with NP-CSMA, can be used at both the design stage of communication channels and vehicular networks, and during operations. This will enable adaptive slot allocation and dynamic transmission of messages, which will result in collisions reduction and better channel utilization.

654

Authors: Ali H. Mhmood, Hazem I. Ali, Amer B. Rakan, Mohammed R. Subhi, Yaseen Kh. Yaseen, Hameed A. Mohammed

This study proposes a new approach for controlling COVID-19 through vaccination, where an adequately descriptive mathematical model is created for COVID-19 using the susceptible exposed infectious recovered (SEIR) model of epidemic diseases. The presented control approach is synthesized using a combination of feedback linearization and H_∞ control, and incorporates model reference control to achieve optimal time responses with the aid of the black hole optimization (BHO) algorithm. The effectiveness of the designed control law is evaluated using data from the lombardy region of Italy. The results of the simulation show that the proposed control approach is able to effectively control the COVID-19 outbreak by accurately implementing the desirable vaccination, while effectively addressing nonlinearity and uncertainty in the COVID-19 system with a desirable control action. The control method has achieved the required immunity of 6.6 million individuals after approximately 25 days with a transmission rate reduced to zero in a short time, and a vaccination rate of 170 thousand people per day.

677

Authors: Karnam Sreenu, Sreelatha Malempati

From the past few decades, the cloud computing has been arisen as an extensively used platform to provide storage, compute and analytics services to organizations and end users on the basis of pay-as-you-use. This enabled the organizations as well as individuals to access the larger set of resources without establishing a costly and high performance computing platform. However, the major issue in cloud computing is the task scheduling which is declared as NP-hard problem and most of the researchers applied meta-heuristic algorithms to solve it. However, they experienced a slower convergence speed which has a direct impact on the efficiency of cloud computing environment. To achieve a faster convergence along with efficient quality of Service, this paper proposes a simple and effective task scheduling mechanism based on multiple resource attributes and conditional logic. This method considers totally four resource attributes such as Resource reaction time resource location, resource availability and resource reliability rate. Based on these four attributes, the proposed method constructs an index called as task scheduling index (TSI) and assigns tasks for resources based on their TSI value. The TSI is constructed through the proposed conditional logic and high priority is given for resources those have higher TSI. For experimental validation, we used two benchmark datasets such as GOCJ and synthetic dataset. Three performance metrics namely Makespan, Throughput and convergence speed are measured and compared with state-of-the-art methods like PSO, GA and GWO. On an average, the reduced Makespan of proposed method is 48.22% and 46.87% for GOCJ and Synthetic datasets respectively.

709

Authors: Muntasser A. Wahsh, Zainab M. Hussain

This paper presents a novel method for optimizing rectangular boundaries in stitched images using a combination of deep learning and genetic algorithms. The proposed method utilizes an enhanced structure of convolutional neural network (CNN) in order to extract features for residual regression and a genetic algorithm to tune CNN parameters. The performance of the proposed method was evaluated using various metrics, including mean squared error, peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and Fréchet inception distance (FID). The outcomes from the conducted experiment demonstrated that the suggested method achieved high accuracy in rectangular boundary optimization, even for images with complex shapes and orientations. The method was also tested with different parameters, and results indicated that increasing the number of generations and population size led to improved performance. Also, the combination of deep learning algorithm specifically (CNN) and optimization algorithm specifically (genetic algorithm (GA)) led to an increase in the accuracy of the rectified images where the average PSNR reached 23.98, SSIM is 0.8066, and FID is 18.72. The proposed method has potential applications in various areas of study, such as image cropping, object detection, and image segmentation.

14

Authors: Mantena Krishna Satya Varma, Raja Kulasekaran, Nynalasetti Konadala Kameswara Rao

Hyperspectral image (HSI) segmentation and classification is trending research in military and civil applications area. However, HSI classification is facing various challenges in analyzing spectral and spatial regions. In order to improve the performance of HSI classification models, segmentation is essential step. Therefore, this article is focused on implementation of unified HSI segmentation network (HSIS-Net) using active learning. Initially, HSI preprocessing operation is performed to normalize the spectral-spatial regions. Then, joint spatial-spectral boundary extraction operation is performed using spatial information divergence (SID) and spectral correlation mapper (SCM). Finally, segmentation of boundary estimated HSI bands is performed using multi-view active learning network based fully convolutional segmentation network (MAL-FCSN). The simulations revealed that the proposed HSIS-Net resulted in superior segmentation performance with segmentation accuracy (SA) of 0.999, and segmentation F1-score of 0.999 as compared to the existing HSI classification approaches for four publicly available HSI datasets.

35

Authors: Duaa Al-Safaar, Wathiq Laftah Al-Yaseen

Network violations are currently society's major challenge. For networks to be protected against hostile threats, an intrusion detection system (IDS) is important. To create effective IDS, deep learning (DL) is used in various fields, including information security. In this paper, a hybrid deep learning approach is proposed to effectively identify network intrusions using Autoencoder (AE) and Multi-layer perceptron (MLP). We use Autoencoder which can reduce the number of the original attributes based on the number of attributes, we first enter the original data on the Autoencoder and produce new compressed data, then enter it on the MLP classifier. The NSL-KDD dataset is thoroughly evaluated to determine the efficacy of the hybrid AE-MLP model the best outcomes are reached, with an accuracy rate of 87.6% and 81.06% (binary classification and multi-classification). In addition, the proposed hybrid method was compared with various recently proposed DL-based attack detection mechanisms. In terms of performance on the available dataset, it is observed that the proposed model outperformed.

64

Authors: Gouri Bompem, Nagaraju Chiluka, Dhanalakshmi Pandluri

In this research article, an effective model is implemented for predicting the outcome of the Indian general election 2019 and farmer’s protest by utilizing the sentiment analysis of Twitter data. In the initial segment, the raw Twitter data are acquired from the Indian Political tweets 2019 and farmers protest tweets databases. Further, the denoising operations such as removal of unnecessary space, tabs, new-line, hashtag symbols, non-English characters, punctuations, numbers and special characters are used to enhance the quality of the acquired data. Then, the keyword trend analysis and topic modeling utilizing latent dirichlet allocations (LDAs) are performed for better data representation. Next, the extraction of the feature is carried out utilizing skip-gram and term frequency-document level frequency (TF-IDF) techniques, and further, the feature optimization is accomplished using enhanced dragonfly optimization algorithm (EDOA) for selecting the optimal feature vectors. In the EDOA, a Brownian motion is added for improving the probabilistic behaviors. Lastly, the selected features are given to the deep belief network (DBN) model to classify the people’s sentiments into negative, neutral, and positive classes. The experimental evaluations demonstrated that the EDOA-DBN model has obtained 99.22% and 98.83% of accuracy on both Indian political tweets 2019 and farmers protest tweets databases, which are maximum related to other existing models.

92

Authors: Nadia A. Khalaf, Hazim G. Daway, Baida M Ahmed

The use of hazy images is amongst the fastest ways to identify natural resources. Aerial photography has overcome the obstacles faced in land areas, such as rugged areas, swamps and sand dunes. It has important applications in tracking and monitoring studies during wars and peace. However, aerial images often suffer from the lack of contrast due to dust and pollutants. Therefore, this study attempts to improve aerial images by modified dark channel prior (DCP) and adaptive histogram equalisation (AHE) algorithms with YIQ. The images are first improved through DCP and then converted into YIQ colour space that saves the chromatic compounds (IQ). The lighting component comes out after improving the original aerial image via the stretch technique and AHE. Afterwards, all the compounds are merged, and reverse transformation is performed. The proposed method for improving aerial photographs was compared with several other methods. By analyzing the results, the proposed method obtained the best measures of quality, including an entropy value of (7.771), an average gradient value of (14.207) and a naturalness image quality evaluator value of (3.700).

115

Authors: Waleed Kareem. Ahmed, Rana Saad Mohammed

The Internet of Things (IoT) was a new concept that connected various physical things to the internet. The IoT is fast expanding and will soon significantly impact everyday life. Although the increasing number of connected IoT devices is simplifying life, they risk our data. Radiofrequency identification (RFID) assists in the automated identification of connected IoT devices. Furthermore, the primary challenges for RFID tag-connected technology are privacy and security. With the improving safety of solutions RFID for many apps, RFID necessitates a centralized database widening. Blockchain techniques are speedily establishing themselves together as a new distributed and decentralized replacement. And it provides enhanced transparency, data security, immutability, dependability, and lower maintenance costs. Because of its inherent benefits, RFID is projected to play a significant part in enabling identifying technology on the IoT. However, due to its association with sensor technology, it has the potential to be applied in a wide range of industries. Security, however, is one of the more complex components of creating an RFID system. RFID Security focuses around authentication and privacy concerns. From the previous works, we found that the computing cost and uptime are higher. To solve these problems, we offer improving lightweight authentication using Arnold Chaotic map and Markov chain for IoT applications. The proposed method uses random algorithms (Markova chain, Arnold map chaotic, and ECC), which add high security. We also used blockchain technology to transfer keys through a non-security channel. Based on parameters such as (communication cost, throughput rate, storage requirements, transmission delay, and computational cost) the results of the proposed RFID protocol were compared with several RFID-based security solutions. As a result, the proposed lightweight scheme is more efficient and secure than existing RFID systems in terms of performance, according to simulation results, and is fully compatible with actual applications.

142

Authors: A Sequence-to-Sequence Text Summarization Using Long Short-Term Memory Based Neural Approach

The massive development in the different fields of engineering and science brought a significant change in the modern years. In the field of natural language processing, automatic text summarization is known as one of the important research directions. The text's primary concepts and flow should be considered to provide a solid summary that limits the repetition in the text as a summary. A sentence is referred to as a basic language unit and there are many semantic links present among the prison terms like example-of, cause-effect, sequential, etc. within a meaningful text. In this research manuscript, an automated text summarization model is developed and its performance is validated on the two challenging datasets such as daily-mail and Gigaword. In this manuscript, a transformer: long short term memory (LSTM) neural network is used along with the Huber loss function and Adam optimizer. The transformer used in this research (LSTM with Huber loss function and Adam optimizer) includes the advantages like requiring limited memory, computationally effective, and easy implementation. The proposed neural network mainly aims in exhibiting the summaries that are composed by the groups or paragraphs that includes more keywords or phrases than summaries composed by sentences. The non-differential evaluation metrics are utilized in the proposed sequence-to-sequence model to provide semantic information of input text and it stores important features for effective text summarization. The size of the parameters is specified by the maximum number of sentences in the same group. The proposed LSTM with Huber loss function and Adam optimizer has 44.51 ROUGE-1, 20.43 ROUGE-2, and 40.08 ROUGE-L on dailymail dataset. Experimental results of the trial demonstrated that the proposed model outperformed the existing models like dynamic residual network, convolutional neural network with LSTM, and knowledge powered topic level attention model in terms of the rouge parameter.

166

Authors: Kelly Rossa Sungkono, Riyanarto Sarno, Maisie Chiara Salsabila, Chintya Prema Dewi

A process model describes business process flow as the activities that employees must carry out. Nowadays, many companies have similar business processes, so they do not establish their process model from scratch but build the model based on an existing process model or a combination of some process models. Several process mining methods approaches matching rules to define similarities of a model, and others consider the semantic side; however, none use the similarity to merge some business process models. This paper proposed graph-based semantic similarity, a method that merges two process models considering the semantic similarity between those activities. The utilized semantic similarity methods are SBERT and TF-IDF. The evaluations compare SBERT and TF-IDF with other methods and use a similarity method with the highest score in graph-based semantic similarity. Based on the semantic similarity score, graph-based semantic similarity with SBERT has higher similarity scores than existing graph-based semantic similarity, i.e., node similarity and Jaro-Winkler distance. With the highest similarity scores among existing methods, the evaluations also prove that graph-based semantic similarity with SBERT correctly combines business process models based on semantic similarity.

192

Authors: Zahraa Ch. Oleiwi, Ebtesam N. AlShemmary, Salam Al-augby

Arrhythmia is one of the cardiovascular disease types that affect humans and often leads to death. generally, ECG signals uses to diagnose the patient’s heart state where the ECG illustrates the electrical activities and physiological state of the heart. This paper proposes ECG classification model to classify four types of heartbeats for the early detection of Arrhythmia. Detail wavelet coefficients of ECG were extracted using discrete wavelet transform (DWT) to produce new datasets of ECG with for the dimensions and processed information to ensure the efficiency of proposed classification techniques. In addition, power spectral density (PSD) has been calculated for approximate wavelet coefficients of ECG to extract more relevant features that improve the performance of the classifier. The two classifier models use, the convolution neural network (CNN) utilizes for deep learning networks with artificial neural network (ANN) and the Random forest ensemble method. The experiments and results show that, the proposed model with RF archives 98.5% classification accuracy considering all decomposition levels of DWT. Additionally, the solution with CNN-ANN achieves 96% classification accuracy at the third decomposition level. Therefore, the results show the impact of the proposed solution with high efficiency in terms of fewer dimensions and high accuracy.

