Search

As symbolic music generation evolves, research interest is shifting toward more controlled and steerable generative processes to support creative decisions. Previous methods focus on global conditioning or fine-grained control through input sequences but often limit flexibility for real-time interventions and require modifications to the model’s architecture. We introduce a novel symbolic music generation framework by combining a Transformer encoder-decoder with probe models, which enable us to interpret the encoder hidden state using pre-defined non-differentiable musical features, and subsequently manipulate the hidden state to achieve a set of desired attributes in the generated music. This method allows fine-grained control over specific musical features without altering the underlying model architecture. Probes can be trained jointly with the generative model or applied post-training, enabling adaptable control without retraining the model. Our experiments demonstrate that this intervention effectively influences the model output without hindering the music quality. This approach enhances both the flexibility and interpretability of symbolic music generation, enabling better real-world applicability for music generation models.

For an extended period, a technological architecture known as cloud IoT links IoT devices to servers located in cloud data centers. Real-time data analytic are made possible by this, enabling better, data-driven decision making, optimization, and risk reduction. Since cloud systems are often located at a considerable distance from IoT devices, the rise of time-sensitive IoT applications has driven the requirement to extend cloud architecture for timely delivery of critical services. Balancing the allocation of IoT services to appropriate edge nodes while guaranteeing low latency and efficient resource utilization remains a challenging task. Since edge nodes have lower resource capabilities than the cloud. The primary drawback of current methods in this situation is that they only tackle the scheduling issue from one side. Task scheduling plays a pivotal role in various domains, including cloud computing, operating systems, and parallel processing, enabling effective management of computational resources. In this research, we provide a multiple-factor autonomous IoT-Edge scheduling method based on game theory to solve this issue. Our strategy involves two distinct scenarios. In the first scenario, we introduced an algorithm containing choices for the IoT and edge nodes, allowing them to evaluate each other using factors such as delay and resource usage. The second scenario involves both a centralized and a distributed scheduling approach, leveraging the matching concept and considering each other. In addition, we also introduced a preference-based stable mechanism (PBSM) algorithm for resource allocation. In terms of the execution time for IoT services and the effectiveness of resource consolidation for edge nodes, the technique we use achieves better results compared with the two commonly used Min-Min and Max-Min scheduling algorithms.

Generative adversarial networks (GANs) have drawn considerable attention in recent years for their proven capability in generating synthetic data which can be utilised for multiple purposes. While GANs have demonstrated tremendous successes in producing synthetic data samples that replicate the dynamics of the original datasets, the validity of the synthetic data and the underlying privacy concerns represent major challenges which are not sufficiently addressed. In this work, we design a cascaded tabular GAN framework (CasTGAN) for generating realistic tabular data with a specific focus on the validity of the output. In this context, validity refers to the the dependency between features that can be found in the real data, but is typically misrepresented by traditional generative models. Our key idea entails that employing a cascaded architecture in which a dedicated generator samples each feature, the synthetic output becomes more representative of the real data. Our experimental results demonstrate that our model is capable of generating synthetic tabular data that can be used for fitting machine learning models, as CasTGAN’s classification performance only falls under the real training data’s PR-AUC score by 4.88% on average for classification datasets, and exhibits an average reduction of the real training data’s $R^{2}$ score by 0.139 for regression datasets. In addition, our model captures well the constraints and the correlations between the features of the real data, especially the high dimensional datasets. Assessing the generation of invalid records demonstrates that CasTGAN reduces the number of invalid data observations by up to 622% in comparison to the second best performing baseline tabular GAN model. Furthermore, we evaluate the risk of white-box privacy attacks on our model and subsequently show that applying some perturbations to the auxiliary learners in CasTGAN increases the overall robustness of our model against targeted attacks.

In this paper, an energy management system (EMS) using intelligent Lyapunov based adaptive fuzzy controller is designed for standalone microgrid having photovoltaic and wind turbines as primary sources, while battery storage system and diesel generator are available as secondary sources. Moreover, any remnant power available in the system is guided by the controller towards the dump load. The centralized controller based on intelligent adaptive fuzzy system helps to optimize the power flow in the system during high and low penetration of renewable energy resources under variable load conditions. Additionally, voltage regulation is carried out using nonlinear Lyapunov controller which kept the output voltages in desired range. The proposed control system helps in increasing the efficiency and reliability of standalone microgrid. Furthermore, the proposed system is validated by comparing the simulation and hardware results with the PID and sliding mode controller, which are consider as benchmark controller techniques for standalone microgrid. The outcome of the system validates the superiority of proposed techniques and achieved global asymptotic stability under variable load conditions.