215

Authors: Jaafar Ahmed Abdulsaheb, Dheyaa Jasim Kadhim1

This work considers the best path planning algorithm for a mobile robot that travels independently in an unknowable environment. To get around the limitations of unstable searches in the conventional Hunter-Prey Optimization Algorithm (HPO), the improved HPO optimization algorithm is used and introduces a new control parameter referred to as randomization adjustment in order to avoid stagnation and early convergence. The absence of the transfer parameter from exploration and exploitation is a significant flaw in the HPO algorithm, which results in an unstable search and additional time waste. Another new parameter called the changing parameter (CP) is used to address this flaw. It is used in an environment with erratic static and dynamic obstacles and a static and dynamic target. Finding a collision-free path that is also the objectively shortest path and the smoothest path can solve the path-planning problem. The proposed algorithm attempts to mimic the real world by taking into account the actual size of the mobile robot, a kinematic model, and the robot's specifications. The proposed algorithm is evaluated by comparing it on 30 dimensions using 13 benchmark test functions. The performance of the proposed algorithm is evaluated against the results of five swarm optimization algorithms. According to the results of the standard deviation tests, the proposed algorithm performs the best in 92% of the 13 test functions. Furthermore, the average outcomes for three complex maps (10×10) m in size demonstrate the potency of this approach for robot paths from the starting point to the target. The average distance over ten runs for maps 1, 2, and 3 is 12.6436 meters, 12.4961 meters, and 17.6547 meters, respectively. It demonstrated how quickly and easily it could avoid both stationary and moving obstacles.

240

Authors: Mazin Ismael Al-Saedi

In this paper, a modified block backstepping controller was formulated for double lane changing control of a two-degree-of-freedom (2DOF) autonomous vehicle. For a systematic procedure, a 5th-degree polynomial double lane change trajectory was created. Next, the state model dynamics of this underactuated 2DOF autonomous vehicle is presented, which was transformed into the block-strict feedback form. Thereafter, the block backstepping technique was implemented to achieve the control input (steering angle). For enhancing the steady state performance during the trajectory tracking, an integral action was included in the proposed controller. The stability of this controller was established through the Lyapunov theory. To investigate the effectiveness of this proposed controller, we compared it with a human driver's transfer function controller and LQR controller, in addition to the sliding mode controller. Simulation results were achieved using CarSim-Simulink. Later, this controller was implemented to track the same lane change path with a different road coefficient of friction. The supposed controller characteristic was further investigated by inspecting the motion of the vehicle when exposed to an idealized lateral step force disturbance. The simulation results indicated the power of modified block backstepping control and it has higher following accuracy, better yaw rate tracking, and stable front steering angle, at high speed and at high and low friction roads.

268

Authors: Purba Daru Kusuma, Meta Kallista

This paper introduced a novel metaheuristic that is developed based on the tournament mechanism, namely quad tournament optimizer (QTO). As its name suggests, QTO proposes a new approach of metaheuristic in which there are four searches conducted by each agent in every iteration. These searches are: (1) searching toward the global best solution, (2) searching toward the middle between the global best solution and a randomly selected solution, (3) searching relative to a randomly selected solution, and (4) neighbourhood search around the corresponding solution and the global best solution. A solution candidate is generated by each search. Then, a tournament is carried out to find the best candidate. This strategy is novel because most of metaheuristic deploys only single search or multiple searches where each search is conducted sequentially. QTO is challenged to find the optimal solution of 23 classic functions. In this challenge, QTO is benchmarked against five shortcoming metaheuristics: marine predator algorithm (MPA), slime mould algorithm (SMA), golden search optimizer (GSO), hybrid pelican Komodo algorithm (HPKA), and guided pelican algorithm (GPA). The result indicates that QTO outperforms all these benchmark metaheuristics. QTO is better than MPA, SMA, GSO, HPKA, and GPA in solving 22, 19, 21, 20, and 13 functions consecutively. The result also indicates that QTO has successfully found the global optimal solution for nine functions: Sphere, Schwefel 2.22, Schwefel 1.2, Schwefel 2.21, Rastrigin, Griewank, Six Hump Camel, Branin, and Goldstein-Price.

291

Authors: Azza. R. Abd El-maksoud, A. H. Ahmed, S. Hasan

In this paper, for four different types of distributed generation (DG) units, a graph theory-based modified particle swarm optimization (MPSO) algorithm is proposed for the simultaneous optimal distribution network reconfiguration (DNR) and allocation (placement and sizing) of multiple DG units to minimize the total real power loss of the radial distribution network (RDN) while fulfilling all system operational constraints. Different cases are carried out for DNR and DG integration to evaluate the effectiveness of the proposed algorithm. To evaluate the objective function, an intelligent graph theory-based backward/forward (BW/FW) sweep load flow technique that can manage any topological alteration owing to DNR and DG integration is introduced. The proposed algorithm is assessed using IEEE 33-bus and IEEE 69-bus radial distribution systems and results are obtained using MATLAB software. The obtained simulation results show that the proposed algorithm can provide a wonderful solution in terms of real power loss minimization and voltage profile enhancement (compared with the base case, loss reduction of 74.29%, 92.15%, 54.3%, and 45.87% for DG type1, DG type2, DG type3 and DG type4, respectively for IEEE 33-bus, and loss reduction of 84.16%, 97.79%, 70.62%, and 63.01% for DG type1, DG type2, DG type3, and DG type4 respectively for IEEE 69-bus).

319

Authors: Nada Fitrieyatul Hikmah, Rachmad Setiawan, Mohammad Daffa Gunawan

Poor sleep quality is a common sign of a variety of sleep and health problems. Thus, sleep quality assessment is necessary as it can be a first-step predictor of physical and mental health. Several studies were completed for this objective. However, no prior study in sleep quality assessment has explored a comprehensive heart rate variability (HRV) analysis by including feature extraction in the time and frequency domain, and non-linear analysis. This study proposed a full evaluation of sleep quality, by incorporating multiple physiological signs in subjects to detect exhaustion after a period of sleep. The primary contribution was the development of an algorithm to estimate sleep quality based on the combination of electrocardiography (ECG) and electromyography (EMG) signals by using machine learning. HRV analysis of ECG signal included time domain, frequency domain, and non-linear analysis. Mean power frequency (MPF) was extracted from the EMG signal by using spectral analysis. In addition, determination of fatigue level as an indicator of the subject's sleep quality was evaluated with fatigue severity scale (FSS) questionnaire. Based on results, the accuracy values of logistic regression, random forest, k-nearest neighbor (k-NN), support vector machine (SVM), and SVM with grid-search were 93%, 93%, 93%, 40%, and 100%, respectively. The proposed method was applicable for investigating sleep quality.

343

Authors: Fatima Abdul Sattar AL-Taie, Ahmed Sabah Al-Araji

This paper presents a new development of a predictive voltage neural controller to control the stack terminal output voltage of a nonlinear proton exchange membrane fuel cell (PEMFC) system based on a neural network technique and a back-propagation learning algorithm. The main objective of this paper is to precisely and quickly identify the best control action of the hydrogen partial pressure to enhance the nonlinear performance of the fuel cell output voltage under a variable load current. This optimal control action prevents damage to the fuel cell membrane, thereby prolonging the fuel cell’s lifetime. The proposed predictive voltage controller consists of three sub-controllers. The first one is the numerical feed-forward controller (NFFC), which is used to decide the steady-state hydrogen partial pressure (PH2) control action depending on the desired voltage. The second sub-controller is a feedback neural controller that uses a multi-layer perceptron (MLP) and a back-propagation learning algorithm to generate the hydrogen partial pressure feedback control action to track the desired output voltage of the fuel cell during transient conditions. The third sub-controller is the predictive control law equation, which is based on the modified Elman recurrent neural network (MERNN) as an identifier for the PEMFC model and the multi-objective performance index. From the simulation results, the proposed controller, which is composed of the three sub-controllers, has the capability to generate a precisely and quickly timed response to the hydrogen partial pressure control action in order to minimize the tracking voltage error and eliminate oscillation in the output voltage of the fuel cell. Finally, the suggested predictive voltage control strategy's numerical simulation results are then verified by comparison with those of other types of controllers in terms of the minimum number of steps ahead prediction (reducing from 10 to 1 step ahead prediction) and enhancement of the tracking voltage error by 81.8% when comparing with a predictive neural controller and improvement of the tracking voltage error by 87.5% when comparing with an inverse neural controller. Moreover, the oscillation effect in the output voltage is completely eliminated, resulting in a response without any overshoot.

375

Authors: Suneetha Mopidevi, Prasad M.V.D, Venkata Vijay Kishore Polurie, Anil Kumar Dande

Multiview video processing for recognition is a hard problem if the subject is in continuous motion. Especially the problem becomes even tougher when the subject in question is a human being and the actions to be recognized from the video data are sign language. Although many deep learning models have been successfully applied for sign language recognition (SLR), very few have considered multiple views in their training set. In this work, we propose to apply meta metric learning for video-based sign language recognition. Contrasting to traditional metric learning where the triplet loss is constructed on the sample-based distances, the meta metric learns on the set-based distances. Consequently, we construct meta cells on the entire multiview dataset and perform a task-based learning approach with respect to support cells and query sets. Additionally, we propose a maximum view pooled distance on sub-tasks for binding intraclass views. The results of experiments conducted on the multiview sign language dataset and four action datasets show that the proposed multiview meta metric learning model (MVDMML) achieves 11% higher performance than the baselines.

401

Authors: Dian Candra Rini Novitasari, Fatmawati, Rimuljo Hendradi, Yuniar Farida, Ricky Eka Putra, Rinda Nariswari, Rizal Amegia Saputra, Rr Diah Nugraheni Setyowati

Diabetic retinopathy (DR) initially derives from damage to the eye blood vessels which leads to bleeding up-to permanent blindness. The severity of DR is not easily known. Therefore. it is necessary to create a system that is able to identify the severity level of DR. In this study, the identification of DR was conducted using hybrid CNN and ELM method. Hybrid CNN-ELM is useful for obtaining the most effective model in the classification system and computational time. CNN architecture is useful for extracting fundus data in image feature recognition. Several modified ELM methods (KELM, MLELM, DELM) were used to classify the severity of DR based on the results of CNN feature extraction. The classification system was tested with two datasets, namely DRIVE and Messidor. Based on the average value, the best architecture evaluation in extracting fundus data was DenseNet compared to GoogleNet, ResNet18, ResNet50, and ResNet101. Based on the computation time, the KELM method was faster than the MLELM and DELM methods, with an average time of 43 seconds. The results on the DRIVE dataset produced good evaluation values, while Messidor obtained good results on the MLELM method. It showed that the Messidor data was able to be separated well linearly. The modified CNN-MLELM method produced more stable values in both datasets with an average accuracy of 99.21%, a sensitivity of 99.29%, and a specificity of 99.21%.

427

Authors: Abbas A. Kareem, Bashra K. Oleiwi, Mohamed J. Mohamed

Path planning is one of the most crucial aspects of implementing unmanned aerial vehicle (UAV) missions. Therefore, it is essential to figure out the optimal path from the starting point to the target point in different scenarios. In this paper, we propose employing a hybrid algorithm, named FPA-GA, generated by combining the flower pollination algorithm (FPA) and genetic algorithm (GA), to find an optimum path in a real modern building’s environment. In addition, a cubic polynomial algorithm via two points was proposed to make the route adequate and smooth. Because the GA has a good capability for exploring the search space, it is employed for exploration while the FPA and GA are employed to increase the exploitation capability in the proposed algorithm. Five different scenarios are utilized to evaluate FPA-GA's ability to find the optimal path in a variety of situations, and then the proposed algorithm's performance is compared with that of seven other algorithms: GA, FPA, bat algorithm (BA), particle swarm optimization (PSO), whale optimization algorithm (WOA), improved whale optimization algorithm (IWOA), and IWOA-PSO. The best path, the mean path length, the standard derivation, and the worst path length are the four parameters used to compress data. In all scenarios, the results demonstrate that FPA-GA is capable of locating the shortest and most collision-free paths, and the hybrid algorithm FPA-GA is always superior to other algorithms. The proposed algorithm provided the mean path length enhancement in all scenarios, where the maximum enhancement equals (33.6%), (23.1%), (55.5%), (29.8%), (50.5%), (53.9%), and (26.4%) compared with GA, FPA, BA, PSO, WOA, IWOA, and IWOA-PSO, respectively. On the other hand, the proposed algorithm is guaranteed to find the best path in all scenarios, where the standard deviation of FPA-GA is always less than that of other algorithms.

454

Authors: Keerthi Prasad Ganeshaiah, Vinay Hegde

Handwritten character recognition (HCR) is a growing field in the applications of pattern recognition, image processing, communication technologies, and so on. But, the identification of handwritten characters affected because of different styles of writers, or even one writer’s style differed according to the conditions. Moreover, the huge amount of features from the handwritten characters also affect the classification. To address the aforementioned issues, the hybrid feature extraction (HFE) with morlet stacked sparse auto-encoder (MSSAE) based feature dimensionality reduction is proposed for improving the classification of HCR. Here, the HFE is the combination of different textural and shape features such as histogram of oriented gradients (HOG), gray-level co-occurrence matrix (GLCM), discrete wavelet transform (DWT), and skeleton features. The morlet wavelet activation function is used in the MSSAE to enhance the dimensionality reduction ability that used for an effective reduction of feature dimensions. The reduction of feature dimension using MSSAE is used to improve the classification using multi-class support vector machine (MSVM). In this research, the real-time indigenous languages i.e., English and Kannada handwritten characters from the Chars74K dataset are used for the analysis i.e., offline recognition. On the other hand, the online recognition of Kannada characters is also done for analyzing the HFE-MSSAE method. The HFE-MSSAE method is analyzed in terms of accuracy, precision, recall, CSI and F-measure. The existing researches namely convolutional neural network (CNN), hybrid feature based long short term memory (HF-LSTM) and elephant herding optimization - long short term memory (EHO-LSTM) are used to evaluate the HFE-MSSAE method. The accuracy of the HFE-MSSAE for Kannada is 96.73% which is higher than the HF-LSTM and EHO-LSTM.

477

Authors: Ammar Abdulsalam Al-Asadi, Mahdi Nsaif Jasim

Many deep learning-based recommender systems have been proposed recently. Where they involve all the users in datasets to build the latent representation of input data to be used later for predicting the missing rates. Despite the fact that, users have different interests, and these differences reduce the model prediction accuracy. This paper proposed a novel cluster-based denoising autoencoder model (cluster-based DAE) for rate prediction recommender systems. Instead of building a single model, it builds K models by using k-means algorithm to divide the users into groups based on their preferences. Each group trains a DAE model to generate recommendations for its members. The proposed method was trained and tested with MovieLens (100K, 1M, and 10M) datasets where 80% of the data are used for training and 20% for testing. The performance of the proposed method compared against other state-of-the-art methods that use deep learning to build rate prediction models. It outperformed the other compared methods in term of mean absolute error (with 12.9%, 14.7%, and 22.3%) and root mean squared error (with 24.2%, 18%, and 21.1%) using MovieLens 100K, 1M, and 10M datasets respectively.

497

Authors: Aruna Kumar Bananaik Thulajanaik, Naganaik Manjanaik

The mmwave (millimeter wave) with massive MIMO (Multiple-input and multiple-output) leverages larger antennas with spatial diversity and multiplexing gain brings about positive vibe for future generation wireless networks for provisioning bandwidth hungry multimedia applications However, the fronthaul link that links a remote radio head (RRH) and carries video frames as bitstreams to the baseband unit (BBU) can fluctuate the overall speed and capacity of massive MIMO if suitable data compression (encoding-decoding) techniques are not functional. This paper presents an error detection aware compression (EDAC) scheme which first performs encoding using low-rank approximation considering both spatial and time domain to reduce video frame size; later, the encoded video frames are compressed using Huffman codewords. Experiment outcome shows the proposed video compression model achieves much better compression efficiency with improved bit error rate (BER), symbol error rate (SER), error vector magnitude (EVM), and compression ratio when compared with the fixed code dictionary huffman video coding (FCDHVC), exiting video compression (EVC) technique and huffman video coding (HVC). The proposed EDAC model achieved compression efficiency of 30.52 whereas the FCDHVC, EVC and HVC model achieved compression efficiency of 26.50, 21.75 and 22.35 respectively.

517

Authors: Kavitha Chandrashekar, Anitha Tuluvanooru Narayanreddy

The use of machine learning (ML) within medical field is on the rise, notably as a means to enhance both the speed and precision of diagnosis. Through evaluating large volumes of patient information, machine learning is able to provide disease prediction, giving both patients and doctors more control over their health. Predicting and preventing heart disease has become a major area of study in medical data processing as a result of the increased expense of therapy. Since there are so many factors that come into play, estimating one's heart disease risk manually is a challenging task. Moreover, there are very few methods which provide better accuracy for the prediction of the heart disease. Hence, by using openly accessible cleveland heart disease dataset, this research aims to design and evaluate several advanced technologies constructed employing machine leaning algorithms for diagnosing if an individual is going to get heart disease or not. In this paper, we propose an ensemble feature optimized (EFO) learning method which uses an enhanced extreme gradient boosting tree and feature level cross validation scheme for effective heart disease (EHD) prediction. The presented EFO prediction algorithm and other existing machine learning algorithms have been used for the prediction of the heart diseases. The performance of the existing algorithms (XGB-based, ensemble tree hyper optimization (ETHO), and MLP-PSO) and proposed EFO algorithm has been evaluated using the classification metrics. When compared with the XGB-based, ensemble tree hyper optimization (ETHO), and MLP-PSO algorithm, the EFO algorithm has attained an accuracy of 98.61%. The EFO algorithm provides the doctors to able to predict the heart disease more efficiently and effectively.

536

Authors: Varsha Pravin Hole, Nikahat Mulla, Pravin Basavraj Hole

Cloud storage has gained extensive attention due to its low cost, global scalability, and improved agility. Data privacy is an important issue in cloud storage because breaches of personal information or business secrets are occurred by malicious hackers or unauthorized cloud operators. To address this issue, a Modified Key Sifting Scheme (MKSS) is proposed for developing an effective Quantum Key Distribution (QKD) with Two-Way Identity Authentication (TWIA) to provide security in the cloud storage. The Tree Parity Machine (TPM) is used for proposing a MKSS which used to minimize the key distribution time based on the synchronization. This MKSS-QKDTWIA considers four different objects such as data user, data owner, server and data storage in a communication scenario. The performance of the MKSS-QKDTWIA is analyzed by means of key computation time, processing time, encryption & decryption time, and memory utilization. Existing research such as Modified Random Fibonacci Cryptographic (MRFC), Software Stack-QKD (SS-QKD) and Enhanced Cloud Security Model based on QKD Protocol (ECSM-QKDP) are used to evaluate the performance of MKSS-QKDTWIA. The key computation time of MKSS-QKDTWIA for 50 attribute sets in each key is 16.7s which is less when compared to the MRFC, SS-QKD and ECSM-QKDP.

561

Authors: Narumon Wannoi, Chaisit Wannoi, Chai Chompooin-inwai, Nirudh Jirasuwankul, Churirat Boonkhun

This article presents the impact study of electrical vehicle charging stations (EVCS) loading to the Thailand’s North eastern power system. Main purposes of this paper are to find the maximum number of EVCS that can be loaded to power system and to rank the critical areas under EVCS load changing by using the repeated power flow calculation technique. Results from the calculation are for the integration planning of distributed renewable generators (DRG) to reduce system congestion. Typically, the critical areas are buses or devices that violate of system control limits. The system control limits are consisting voltage limit and load percentage limit of power transfer devices. All critical areas are needed to improve system stability and reduce system congestion for support the EVCS growing. In this study, The EVCS have been installed on main roads in Thailand’s North-eastern area from Bangkok to Nong-Khai province with a length of 624 km. The EVCS were installed every 50 km, totally 11 stations, and connected to 115 kV substations in the system. The impact study results found that the system experienced with congestion problem on following elements: a) 11 critical buses at 115kV network, b) one critical bus at 230kV network and c) one transmission line is overload to 100 percent at 115kV network when each of EVCS station load increase to 30 MW. In addition, study results reveal that when installed DRG at the right and proper locations, these can reduce congestions situation, reduce system losses and enhance system stability.

580

Authors: Rana H. Saloom, Hussein K. Khafaji

Molecular biology and bioinformatics are fields that alter knowledge and abilities for gathering, managing, storing, analyzing, interpreting, and disseminating biological data. In order to get insight into the design and testing of life sciences, it is necessary to make use of high-performance computers, cutting-edge software tools, and unique algorithmic approaches for data analysis, interpretation, and prognostication. The computing challenges of processing biological sequences are used in this research to describe a significant technique for producing DNA sequence alignments. One of these challenges is to put forward computational models that aid in speeding up the computation of the substitution matrix of the needleman-wunsch algorithm. Another difficulty is the implementation of a parallel program for two threads to find the values of the matrix concurrently in order to achieve high efficiency and find the best sequence alignment in less time. This paper introduces a novel approach to constructing the substitution matrix of the Needleman-Winch algorithm. This approach enhances the overall performance of the algorithm by adjusting the values of the penalties imposed for the gap, mismatch, and match in addition to computing the matrix using two parallel threads depending on the progress made in the devices, especially with regard to the capabilities and performance of the GPU. Numerous tests have been run using different data sets and different lengths of sequence. A comparison was made between the amount of time needed to implement the Needleman method and the amount of time needed to implement the suggested model using the same data in order to evaluate the performance of the proposed model. Processing time and speedup for parallel performance are experimentally calculated. The recommended model has high scalability in terms of workload and machine capacity, according to the performance evaluation and scalability assessments. The proposed approach reduced the execution time ratio by 47% to 90%, and this improvement rate rises as the length of the reference and query sequences does.

607

Authors: Ajay Mamindla, Ramadevi Y

One of the most prominent indicators for the detection of breast cancer is a breast mass. In this regard, effective mass segmentation for any type of mammography is crucial for improving breast cancer detection accuracy and lowering mortality. In order to pace up the process of mammogram segmentation for breast mass, an ABC3D (artificial bee colony based 3 dimensional) Otsu method is proposed in this paper. Firstly, convergence speed of bees in basic artificial bee colony (ABC) is improved by adopting the epsilon greedy method for scout bees. Secondly, proposed improved ABC method is paired with optimal 3D Otsu multilevel thresholding technique to get the better thresholding set for medical mammogram images. Epsilon greed based scout bee technique streamline the exploration-exploitation problem of standard ABC while searching for best threshold set in 3D space. The proposed ABC3D is tested on eight mammogram images collected from the authoritative and publicly available database mini MIAS (mammographic image analysis society). PSNR (peek signal to noise ration), SSIM (structural similarity index) and time cost are measured to record the effectiveness of ABC3D method. The results of experimentations indicate that the proposed ABC3D achieve superior segmentation results than the teaching learning ABC (TLABC), ABCDS (directed scout), gbest guided ABC (GABC), improved particle swarm optimization (IPSO) with 3D Otsu as objective functions.

13

Authors: Omaima El Alaoui-Elfels, Taoufiq Gadi

Capsule Network has proved its performance against image deformation, through its ability to preserve space relations among features. However, the Capsule Network underperforms at classifying complex images datasets. Through an in-depth examination of Residual Neural Network, we present a novel Capsule Network named ER-caps to improve the model performance toward complex images classification and reconstruction. The architecture of ER-caps uses customized residual blocks to enhance feature extraction without using any pooling methods to conserve spatial information among features and preserve the Capsule Network equivariance. Our model adopts the RELU and the ELU activation function to ensure fast convergence and uses an advanced decoder to allow better reconstruction for complex images. The experiments show that the suggested model outperforms the baseline Capsule Network with dynamic routing on a variety of benchmark datasets CIFAR-10, CIFAR-100, and SVHN by approximately: 17 %, 9 %, and 4 % respectively. Moreover, ER-caps reduces the number of training parameters.

38

Authors: Sudhir Dagar, Geeta Nijhawan

Robust Principal Component Analysis (RPCA) has recently been active everywhere for dimension reduction in image processing, display and pattern recognition. Methods based on low-rank sparse representations, which make some specific significant assumptions, have recently received a lot of attention in background modelling. Meanwhile, a powerful analysis framework is needed to handle background areas or foreground motion at various scales. this paper presents a hybrid approach along with total variation L1 (TV-L1) features and reproductive RPCA model in low rank background subtraction modelling and sparse matrix with mixture of Gaussians (MoG) as foreground modelling. The hybrid structure with TV-L1 features imposes a hierarchical RPCA on the singular values of the low-rank component and MOG sparsity indicators. The proposed work was evaluated on the CDnet2014 (ChangeDetection.net) dataset, obtained result as accuracy was 92.9%, 86.7% 95.7% for Highway, Escalator, and Indoor respectively. The proposed method is compared with traditional methods and obtained relative reconstruction error is 0.01529 as a lower side.

67

Authors: Karthik Sekar, Rajdeepa Balasubramanian

Software code clones (CC) in software programs are degrading the performance of software systems. many code clones detection (CCD) methods proposed in the literature detect only individual cloned types efficiently. This paper proposes a collaborative code clones detection (CCCD) method by utilizing lexical, syntactic, semantic and structural features for effectively identifying all types of clones including type-4. Initially, a large variance mapper (LV-mapper) is utilized to identify clone pairs (CPs). Then CPs are converted into lexical features by directly applying Word2vec. The synonyms of CPs are obtained using the WordNet tool and converted as semantic features by Word2vec.Additionally, code size metrics (CZM) and object oriented metrics (OOM) are additionally measured as structural features of a program’s code blocks (CBs). The syntactic features are extracted through the abstract syntax tree (AST) from the source code. Finally, the joint feature vector is generated by combining all the features together. In order to detect CCs in any new software, the joint feature vector of known clone type source codes is generated first (training set) and then the joint feature vector of unknown clone types source codes is generated next. The Euclidean distances between training and testing of joint feature vectors determine the clone type of test features. Finally, the experimental outcome demonstrates that the proposed CCCD technique has an accuracy of 87.8%, 92.3% and 95.5% for the dataset Apache Maven 3.8.3, Appache ant 1.10.12 and Opennlp-master 1.9.1, respectively, compared to the existing LV-CCD, ES-CCD, TBCNN-CCD and CPVDetector methods.

91

Authors: Manohara M, V. C. Veera Reddy, Vijaya Kumar M

Radial distribution systems (RDSs) can work better from a technical, economic, and environmental point of view if they have the right mix of renewable energy-based distribution generation (DG) and reactive power sources like and distribution-static compensator (DSTATCOM). In this paper, the Northern Goshawk Optimization (NGO), a new meta-heuristic that mimics the hunting behaviour of the northern goshawk, is introduced as a way to figure out where and how big different types of active and reactive power DGs in RDSs should be. For the suggested multi-objective function, NGO is used to find the best places and sizes for photovoltaic systems (PVs), wind turbine systems (WTs), and DSTATCOMs. The main goals of the suggested method are to reduce distribution losses, improve the voltage profile, increase the voltage stability margin, and reduce greenhouse gas (GHG) emissions. Five different case studies are simulated to figure out how well the suggested method works with computers. Simulations use IEEE 33-bus feeder. When compared to the literature and other heuristic algorithms, the results show that NGO works. The computational efficiency of NGO is compared for four cases: (i) only PV ((ii) only WT (iii) only DSTATCOM and (vi) simultaneous PV, WT and DSTATCOM allocation. The results also show that the proposed NGO much efficient than other algorithms and also improves all techno-environmental factors and RDS performance in a big way.

110

Authors: Amit Kumar Pandey, Prashant Singh, Dhyanendra Jain, Anupam Kumar Sharma, Ashu Jain, Anjani Gupta

Network Intrusion detection performances are highly affected by imbalance data problems due to the presence of less number of attack information in the dataset. Deep learning models are applied in existing methods to improve the efficiency that has limitations of overfitting problems. The Generative Adversarial Network (GAN) – Bayesian optimization Multi-class Support Vector Machine (BMSVM) is proposed to overcome imbalance and overfitting problems in intrusion detection systems. The Min-Max Normalization method is applied to normalize the input data to reduce the differences in features. GAN model is applied to generate minority class to balance the data instances to train the model. The proposed GAN-BMSVM model is compared with the classical sampling method, optimization, and classifier in the intrusion detection model in terms of Accuracy, Detection Rate (DR), and False Alarm Rate (FAR). The classical sampling methods are Near-miss, SMOTE and Autoencoder; traditional classifiers are KNN, RF, SVM, DNN and LSTM, and classical optimizations are PSO and WOA. The existing researches such as HCRNN, HLD, DONN, FL-NIDS and CNN-LSTM are used to evaluate the efficiency of GAN-BMSVM model. The GAN-BMSVM model has achieved 99.58% and 85.38 % accuracy for NSL-KDD and UNSW-NB15 dataset respectively, which is higher than the existing CNN-LSTM model.

131

Authors: Rana Dhia'a Abdu-Aljabar, Osama A. Awad

Lung cancer is more likely to relapse in the first five years following surgery; even though the operation may have been a complete success, there remains a chance that the lung cancer could return. This return may lead the patient to die after a successful surgery. Because there are no symptoms of lung cancer in its early stage, many researchers use intelligent systems to predict the relapse of lung cancer in its early stages. The outcome of previous works considering this issue still suffers from low prediction accuracy. This study proposed a method to predict lung cancer relapse more accurately. This method has multiple stages: 1st optimization system, feature selection stage, 2nd optimization stage, and extreme gradient boost (XGBoost) classifications stage. It used two datasets (GSE8894 and GSE68465) of a gene expression microarray for NSCLC with its clinical information on relapse state. We obtained three probes (3 genes) with clinical data combinations that can get good prediction results. These genes included 225389_at (BTBD6), 220239_at (KLHL7), and 204832_s_at (BMPR1A). A comparison between the proposed model and the original XGBoost with PSO and Hyperopt as hyperparameter optimization for the XGBoost classification model is performed. Extensive comparisons with four machine learning algorithms, including Deep Forest, K-nearest neighbor (KNN), Support Vector Machine (SVM), and Naive Bayes, are conducted. The proposed model accuracies are 0.93 for the GSE8894 dataset and 0.81 for the GSE68465 dataset.

154

Authors: Ahmed Kamil Hasan Al-Ali, Fadhil Sahib Hasan

In this paper, the security of image transmission through space time block coding (STBC)-coded orthogonal frequency division multiplexing (COFDM) is enhanced using two stage algorithms. In the first-stage, the samples scrambling algorithm using one dimensional (1D) chaotic map is proposed. In the second stage, the bits scrambling algorithm using chaotic map is proposed. The sample scrambling algorithm is designed depending on three types of 1D chaotic maps that are logistic-sine-cosine (LSC), sine-tent-cosine (STC), and tent-logistic-cosine (TLC) maps. In the bit scrambling algorithm, the stream bits are XORed with a pseudo-random bit generator (PRBG) produced by these chaotic maps using threshold method. Simulation results demonstrate that the image is retrieved with high quality for STBC-COFDM system and low signal to noise ratio (SNR). The image is unrecovered when the correct key is unavailable. For one level of security, the system using TLC chaotic map has the best performance compared with other chaotic maps. The bit error rate (BER) performance of the suggested system achieved 10-4 at SNR=1, and 2 dB for (TLC, LSC) and STC, respectively. For two stages of security, the proposed system using STC chaotic map improves the BER performance with a gain of about 2 dB at BER=10-4 compared with one stage of security.

173

Authors: Maysam Alwan Hasson, Taha Mohammed Hasan, Zaidoon Kamil Maseer

A numeral of deep learning models has been suggested for COVID-19 examination in computed tomography (CT) scans as an automation tool to help in diagnosis. Although, deep learning models achieved high accuracy, but training approaches are still inefficient to detect injections due to some deep learning models did not meet the requirement of a generalization term in deep learning. Furthermore, other traditional algorithms achieved low detection for 3D CT. Therefore, is high time to develop a deep learning model to diagnose COVID-19 infections in a regularization mode. In this research, greedy learning approach (GLA) is utilized to design and implement the 3D convolutional neural network (3D CNN) model, greedy learning approach is consistings of two stages; the first stage generates many 3D CNN models based on the randomness in the layers, for providing many movements toward solving one problem which is diagnosing COVID-19.Then, the second stage selects an optimal 3D CNN model based on high accuracy of 3D CNNs obtained in the first stage, optimal 3D CNN model is to be significate solution among them. We evaluate the proposed approach on the 3D Mosmed-1110 and 2D SARS-CoV-2 CT Datasets, the best accuracy scores obtained by the present approach are 1.00% and 98.87% respectively on the said datasets in terms of metrics, such as accuracy, precision, recall, and F1. The proposed system also exhibited good generalization and robustness, when it was trained and tested using a portion of data (80%) and (20%).

202

Authors: Sumanth Yamparala, L. Lakshminarasimman, G. Sambasiva Rao

One of the wind energy industry's fastest-growing technologies is the doubly fed induction generator (DFIG). Wind energy resources (WER) must be connected to the network both during and after grid voltage disturbances, according to the most recent grid codes. For improving the low-voltage ride through (LVRT) capability of DFIG without an additional hardware equipment during voltage sag, a control strategy is developed in this research, which is based on fractional order proportional integral derivative (FoPID) controller. A sea Lion adapted grey wolf optimized (SLnO+GWO) hybrid algorithm is employed for optimal selection of controller parameters. Consequently minimizes the variations in active, reactive powers, rotor current harmonics and dc link voltage in DFIG. Furthermore the proposed method is simulated in MATLAB/SIMULNK and has 2.56% THD in rotor current whereas existing methods has more. The performance of chosen method is compared with the methods available in literature RTO, MEHRFA, GSA and is found to be superior.

224

Authors: Purba Daru Kusuma, Astri Novianty

This research introduced a novel simple metaheuristic called as total interaction algorithm (TIA). TIA is a swarm intelligence which relies on the interaction among solutions in the population. The core and distinct concept of TIA is that each solution interacts with all other solutions in every iteration to find the best possible solution. Then, this new alternative replaces the current solution if it is better than the current solution. In this research, TIA is tested to solve 23 functions that represent both unimodal and multimodal problems. TIA is benchmarked with five metaheuristics: particle swarm optimization (PSO), marine predator algorithm (MPA), golden search optimizer (GSO), guided pelican algorithm (GPA), and driving training-based optimizer (DTBO). The result indicates that TIA is superior to other benchmark metaheuristics, especially in solving the high dimension functions. TIA is better than PSO, MPA, GSO, GPA, and DTBO in 22, 21, 16, 11, 13 functions. The result also indicates that the increase of the maximum iteration improves the performance of TIA mostly in solving high dimension unimodal functions. Meanwhile, the increase of the population size is less significant to improve its performance. Overall, this research resumes that the interaction with as many as possible individuals is proven better than with only selected individuals as implemented in many other metaheuristics.

246

Authors: Hind A. majeed, Ahlam M. Kadhim, Hazim G. Daway

The enhancement of colour medical microscopy images plays an important role in biological fields. Colour medical microscopy images have the lack of contrast and lighting. In this paper, a robust algorithm for enhancing microscopy images was proposed. In the proposed method, the image was first improved by using Contrast-Limited Adaptive Histogram Equalisation (CLAHE). Then, the luminance channel in the colour space was enhanced on the basis of sigmoid transformation and fuzzy logic. Data containing 50 medical microscopy image forms were studied, and the suggested method for enhancement was compared with several other methods, such as principal component analysis using the reflection model, fuzzy logic by stretch membership function, CLAHE, modified colour histogram equalization, contrast-enhancement based on median–mean based sub image-clipped histogram equalization and retinex method by colour restoration. The collected results illustrated that the suggested method gets excellent quality averages in terms of entropy value (7.96), average gradient (14.57) and mean of standard division (62.80).

265

Authors: Tody Ariefianto Wibowo, Nana Rachmana Syambas, Hendrawan, Leanna Vidya Yovita, Ade Aditya Ramadha

Due to named data network (NDN) 's advantages as a content-based architecture, it is a strong candidate for future internet architecture. Extensive research has been done on NDN implementation in the real world. However, the method widely used has advantages and disadvantages in reaching the goal. Simulation gives simplicity but is unlikely to provide reality in node implementation. The testbed method is close to real-world performance but needs higher effort. To solve the problems, building an environment with nature closer to the implementation world is required. The method also gives ease of use and observation, like simulation/emulation. Due to these needs, we proposed a framework that combines the advantages of emulation and a testbed that occupies a UNetLab-based virtualization environment for extensive observations for the NDN researcher. We build a test case in the PNETLab and Mini-NDN for gives a perspective on how our proposed method works. PNETLab framework evaluates NDN network’s nodes performance, giving 7,59% in RTT delay, and 0,89% in nCHR deference result to Mini-NDN, and provides intensive hardware performance parameters result which is very important for NDN implementation.

290

Authors: Rana H. Saloom, Hussein K. Khafaji

Numerous variables, such as changes in the environment, toxins, spontaneous mutations, and replication mistakes, can have an impact on DNA. It is known as a gene mutation when this occurrence permanently alters the DNA sequence that forms a gene, changing it from the sequence seen in most people. Alleles, which are small variations within the same gene, are produced as a result of mutations. Each person is different due to these minute variations in their DNA sequence. In particular, some mutations affect just the carriers, whilst others affect both all children and the carrier organism's progeny. Changes in the DNA sequence (genetic mutations) are one of the reasons that lead to life-threatening disorders such as cancer and other diseases, so it has become necessary to detect these mutations early and know their types and their impact on the DNA sequence The DNA modifications developing in the cells of the following generation are what bioinformatics is most concerned about. In this paper, an efficient approach based on the architecture of long-term memory (LSTM) recurrent neural networks is presented. The method involves locating mutations using the Needleman algorithm that compares the reference DNA sequence with the mutant sequence. The deep neural network is then trained on the Cancer Cell Lines portal (CCLE) dataset to classify the different types of genetic mutations that have been identified. The algorithm reports the mutation type and locus for each pair of DNA inputs. Finally, the simulation output demonstrates the effectiveness of the algorithm. By testing the method, it was found that it was able to identify the mutation type with 100% accuracy.

314

Authors: Oger Amanuel, Yarub Alazzawi

It has been shown recently that there is a need to design smart structures, such as smart houses, in order to be controlled in different ways. That will be in high demand due to its usefulness for some people who are incapable of reaching some control units that require direct interaction with human beings. In this paper, we propose and develop a new enhanced electroencephalography (EEG)-based smart structure setup that can be utilized to assist people, with or without disorders, to control devices in an easy and comfortable way. Ten people of a wide range of ages (20–65) and both genders actively participated in this research. Consequently, eight EEG channels are employed in this study to cover most of the brain’s regions, and the protocol utilized is suitable for people with disabilities and immobility. Finding the standard or common features for the wide range of participants is a challenge. To mitigate this, reconstruction independent component analysis (RICA), which is a modified technique of the conventional independent component analysis (ICA), was used to obtain the optimum features. In addition, the proposed modified support vector machine (SVM) model classifies the selected features into different classes with the capability of removing the high noise and overlaps that cause misclassification. The identified classes are responsible for actuating the smart house’s actuators based on participant status. Real-time classification of multi-channel EEG data into brain wave components, visualization of results, and control of the devices are carried out using MATLAB and an embedded system. With the proposed model, there is only one case of overlap between the classes, compared with 74 cases with conventional SVM. Consequently, the results of misclassification reach zero, and the proposed model enables control of a smart room based on brain waves, achieving an overall accuracy of 98%. With future improvement, the acquired findings would urge the usage of the suggested EEG-based smart structure, which might be helpful for immobile individuals.

339

Authors: Komagal Yallini Shanmuga Sundaram Kanagaraj, Mahendiran Nallappan

In data engineering, multi-label learning (MLL) has emerged to classify the instances through a specific characteristic that associates with the set of class labels (CLs). Mostly, the learning design was adaptive and newer views may exist in a data stream (DS); so, MLL has to classify the features with newer CLs. To tackle this issue, the MLL with emerging new labels (MuENL) and handling high-dimensional DSs (MuENLHD) technique has been adopted which considers the CLs in the test data were similar to that in the learning data. But, it was not able to deal with the adaptive situation where multiple newer CLs exist since it can manage only a single newer CL in one iteration. Hence this article proposes an MLL with emerging multiple new labels (MuEMNL) and MuEMNLHD versions to combat the issues in a complex scenario wherein several newer CLs are found. The main idea of this technique is to split the newer CL group into many newer CLs independently for changing the complex scenario. In this technique, four different steps are executed: i) creates a linear classification model to adjust the pairwise CL sorting error and the categorization error on the given CLs, ii) develops a novel outlier identifier depending on the primary and test DS, iii) discovers the cluster for the MuEMNL and MuEMNLHD depending on the OPTICS clustering and iv) employs a classifier updating scheme to integrate newer CLs for designing a robust classifier. Finally, the experimental outcomes exhibit that these techniques on low-dimensional databases attain an overall mean precision of 67.08% and an overall F1-score of 64.02% compared to the classical MLL techniques. Similarly, these techniques on high-dimensional databases achieve mean precision of 64.6% and a mean F1-score of 63.1% compared to the classical MLL techniques.

364

Authors: Priyanka Y, Raghu Ram

The distribution systems act as the primary connecting point between the power utility and the end users of the electricity. It is vital to assess the behaviour of the network when subjected to a variety of various types of loads and any new D-FACTS devices that may be developed in the near future, in order to improve the operation and planning of electrical distribution systems. The authors of this article performed an analysis of the distribution system when unified power quality conditioner (UPQC) was present and acting as a Distribution FACTS device. The main objective of this article is to minimization of total investment cost of UPQC as well as system operating considering distribution network reconfiguration (DNR). For optimal solution, an improved whale optimization algorithm (IWOA) has been introduced. Besides, impact of voltage dependent load models has been studied. In order to evaluate the practicability and usefulness of the suggested strategy, a number of potential outcomes involving the DNR and allocation of UPQC under a voltage dependent load model have been considered. The active and reactive power losses have been reduced by 44.67 % and 37.42 %, respectively, as compared to conventional schemes (such as UPQC allocation solely). In addition, the proposed methodology results in cost reductions of around 33.68 % annually.

388

Authors: Mustafa A. Kamoona, Omer Cihan Kivanc, Oday A Ahmed

An energy management system (EMS) for hydrogen fuel cell hybrid electric vehicles (FCHEV) based on artificial intelligent (AI) technique is presented in this paper. In order to achieve a fast dynamic response and maintain high efficiency of energy storage resources, the fuzzy logic controller (FLC) and artificial neural networks (ANNs) are utilized for purpose of intelligently managing the system's power flow. Moreover, the feed-forward wavelet neural network linked with proportional-integral (PI) controller named (WNN-PI) and the recurrent wavelet neural network linked with PI controller named (RWNN-PI) are tuned by particle swarm optimization (PSO) algorithm, both are aimed at the operating of the resources at high efficiency with respect to their mechanism performance, meeting the load power demands efficiently, and reducing hydrogen usage. Finally, a comparison of the simulation outcomes is presented to choose the best of the proposed AI controllers where the results showed optimum power flow between power sources and load power of FCHEV then as consequence, the BAT and UC are run in a safe manner and extend their lifetime, also, the average efficiency of the FC stack has been increased, and the amount of usage of hydrogen fuel is reduced. The simulation of AI EMS has been carried out by MATLAB/Simulink R2022a, with various vehicle driving cycles by the advanced vehicle simulator (ADVISOR).

410

Authors: Hussein Fouad Almazini, Ku Ruhana Ku-Mahamud, Hassan Almazini

Anomaly detection deals with identification of items that do not conform to an expected pattern or items present in a dataset. The performance of the various mechanisms that are employed to execute anomaly detection is strongly dependent on the set of features that are utilized. Thus, not every feature in the dataset may be employed in the classification operation since certain characteristics may result in poor solution quality. Feature selection (FS) may reduce the size of high-dimensional datasets by eliminating unimportant features. Modified binary grey wolf optimizer (MBGWO) is a successful metaheuristic that has been used for FS in anomaly detection. Nonetheless, the MBGWO is a randomization population-based algorithm that has an issue in finding a good quality solution during the initial population procedure. Thus, this study proposes a heuristic modified binary grey wolf optimizer (heuristic MBGWO) algorithm for FS in intrusion detection to enhance the initial population of the MBGWO using a heuristic-based ant colony optimization algorithm (ACO). The heuristic MBGWO algorithm was evaluated on NSL-KDD benchmark dataset from the University California Irvine (UCI) repository against five (5) benchmark metaheuristic algorithms. experimental results of the heuristic MBGWO algorithm on the NSL-KDD dataset in terms of the number of chosen attributes and classification accuracy are superior to other benchmark optimization algorithms, where it obtained the best features with 99.85% classification accuracy. The proposed heuristic MBGWO algorithm can be used for FS in anomaly detection tasks that involve any dataset size from various application domains.

435

Authors: Sreenivasulu Ummadisetty, Madhavi Tatineni

The absence of P waves through electrocardiogram (ECG) tracing causes atrial fibrillation (AF) affecting around 1% of the global population. In recent years, wearable and portable devices have made mobile healthcare much closer to reality. The main purpose of this article is to develop an automatic AF detection system based on short single lead ECG signals. Also, AF is one kind of arrhythmia that change the rhythms in the heart and have the potential to alter the characteristics of morphology in ECG tracings. For feature extraction, heart rate variability (HRV) and frequency analysis are adopted. The novel contribution of this work is to deploy a modified Moth Flame Optimization algorithm for detecting AF in a short ECG recording. The obtained results were validated with the available public data set comprised of short ECG recordings by genetic algorithm and modified moth flame optimisation algorithm. For N versus A classification, accuracy varies from 94.2% to 97% under noise levels ranging from 0 to 30 dB. For N versus A versus O, maximum accuracy of 84.3% is obtained. The obtained experimental results suggest that HRV is efficient and robust for AF detection for relatively short ECG recording.

474

Authors: Syamala Misala, Satya Anuradha Mosa

A compact rectangular dielectric resonator antenna (RDRA) with a surface-mounted short rectangle (SMSR) having an E- a shaped microstrip feed of wide aperture slot investigated for C-band operation in applications of Radar as well as wireless. The surface-mounted short rectangle improved the gain and efficiency characteristics. The size, material, and cost are essential considerations in designing an antenna due to advancements in C-band applications. The RDRA designed on a small-size substrate is 46x46x1.6 mm3 made with FR4. It is rugged and can withstand in all pressure conditions. The proposed design utilizes copper has high conductivity and low cost for SMSR enhancement of return loss up to -47 dB, additionally gain to 9.4 dB, and observed bandwidth impedance of 19.6 % in the wide frequency of (5.75– 7.5) GHz. Furthermore, 3dB beam width attained in E – plane 108.430 whereas in H – plane is 64.360. The proposed SMSR design outcomes the conventional counterparts in all salient parameters, including return loss, impedance bandwidth, Gain, and Radiation efficiency with small size and low cost.

493

Authors: Marouane Rayyam, Malika Zazi

The design of electrical drives capable of tolerating sensor faults has recently become very popular due to their possible use in Fault Tolerant Control systems. This paper introduces a virtual speed sensor design based on a nonlinear state estimation algorithm for fault-tolerant vector control for Induction Motor Drives after speed sensor faults. The rotor speed estimator is realized by integrating the Unscented Kalman Filter (UKF) and the Sliding Mode Observer (SMO) based only on the measurements of the stator current and voltage. This proposal is dedicated to maintain the proper behaviour of the Indirect Rotor Field Oriented Control of the induction motor drive, where it is necessary to switch to the nonlinear estimators as a result of damage to the speed encoder. Simulation studies of the proposed fault tolerant design are performed in the case of total and partial losses of the speed sensor in a closed loop control structure for different operating conditions of the drive system. Fault-tolerant designs based on the UKF and the SMO estimators show good performance in the presence of speed sensor faults in terms of stability, overshoot rate (less than 10%), and trajectory tracking over low and medium speed range. However, the SMO performance is reduced for high-speed operation (⍵r >1400rpm) over the UKF which maintain good performance over the entire speed range.

518

Authors: Alaa Thamer Mahmood, Raed Kamil Naser, Sura Khalil Abd

YouTube platform provides paid/ unpaid video services for different users for promotions, education, learning, information sharing, etc. Accessing this platform's content/ channel-based advancements is based on user identification through registration/ credential sharing. In recent years, accessing information through fake accounts has increased other users' illegitimacy and security demands. The article introduces a temporal behavioral method (TBM) using the classified learning (CL) technique to address the impact of fake users on YouTube-like platforms. The proposed method eyes on user credentials, interactions, and access information under different usage series. The data used for the classification learning identifies the active, inactive, and false searches/ information accessed by the user. With the recent meetings removed, the static and inaccurate searches are segregated using the learning process from the last known access session. Rather than applying new rules to the identified account/user, false user flagging reported data sharing and illegitimacy is all avoided. Based on the different session logins and the recommendations by the learning process, the fraudulent users are restricted from communicating with other users and legitimate information. Maximizing data sharing and user detection is facilitated by the better recognizability afforded by the prolonged inactive session categorization following in succession. The suggested TBM-CL strategy addresses the impact of fake users on YouTube-like platforms and increases the maximums for user detection (14.93%), information sharing (11.3%), and classifications (15.93%) across all sessions. It decreases the erroneous ratio by 8.03% and the delay by 9.07% based on the dataset.

548

Authors: Saja A. Nahoo, AllaEldin H. Kassam

A key field of research for addressing complicated decision problems is multiple criteria decision-making (MCDM), which has grown quickly. The purpose of the paper is to explore the performance appraisal model for educational institutions. In this paper, Analytic Hierarchy Process (AHP) is used to calculate the subjective weights and Shannon’s method is used to calculate the objective weights, then VIKOR method is carried out using these weights which finally results in the ranking of the alternatives. To explore the feasibility of the developed model, the performance evaluation of twelve Iraqi medical colleges was considered as alternatives applied based on eight academic criteria. The results indicated that the sixth choice received the highest rating, while the fifth alternative received the lowest rating. Finally, sensitivity analysis (SA) is used to determine how the alternatives change depending on how important the criteria are. SA results the eighth criterion (Scientific output) was discovered to have the greatest influence on changing the priorities of the alternatives. The proposed method is receiving critical attention from higher authorities because it is considered a reliable scientific method of performance appraisal.

572

Authors: Parag Rastogi, Devendra Singh, Sarabjeet Singh Bedi

The medical technology is being evolved for enhancing the healthcare information related to medication, diseases, medical images etc., by using Electronic Healthcare Recording (EHR). The details such as gender, age, and weight of patients are sensitive, and require protection against unauthorized access. Therefore, providing a robust security for medical data is indeed a challenge. The present research work proposed a Fully Decentralized Block chain (FDBC) model with the Proxy Re-Encryption Algorithm (PReA) with an Ethereum smart contract. The proposed FDBC-PReA allows the data to be visible only for the owner and the smart contract as the model is implemented in an Ethereum based test bed. Each time when the data is transformed, the data loss is obtained for the incorrect data which is taken care by the decentralized BC. The fully decentralized BC stores and checks whether each of the entity shares or accesses the real time data. The proposed FDBC-PReA Ethereum smart contracts, and proxy re-encryption algorithms are executed with a faster rate that consumed average block time of 5.48s, which is lesser when compared to average block times of Smart-Contract Ethereum Distributed Ledger-14s, Searchable Encryption-48.125s, and Decentralized Security Architecture-Software Defined Networking (SDN)-10s.

597

Authors: Maura Widyaningsih, Tri Kuntoro Priyambodo, Moh Edi Wibowo, Muhammad Kamal

Contrast improvement and noise reduction are needed to improve aerial images prone to interference during data collection. Contrast limited adaptive histogram equalization (CLAHE) is a histogram equalization (HE) development method that is commonly used for contrast improvement. Median blur (MB) is one of the methods used for noise reduction. Combining these two techniques helps optimize the preprocessing process before semantic segmentation analysis is carried out in image maps. The results of testing experiments with U-Net VGG-16 and VGG-19 on image maps show a detailed representation of the predicted pixel class. Comparing accuracy with state-of-the-art methods shows that contrast enhancement and noise reduction are better than the previous method. The highest average result for combining CLAHE+MB with U-Net VGG-16 was 76.5 and VGG-19 was 73.8, and the highest accuracy for image sample testing was 87.94 with U-net VGG-16.

11

Authors: Manaswini Pradhan, Alauddin Bhuiyan, Subhankar Mishra, Thanh Thieu, Ioana L. Coman

In the recent decades, the lung cancer is the most dangerous disease that causes 1.6 million deaths per year. Therefore, the early recognition of lung cancer is an effective way to reduce the mortality rate. Compared to other imaging techniques, the histopathology images are effective in identifying the location and level of cancer. In this manuscript, a novel model is implemented for an automatic classification of histopathological images related to lung tissues. Initially, the color normalization technique is applied for improving the contrast of the histopathological images, which are acquired from the LC25000 lung histopathological image dataset. Additionally, the cancer segmentation is accomplished utilizing saliency driven region edge based top down level set (SDREL). Further, the feature descriptors: Alexnet and gray level co-occurrence matrix (GLCM) features were used for extracting the feature vectors from the segmented histopathology images. Lastly, the enhanced grasshopper optimization algorithm (EGOA) and random forest classifier were used for optimal feature selection and lung tissue classification. The simulation result shows that the EGOA-random forest model obtained 98.50 % of accuracy on the LC25000 lung histopathological image dataset.

35

Authors: Valliammal Narayan, Sudhamathy Ganapathisamy

Machine learning methods are widely applied in credit card fraud detection to improve its efficiency and automate the process. The existing methods in credit card fraud detection have limitations of imbalance data problem. In this research, Hybrid Sampling (HS) - Similarity Attention Layer (SAL) – Bidirectional Long Short Term Memory (BiLSTM) is proposed to improve the classification performance of credit card fraud. Two datasets such as European data and Revolutionary analytics are applied to evaluate the SAL-BiLSTM model. Hybrid sampling of SMOTE-ENN model over-sample the minority class and under-sample majority class that helps to reduce difference between original data and generated data. The SAL is introduced to measure similarity of sequence of data to provide weight to unique features and reduces overfitting problem in classification. The proposed method has the advantage of removing the overlapped samples in the majority class and increasing the instances in the minority class. The SAL method in the proposed method helps to focus on unique features in the datasets to improve the classification performance. The BiLSTM model performs forward and backward analysis to find the relevant features for classification. The proposed HS-SAL-BiLSTM has a 99.2 % recall value and the existing RF-SMOTE-Support Vector Machine (SVM) has a 97.7 % recall value.

54

Authors: Dedy Kurnia Setiawan, Widya Cahyadi, Bambang Sri Kaloko

Uninterruptible power supply (UPS) is a form of electrical supply that is frequently used to regulate electrical current and perform as a backup power source when the primary power source fails. Apart from providing backup power, the inverter system proposed in this study can reduce harmonic effects and increase power factor. The control system is based on a synchronized reference frame (SRF) and uses online combined error adaptive fuzzy (OCEAF). The load voltage and current are decomposed into their symmetrical components: positive, negative, and zero. These components are then transformed to a synchronous reference frame (dq), which OCEAF controls. Controlling the inverter is accomplished by converting the controller output to abc coordinates. OCEAF combines the delta error adaptive fuzzy (DEAF) and the absolute error adaptive fuzzy (AEAF) techniques with the NN adaptor. This adaptive fuzzy-PI automatically modifies the Kp and Ki values when the system changes, significantly when the source value alters or is faulty. According to power factor measurements, OCEAF performance is superior, more consistent, and closely unity power factor. As for evaluating network harmonics, systems with OCEAF controllers may reduce network harmonics from 6.24 % to between 1 and 1.5 %. The DEAF and CEAF techniques reduce harmonics by 2-2.5 and 1.5-2 %, respectively.

74

Authors: Yenni Astuti, Risanuri Hidayat, Agus Bejo

Speaker recognition system is intended to recognize a person’s identity. This task can be done by knowing the feature of the contained voice signals. The feature can be extracted using a feature extraction technique. One of the most popular feature extraction techniques is Mel-Frequency Cepstral Coefficient (MFCC). Until this time, the MFCC is still a challenging technique to be developed. In this work, a proposed technique is developed based on the MFCC. The original MFCC is modified in the process of filter bank and by adding spectrogram. The target of this work is to obtain a better recognition accuracy than the original MFCC. This target can be done by applying a technique called Mel-weighted spectrogram. The output of the proposed technique are spectrogram images which contain the feature of the voices. The spectrogram result is then classified using dissimilarity space based on Euclidean distance to identify the person’s identity. For the dataset, this work uses 315 recorded voice signals, consisting of 3 speakers, each pronouncing five words repeatedly 21 times on seven different days. The performance of this system is evaluated by comparing the percentage of accuracy among the proposed technique, the original MFCC, and two other MFCC-based techniques. In this work, the proposed technique is better than the three other techniques with an accuracy up to 88.57%. From these results, the Mel-weighted spectrogram can be considered as a recommendation for obtaining a higher recognition rate in speaker recognition system.

91

Authors: Preethi Srinivasalu, Aishwarya Palaniappan

In recent times, brain tumor detection become an important task in medical image processing applications. The early detection of brain tumor improves the treatment process and increases the survival rate of the patients. However, the manual segmentation and classification of brain tumor is a complex and time consuming process. Therefore, a new automatic brain tumor detection model is implemented in this manuscript for effective brain tumor detection. After collecting the scans from the cloud, the normalization and adaptive histogram equalization techniques are used to enhance the acquired brain scan quality. Further, the tumor regions are segmented by integrating the Adaptive Kernel Fuzzy C Means (AKFCM) clustering algorithm with the Otsu thresholding technique. Next, the deep and textual feature values are extracted from the segmented regions using the Gray Level Co-occurrence Matrix (GLCM), Local Ternary Pattern (LTP), and LeNet-5. The dimension of the extracted feature vectors is optimized by using the Modified Particle Swarm Optimization (MPSO) algorithm, which are given as the input to the Multi-Support Vector Machine (MSVM) for tumor classification. The experimental outcomes confirmed that the MPSO-MSVM model obtained high accuracy of 98.89%, which is superior related to the existing machine learning techniques.

111

Authors: Thuan Thanh Nguyen, Huu Truong Trinh, Hoai Phong Nguyen, Tan Loc Vo, Thanh Long Duong

This paper demonstrates an artificial ecosystem optimization (AEO) for the multi-goal electric network reconfiguration (ENR) problem (ENRP). The membership goal functions of the ENRP comprise power loss reduction, voltage deviation reduction, reduction of load unbalance index among branches and reduction of number of switch operations. To show the advantages of AEO for the multi-goal ENRP, the multi-goal ENR approaches relied on particle swarm optimization (PSO), genetic algorithm (GA) and cuckoo search algorithm (CSA) are applied to contrast with the AEO method. The result comparisons among methods on the 33-node distribution network show that the AEO method can archive the optimal solution with the higher success rate and the lower average value and standard deviation of the fitness values than GA and PSO. In addition, the result comparisons with other previous approaches also show the reliability of the proposed AEO method for the single and multi-goal function ENRP. Therefore, the AEO can be an effective and useful approach for the ENRP to optimize the single and the multi-goal functions.

132

Authors: Anjani Parvathi K, Kotaiah N.C, Radha Rani K

One of the challenging tasks in an active distribution network (AND) embedded with intermittent renewable energy sources (RES) under islanding conditions is the maintenance of frequency and voltage profiles within tolerable limits. Failing to maintain these operational requirements may lead to voltage collapse or complete blackout of the network. In order to avoid this scenario, ADNs may function with contemporary load shedding schemes but these schemes may result in an inadvertent and excessive amount of load shedding, thereby consequently causing unsatisfactory and low reliability at the consumer level. Thus, this paper used Pelican optimization algorithm (POA) approach for minimising the amount of load to be shed by determining the optimal amount of load to be shed within the consumers’ specified limits. Simulations are performed on IEEE 33-bus by assuming different kinds of renewable and distribution generation (DG) units. The computational efficiency of POA is compared with literature works and its performance is also characterised based on 25-indipendent runs of each case. The results obtained by POA are observed to be superior in terms of global optima and less run time. Also, the ADN is observed as satisfactory operation with fewer distribution losses, an improved voltage profile, and enhanced stability margins by imposing optimal load controls to the consumer.

152

Authors: Sumaya A. Ibraheem, Omar Y. Shaaban

Nowadays, latency factor in network virtualization takes a serious attention in 5G networks upgrading. Flex-grid/elastic optical network (EON) virtualization recently become known as the fifth-generation (5G) enabling technology for the optical network slicing efficiency. As a consequence, the solution of the virtual optical network embedding (VONE) problem over EON must be solved with the latency targets guarantee. In this work, a new loading factor was suggested after adding the latency effect to the existing EON constraints by identifying its contributing factors in the network. Then, a virtual network mapping algorithm was presented to solve the VONE problem including latency constraint. Network topologies (NSFNET, USNET) were used under the dynamic scenario and four optimization methods which are; alignment and consecutiveness-aware virtual network embedding (ACT-VNE), the local resource capacity and the shortest path first fit (LRC-SP-FF), greedy algorithm and the shortest path first fit (Greedy-SP-FF), load and resource-aware based on ant colony optimization (LRA-ACO) were used and results have been compared to get the optimal solution. The blocking probability on average was increased by approximately 50 % in both topologies affected by adding latency (between 20-50 ms) as compared to the non-latency cases of previous works. Also, during all simulations, the LRA-ACO gave the best embedding performance.

179

Authors: Purba Daru Kusuma, Anggunmeka Luhur Prasasti

This paper presented a new metaheuristic technique, namely the guided pelican algorithm (GPA). GPA has the improvements for a shortcoming algorithm, namely the pelican optimization algorithm (POA), that mimics the behaviour of pelican birds during hunting prey. It improves the original POA in three ways. First, GPA replaces the randomized target with the global best solution as a deterministic target in phase one. Second, GPA replaces the pelican's current location with the search space size in determining the local search space size in phase two. Third, GPA implements multiple candidates in both phases rather than a single candidate as it is used in the original POA. Simulation is implemented to observe GPA’s performance in optimizing both theoretical and real-world problems. GPA is compared with four algorithms: marine predator algorithm (MPA), particle swarm optimization (PSO), komodo mlipir algorithm (KMA), and POA. The result shows that GPA outperforms all sparing algorithms in optimizing most benchmark functions. GPA is also implemented to optimize the portfolio problem. The result shows that GPA outperforms all sparing algorithms in optimizing the portfolio problem. It outperforms three sparing algorithms in optimizing the portfolio problem. Its performance is 9%, 11%, and 13%, better than PSO, MPA, and KMA consecutively. Meanwhile, its less than 1% worse than POA.

201

Authors: Periya Nayaki Anbarasu, Thanabal Muniyandi Suruli

Nowadays, clinical decision support system (CDSS) is emerged broadly due to the growth of smart applications. Yet, labeling a massive amount of clinical data is expensive and needs domain expertise. Also, how to design a robust CDSS model in a healthcare platform with a minimal number of annotated clinical data was an essential problem. To tackle this issue, a generative adversarial network (GAN)-based semi-supervised training model was developed, which increases the quantity of annotated data and solves the unbalanced annotated labels. As well, support vector machine (SVM) and K-nearest neighbor (KNN) classifiers were applied as the base learners to predict the label for unannotated data. But, the result from both base learners was not often similar, which impacts the prediction efficiency. So, this article introduces a deep ensemble learner model (DELM) for predicting labels to the unannotated data by stacking multiple machine learning classifiers as base learners. In this model, two levels of prediction: the initial level and the second level. During the initial level, multiple base learners including SVM, KNN, naive bayes (NB) and random forest (RF) learn the dataset with both the annotated and unannotated samples independently to predict the class of the unannotated instances. Then, those prediction outcomes are merged during the second level, which applies the deep neural network (DNN) as an ensemble classifier to get the final predicted class of the unannotated instance. Finally, the experimental results exhibit that the DELM with GAN achieves 86.54 %, 84.83 % and 86.72 % accuracies on SPECT, WDBC and Hallmarks databases, correspondingly, compared to the Fuzzy-AHP+ANN, HTM+LSTM, DBSCAN+SMOTE-ENN+XGBoost and GAN-based semi-supervised models.

224

Authors: K. Soumya, Vijay Kumar Garg

Detecting suicide ideation from social media texts is challenging as it necessitates analyzing the content and context of the text utterances. Modeling the temporal trajectory of suicide through various stages like stress, depression, thought and strengthening of the thought becomes difficult with content features alone and this needs context cues too. This work proposes a hybrid feature representation incorporating both content and context features. The richness of features is enhanced using adversarial learning. Neural machine translation is done on hybrid feature representation to provide named entities with emotional intensity tagging. The sequence of emotional intensity tagging is mapped by Bi-directional Long short term memory (LSTM) to suicidal ideation label. The proposed integration of emotion tagging to SI detection with Bi-directional LSTM provided an SI detection accuracy of 95.54% which is atleast 3% higher compared to existing works.

247

Authors: Saif K. Jarallah, Sawsen Abdulhadi Mahmood

Video summarization technologies strive to provide a succinct and thorough description by selecting the most informative frames from the original video. The essential concept of query-based video summarization methods is represented by constructing a video summary related to user query in term of user interest. This paper aims to address the problem of query-based video summarization construction task by adopting a lightweight deep learning model for object detection task in order to evident how concerning a frame or shot is to a given query. Both YOLOv3 and Tiny YOLOv3 deep learning models are used separately in the proposed system to train the networks using images and videos dataset including diverse types of objects such as; (motorcycles, bicycles, cars, buses, and trucks). Subsequence, the most relative frames to a given query are selected and assembled as keyframes using a modified K-mean clustering scheme to provide different interesting summaries from the original video. Based on the experimental results of object detection phase, we have obtained an average object detection accuracy around 93%, 83% based on YOLOv3 and Tiny YOLOv3 deep learning models respectively. Comprehensive experiments were performed to evaluate the proposed video summarization system, which exhibited an efficient summarization rate close to 33% of the original video. Further experiments were conducted using standard UTE dataset to exhibit the competitive performance of the proposed method compared to state of art query-based video summarization methods.

273

Authors: Neny Sulistianingsih, Edi Winarko, Anny Kartika Sari

The Girvan-Newman (GN) method is one of the most popular methods for detecting communities. However, the method is applied to graphs with only positive weights, while graphs with positive and negative weights are in the real world. This paper proposes improving the GN method to work on graphs with positive and negative weights. The method is called Girvan-Newman Parallel Positive Negative (GN-PPN). Other than the benefit related to the use of negative weights, GN-PPN has a parallelization process that accelerates processing time. This method also uses edge betweenness and pair dependencies as calculations. The GN-PPN method is evaluated using modularity score, Normalized Mutual Information (NMI), and processing time. Our experiments show that the modularity score of the GN-PPN is similar to GN, modified GN (mGN), and Parallel GN (PGN). The accuracy of the GN-PPN method in detecting communities evaluated by NMI shows that the GN-PPN is better than similar methods such as Infomap, which ranges from 0.651 to 0.937. The processing time of the GN-PPN method shows a significant acceleration, which is a decrease of around 45% to 95% compared to other methods such as GN, mGN, and PGN.

293

Authors: De Rosal Ignatius Moses Setiadi, Rahmawati Zulfiningrum, Eko Hari Rachmawanto, Pulung Nurtantio Andono

Telemedicine is a communication in the medical field that requires privacy and confidentiality, so encryption is needed as a security method. The development of encryption methods must accompany technological developments. This research proposes a cryptography method developed and tested on medical images. This method consists of bit plane slicing, system chaos, and hash function. Bit plane slicing to break up the image bits so that they can be scrambled independently and dynamically with the chaos method. The hash function encrypts key 1, which is user input. The chaos method is run with dynamic parameters generated with a hash function to perform the confusion process on each bit plane. Then bit-plane join is performed, followed by the diffusion process using a logistic map with the second key for the parameter. The method is sensitive to slightly modified plain images or keys. Several measurement tools use to prove the method's performance. The result is a mean value are 7.999 for entropy information, 99.6027 for the number of pixels change rate (NPCR), 33.0925 for unified average changing intensity (UACI), 0.0021 for correlation coefficient, 231.8679 for chi-square and histogram, 50.0550 for avalanche effect, and perceptual analysis have proven the proposed method has satisfactory performance. The decryption is also done perfectly, as verified by the structural similarity index (SSIM)=1, peak signal-to-noise ratio (PSNR) is ∞, and bit error ratio (BER) is 0. Finally, this method can also work very fast in less than one second, so it can be implemented in real-time encryption.

315

Authors: Nova Rijati, De Rosal Ignatius Moses Setiadi, Pulung Nurtantio Andono

Tamper detection and localization are essential things in fragile image watermarking to carry out the authentication process. The tampering method continues to evolve from general and complex tampering. This study proposes a tamper detection and localization technique that can withstand a variety of standard and complex attacks with a combination of bidiagonal singular value decomposition, blockwise and group block for authentication. The least significant bit method is used in the embedding process to increase imperceptibility. By combining these methods, fragile image watermarking is produced that is robust against various common tampers such as type I and type II copy-paste, rotate, text addition, and noise addition, as well as various complex attacks such as vector quantization, collage, content only and constant feature. The tamper detection rate is more than 0.99, and the average PSNR value is 51.85 dB. This shows that the proposed method is robust against various tampers and has excellent imperceptibility.

337

Authors: Abousserhane Zakaria, Abbou Ahmed, Bouzakri Hicham

The aim of this paper is developing and validating the sliding mode control application (SMC) for hybrid renewable energy system (HRES) featuring stand-alone and on-grid operations modes. The proposed system composed by a photovoltaic (PV) and wind turbine (WT) as principal sources. The PV is connected to the overall system via boost converter which controlled using MPPT algorithm depending to Adaptive Perturb and Observe (APO) method managed by SMC to extract maximum of power of PV modules. The WT depends to a Permanent Magnet Synchronous Generator (PMSG) and a controlled rectifier to maximize the operating point of the WT. Therefore, the battery and supercapacitors are connected to the system via a bidirectional converter to ensure the charging modes operations. The voltage source converter (VSC) is controlled by SMC had an aim to control the frequency and amplitude of the system outputs injected into the grid and load utility. To expose the performance of the proposed sliding mode control approach, we propose a different comparatives tests include a conventional PI control and hysteresis current control under two operating grid modes depending to different climate scenarios. The developed control strategy for HRES have shown great results such as the control performance, effectiveness and delivers low harmonic distortion values (Max of THD are 3,43% and 4,88%) compared with hysteresis current controller (Max of THD are 4,63% and 10,98%) and PI control (Max of THD are 4,92% and 12,86%) featuring OFF and ON grid, respectively. Finally, the proposed control enhances the HRES quality and stability.

364

Authors: Eman K. Ibraheem, Ekhlas K. Hamza

The increasing number of mobile devices challenges the current radio frequency (RF) networks. A hybrid RF/VLC network (HLRN) is proposed to mitigate the spatial fluctuation of data rate, offering a system throughput greater than that of standalone visible light communications (VLC) or radio frequency (RF) networks. In hybrid networks, the main problem is load balancing, which reduces network performance. Therefore, to solve this problem, the load balancing (LB) schemes in HLRNs are studied and focus on the AP assignment for the users. In that, efficient spectrum sensing is found essential to prevent the wrong holes detection in the band using five technologies divided into three stages, namely: Received signal strength (RSS), particle swarm optimization (PSO) and deep learning methods: Feed forward neural network (FFNN), convolutional neural network (CNN) and cascade back propagation neural network (CBPNN). An event such as high energy noise availability or primary user fluctuation (mobility) impact on the throughput is studied, which was carried out in three stages depending on MATLAB. In the first stage, the received signal strength (RSS) algorithm is proposed, which is used the SNR to calculate throughput for fixed and mobility user and user satisfaction. In stage two, the particle swarm optimization (PSO) algorithm is proposed to achieve a better performance than the RSS algorithm. At last, in stage three, the DL algorithms (FFNN, CNN, CBPNN) are used to enhance the performance's accuracy and compare them. The results in stage one show that throughput using RSS technique is decreasing when the number of users increases by achieving 390 Mbps approx. is detected with single user existence while 180 Mbps approx. is detected with 10 users' existence. And it can be noted that most of the users are connected to wireless fidelity (Wi-Fi) AP. Hence the Wi-Fi is overloaded. Further, only 34 % of the users in RSS based scheme will achieve the desired performance if a user satisfaction threshold of 0.5 is considered for the system. The results in stage two shown that throughput using the PSO technique is outperformed over the others by achieving high average throughput of 960 Mbps approx. Also, it has offered acceptable load balancing for the access points with good user satisfaction reach to 80 % of the users will achieve the desired performance when the user satisfaction threshold is considered to be 0.9 for the system. In stage three, deep learning methods for system performance enhancement are used (as compared with the standard RSS method). throughput is improved, and this improvement is lesser than that in the case of the PSO algorithm. 800 Mbps approx. is achieved using the CNN method, and it is realized that 70 % of the users will achieve full user satisfaction when the user satisfaction threshold is considered 1.2 for the system.

392

Authors: Mina Ibrahim, Muhammad Mahmoud, Reda M. Albadawy, Hatem Abdulkader

Liver cancer is a common type of cancer that causes death, because there are no noticeable symptoms at an early stage, as this disease is not detected in most patients until cancer has reached the advanced stage only. Researchers are developing algorithms that doctors can use to detect liver tumours early by examining images of tissue from a biopsy or an abdominal medical image. The tissue expert must put in the time and effort required at this stage to determine whether or not this tumour is cancerous and in need of treatment. This model can then be used by a histology expert to make an initial diagnosis. Convolutional neural networks (CNNs) are employed in this paper to propose a novel combination of deep learning models that can transfer information from previously trained global models. After that, this information was decanted into a solo model to improve our model approaches to increase the performance in time and accuracy with tuning to an encoder, decoder, shortcuts, and skip connections to custom convolution layers for three classes such as background, origin, and Tumour shape. Regarding semantic segmentation, our model has proved to be a highly effective way to make results more accurate and valuable to assist in the diagnosis of the liver Tumour using CT scans. As a result, we were able to develop a hybrid model that is capable of recognizing CT images of a liver tumour. Our research yielded the greatest possible results, which we got, reaching 99.50% accuracy, a 86.40% precision and a recall of 97.90%. This accuracy with multi-class is higher than that obtained using other previous models that obtained the best accuracy of 0.991 during the annual periodic examination campaigns for liver cancer detection. Using this model, experts in this field can save time and effort while becoming more informed choices. It also keeps the time and effort that would otherwise be required to administer this treatment, especially during the annual examination campaigns.

418

Authors: B. Uma Rani, A. R. Naseer

Underwater Acoustic Wireless Sensor Networks (UAWSN) is a collection of sensor nodes that are strategically placed at different depths under water to monitor the activities of interest. In contrast to Terrestrial Wireless Sensor Networks (TWSN), a UAWSN is constrained by several challenges like limited bandwidth, multipath propagation, path-loss, noise, Doppler spread, etc. Also, other parameters like salinity, depth and pressure influence the propagation speed and in turn the achieved throughput. Most of the existing works in this direction focus on designing new medium access control protocols to minimize collisions and maximize bandwidth utilization. Other existing routing and data forwarding protocols use different metrics to maximize network objectives like network lifetime, increase network throughput, minimize energy drain, etc. In this paper, we propose a routing and data forwarding mechanism using link bandwidth by considering path-loss and multipath propagation effects. These cross-layer metrics allow us to select high-capacity links increasing overall network throughput. In addition, we propose different optimizations to the data forwarding and routing mechanisms to save energy and increase lifetime of the network. The simulation results show that the proposed mechanism improves 6% packet delivery ratio, 10% throughput, and minimizes energy consumption in comparison to state-of- the-art protocol CLOR.

443

Authors: Mohammed Kh. Al-Nussairi, Sadeq D. Al-Majidi, Ali Jber Mshkil, Adel Manaa Dakhil, Maysam F. Abbodm, Hamed S. Al-Raweshidy

An Automatic Generation Control (AGC) is considered a substantial stage in power systems to ensure that the area-frequency response and the tie-line power of an interchanging steady state errors are acceptable values especially at the transient state. Whiles, several techniques have been designed for the AGC, Fuzzy logic control is commonly used. However, it is faced a longer processing time issues to adjust the accurate single in the transient state. In this paper, a Single Input Fuzzy Logic Gain Scheduling PID Controller (SIFL-PID) is designed as a supplementary loop for AGC of two area interconnected power system to reduce the processing time. The lower number of rules and the simple parameters of the SIFL-PID will contribute to minimise the design time required for the Fuzzy Logic Gain Scheduling FL-PID. The SIFL-PID also reduced the signed distance technique when it is generated one input variable simple mathematical model of the system. In comparison with the FL-PID, the number of rules controlling system can be significantly reduced. The main outcome of this work is that it is assessed the effectiveness of the SIFL-PID for two power systems that are interconnected when it is compared with the FL-PID using the MATLAB/SIMULINK environment. The results prove that, the simple design of the SIFL-PID control is validated under various condition tests. Performance analysis standards are also performed. The results of all simulations demonstrate that the suggested controller provides better performances such as like as minimum Performance analysis standards, frequency deviation power deviations and low overshoot and undershoot.

465

Authors: Nguyen Dinh Hung, Le Cao Quyen, Le Van Dai

The primary goal of this paper is to develop and apply an adaptive fuzzy (AF) based on the proportional integral derivative (PID) controller for the automatic voltage regular (AVR) system in the synchronous generator (SG) to improve output voltage stability. This proposed AF-PID is applied to ameliorate the shortcomings of the traditional PID controller by determining the optimal coefficients for the controller. Firstly, the initial coefficients of the traditional PID controller are calculated by using the Ziegler-Nichols method. After that, it develops the AF algorithm based on the fuzzification, rules of the fuzzy controller, and defuzzification of the voltage signals for self-tuning the coefficients to add the traditional PID controller. The transient response of the system is simulated and analyzed by considering the different step changes in reference voltage. The efficacy of the proposed AF-PID controller for the AVR system in the SG is illustrated by comparing it with traditional PID and without the controller based on the obtained results from the case studies. It can be concluded that when applying the proposed method, the transient response characteristics of the system from the perspective of simulating via the time domain under any sudden load changes in system operation, such as the rise time, steady-state error, max peak overshoot, and settling time are greatly improved. The simulation has been performed using the MATLAB/Simulink program.

492

Authors: Dwi Sunaryono, Riyanarto Sarno, Joko Siswantoro, Diana Purwitasari, Shoffi Izza Sabilla, Rahadian Indarto Susilo, Naufal Rafi Akbar

Epilepsy is a common chronic brain disease caused by abnormal neuronal activity and the occurrence of sudden or transient seizures. Electroencephalogram (EEG) is a non-invasive technique commonly used to identify epileptic brain activity. However, visual detection of the EEG is subjective, time consuming, and labour intensive for the neurologist. Therefore, we propose an automatic seizure detection using a combination of one-dimension convolution neural network (1D-CNN) with majority voting and deep neural network (DNN). EEG signals features are extracted using discrete Fourier transform (DFT) and discrete wavelet transform (DWT) which then these features will be selected with XGBoost to minimize features classified with CNN. The proposed method experimental results show that it can detect epilepsy from EEG signals perfectly with an accuracy of 100%. However, the proposed method only yielded classified EEG signals from the University of Bonn Dataset as its results. The performance of the suggested approach might not be similar to other EEG datasets.

511

Authors: G. Krishna Raju, P. Padmanabham, A. Govardhan

Introduction of the World Wide Web (WWW) and progressions in the field of multimedia and computer technology have enlarged the amount of image collections and databases including art galleries, digital libraries, and medical imageries that extend to millions of images. From these large-scale datasets, the retrieval procedure of images is carried out by classical approaches like chi square distance, colour histogram and text-based image retrieval, which however requires more time for getting the desired images. Thus, there is a need of proposing an effective retrieval system for images by handling these large numbers of images. For this reason, a new Content-Based Image Retrieval (CBIR) system has been implemented in this paper for retrieving the desired images of users from the collected images. Initially, different kinds of images like medical images, texture images, and environmental images etc., are collected from the standard image database. The deep feature extraction using Visual Geometry Group network (VGG-16), Inception v3, and Xception are used to get the features of all the images in the database. The parameters in these three deep learning techniques are tuned with an enhanced optimization algorithm of Modified Bypass-Rider Optimization Algorithm (MB-ROA). The features from VGG16, Inception and Xception are used for optimal weighted fused feature selection using enhanced ROA under training phase. During testing with query images, it undergoes deep feature extraction with the same set of techniques and, then fed into the optimal weighted fused feature selection to get the features of query image. Then, the multi-similarity function considering the cosine, Euclidean and Jaccard similarity are checked between the optimal features of database images and query image. The database images with the minimum similarity with the query image are retrieved from the database. The overall precision and recall to retrieve the top ten images related to the provided query image from the Corel dataset are 77.25% and 76.89%, respectively, while 70.15% and 70.26%, respectively from the VisTex dataset. The experimental findings show that the recommended model is effective at retrieving images from the database.

541

Authors: Yunifa Miftachul Arif, Hani Nurhayati

One of the challenges of learning mathematics is making students like the subject to understand the material being taught more efficiently. Therefore, several studies propose the use of gamification elements in the mathematics learning process. Another important consideration in learning mathematics is the accuracy of the subject matter given to students. This study proposes a metaverse-based mathematics pedagogy media (MMPM) equipped with learning material selection (LMS) to adaptively select subject matter according to the level of student knowledge. We use the multi-criteria recommender system (MCRS) to support LMS in providing recommendations for determining the choice of subject matter. The method used to calculate similarity is cosine-based, while the ranking is based on average, worst-case, and aggregate. We built the MMPM system using the unity game engine in the experimental stage. At the same time, the subject matter that becomes the focus of scenario visualization is about beam, cube, prism, and pyramid. The test results show that the LMS system works well in carrying out its duties by adaptively choosing the appropriate subject matter scenarios for students based on their pre-test results. In this study, MCRS-based LMS produces the highest accuracy of 92% for two to three input items and the lowest 90% for four input items.

563

Authors: Hussein A. Hussein Al-Delfi, Fadhil Sahib Hasan

This paper suggests the transmission of secure speech signals through the MIMO-OFDM system over a doubly selective fading channel. The two-dimensional chaotic map like logistic (TDLM), standard (TDSM), and triangle function combination discrete chaotic map (TD-TFCDM) are used to generate pseudo-random bit generator (PRBG) and index permutation (IP) to encrypt the information based on ciphering and scrambling techniques. Different speech quality and residual intelligibility measures including log-likelihood ratio (LLR), frequency-weighted segmental signal-to-noise ratio (fwSNRseg), signal-to-noise ratio loss (SNRLOSS), correlation coefficient (CC) and spectrogram measures as well as Bit Error Rate (BER) performance are used to test the performance of the proposed security system. The simulation results show that the speech signal over the MIMO-OFDM system behaves well with high security. Also, the speech encryption system using TDSM and TD-TFCDM outperforms TDLM in most residual intelligibility measures.

589

Authors: Mie Mie Oo, Nu War

Acoustic scene classification is a difficult issue among artificial intelligence, signal processing, and machine learning. Scene recognition performance has a robust relation with feature learning using deep convolutional networks. In the following research, end-to-end deep residual network embedded channel attention is explored to learn the discriminative features from the audio scene. Log-Mel spectrogram is obtained from input raw audios. It is forwarded to proposed attention network. An extracted feature layer is concatenated with the SoftMax classifier in the proposed attention network. The experimentation is carried out on Detection and Classification of Acoustic Scenes and Events (DCASE) 2016 and 2017 datasets. The proposed channel-attention-based residual network achieves classification results with an average accuracy of 80.27% and 80.82%, respectively.

609

Authors: Saritha Ibakkanavar Guddappa, Rajeshwari Mahabhaleshwar Hegde

Cyber physical system (CPS) is an engineered system that integrates the physical world with the cyber world through enhanced computation, communication and control (3C) abilities. But, the CPS’s sensors are susceptible to malicious threats due to its dynamic topology and open medium of the network. Therefore, an effective secure adaptive cluster-based routing is required to be developed for improving the security of CPS. In this paper, the multi objective-trust based hybrid optimization algorithm (MO-THOA) is proposed to improve CPS security. The MO-THOA is the combination of trust based flower pollination algorithm (TFPA) and trust based ant colony optimization (TACO). In that, the TFPA is used to select a secure adaptive cluster head (ACH) and TACO is used to identify the secure routing path via ACHs. Therefore, the proposed MO-THOA improved the security against malicious attacks while broadcasting the data. The performance of the MO-THOA is evaluated using packet delivery ratio (PDR), average end to end delay (AEED) and normalized routing overhead (NRO). The existing methods namely grey-wolf updated whale optimization algorithm (GU-WOA), moving centroid based routing protocol (MCRP) and multi-objective ant colony optimization (MOACO) are used to evaluate the MO-THOA method. The PDR of the MO-THOA for 20 nodes is 0.981, it is high when compared to the GU-WOA, MOACO and MCRP.

627

Authors: Sara Ahmed, Fatma A. Omara

Cloud computing has gained many attentions worldwide. Workflow systems become a significant method for develop scientific applications. Therefore, workflow scheduling is considered one of the most important issues in cloud computing. It concerns about mapping tasks on cloud resources (i.e., Virtual machines (VMs)), to improve scheduling performance. Because the existing heterogeneous earliest finish time (HEFT) algorithm is considered one of the best and simplest algorithms, many algorithms have been proposed to improve the performance of the HEFT algorithm. According to our previous work, a modification has been done to HEFT algorithm to enhance the performance, called modified heterogeneous earliest finish time (M-HEFT). The goal of M-HEFT algorithm is to reduce make span, maximize resource utilization and increase load balance. According to the work in this paper, an enhancement has been added to our previous M-EFT algorithm to reduce the tradeoff among make span, resource utilization, and load balance, called enhanced modified heterogeneous earliest finish time (EM-HEFT). The enhanced EM-HEFT algorithm consists of two phases; task prioritization and task-VM mapping. In task prioritization phase, a priority will be provided to each task in directed acyclic graph (DAG) by introducing new factors in priority value to be more aware about task requirements. According to task-VM phase, tasks are allocated to resources as in our previous M-HEFT algorithm. To evaluate the performance of the proposed EM-HEFT algorithm, a comparative study has been done among the proposed algorithm and four existed algorithms (HEFT, Efficient scheduling algorithm use critical path and static level attribute, Optimized Min-Min (OMin-Min) and our previous M-HEFT). The experimental results show that the proposed algorithm outperforms other algorithms by minimizing make span by 25%, improving resource utilization by 43% and load balance by 14% in average.

657

Authors: Bakr S. Shihab, Hadeel N. Abdullah, Layth A. Hassnawi

Recently, path planning algorithms have been one of the primary and important functions of unmanned aerial vehicles (UAVs). Path planning algorithms in UAVs focused on path length, average path length, computation time, and standard deviation from the mean path length. In spite of this, it faced many difficulties and problems, such as many obstacles, path segmentation, and the increasing number of obstacles and paths in urban environments. This work proposes polynomial functions for path planning and obstacle avoidance. Since it enables us to plan the path in static internal environments, it enables us to plan the path quickly and with less computing time because it does not require high memory and does not require pre-compute of the path. Instead, the route is plotted in real time, Where the appropriate equation is entered into the program, so that the vehicle follows the curve of the entered equation. An accurate data set and metrics were used to measure the efficiency of the proposed method. The experimental results showed a clear improvement in the work of the polynomial function on A*, PSO and genetic algorithms, as this improvement appears very clearly when compared to the computing time, which was reduced by 15% in the method of polynomial functions where the path calculation took only parts of The second, as well as the path length was halved in the polynomial method as the results showed, which reduces the time of battery and memory consumption, the cost of calculating the path and the time to reach the goal.

682

Authors: Ahmed Wasif Reza, Khairum Islam, Sidratul Muntaha, Omair Bin Abdur Rahman, Rabeya Islam, Mohammad Shamsul Arefin

The emergence of new and improved technological advances created severe problems in the security state of the educational certification system. Throughout this paper, a proposal has been made to improve security. Here, Blockchain technology has been introduced as reliable secure storage for the educational certification system, providing an additional facility to the users. That is the validation and authentication of the student’s academic records. Moreover, for security purposes, Blockchain technology can replace the traditional academic certification system and contribute to a new model for sharing student information. After completion of data inclusion and hashing, the blocks will be inserted into the Blockchain network. This proposed model enhances document security and fraud reduction and additionally reduces a significant amount of authentication time almost up to double the current speed. With this system, we will get a certification process in which all data will be digitalized and secured in an unbreakable database with proper authentication and with a noticeable amount of time efficiency.