Blood smear analysis is often used to diagnose diseases like malaria, Anemia, Leukemia, etc. Morphological changes, such as size, shapes, and color, are receiving much attention in pathological analysis. Existing methods for detecting, diagnosing and analyzing blood smears cannot quantify overlapped, irregular boundaries and complex structures. This work proposes and evaluates a framework that utilizes Generative adversarial networks (GANs) for the segmentation and classification of blood elements, that is, white blood cells (WBCs), red blood cells (RBCs), and platelets (PLTs) simultaneously. The Generator of the network determines the mapping from microscopic images of blood cells to a confidence map. This mapping stipulates the probabilities of the pixel of the microscopic blood cell images with respect to ground truth. The Discriminator of the network is essential to castigate the mismatch between the microscopic blood cells images and confidence map. Additionally, adversarial learning enables the Generator to generate a qualitative confidence map that is converted into segmented images. We have calculated minimum, maximum, and average losses to judge the performance of the proposed model. We measure structural similarity, peak signal-to-noise ratio, pixel classification error, and finally, classified cells. The proposed framework can analyze all the blood cell elements simultaneously. The proposed framework shows a significant improvement in the segmentation and classification of blood cell elements compared to state-of-the-art techniques. During the training process, generator total loss reduces by 12.18%, 5.39%, and 3.62% for RBCs, WBCs, and PLTs, respectively. Our results demonstrate that the proposed framework outperforms existing state-of-the-art techniques, achieving the highest pixel correctly classified (PCC) ratio for the segmentation of blood cells as 99.8%, 93.4%, and 99.9% for WBCs, RBCs, and PLTs, respectively. Our framework attains 95.45% and 88.89% classification accuracy for WBCs on ALL-IDB-I and ALL-IDB-II datasets. The dataset used for this study can be found at https://drive.google.com/drive/folders/1F7kZ1SRWUD9R6aHLMkj3wsjcHnvlGuwP?usp=sharing

Study is being conducted on training Generative Adversarial Networks (GANs) from 2D datasets to generate 3D human body avatars. Numerous applications, such as virtual reality, sports analysis, cinematography, surveillance, and more, have advanced significantly as a result of the promising research in this subject. Aerial photography sensors together with drone active tracking can remove occlusions and enable 3D avatar body reconstruction by avoiding obstacles and generating high-resolution, rich-information multi-view (RGB) photos. Training failures of 3D avatar reconstruction techniques lead to distortions and loss of features in 3D reconstructed models due to several reasons, including limited viewpoint coverage, visible occlusions, and texture disappearance. The recently developed end-to-end trainable deep neural network technique This work presents PIXGAN-Drone, a photo-realistic 3D avatar reconstruction system for the human body from multi-view photos. To create high-resolution 2D models, is predicated on integrating aerial photography sensors (a steady autonomous circular motion system) coupled with active tracking drones into the Pix2Pix GANs training framework. Accurate and realistic 3D models can be created with conditional image-to-image translation and dynamic aerial views. This study used tests on several datasets to show that our approach outperforms state-of-the-art approaches for a variety of metrics (Chamfer, P2S, and CED). Our 3D reconstructed human avatars in RenderPeople were 0.0293, 0.0271, and 0.0232; on People Snapshot (inside), 0.0133, 0.0136, 0.0050; on People Snapshot (outdoor), 0.0154, 0.0101, 0.0063; and on Custom data-drone (collected dataset), 0.0316, 0.0275, 0.0216.

Deep learning techniques have demonstrated significant capabilities across numerous applications, with deep neural networks (DNNs) showing promising results. However, training these networks efficiently, especially when determining the most suitable nonlinear activation functions, remains a significant challenge. While the ReLU activation function has been widely adopted, other hand-designed functions have been proposed. One such approach is the trainable activation functions. This paper introduces a novel neural network design, the SWAG. In this structure, instead of evolving, activation functions consistently form a polynomial basis. Each hidden layer in this architecture comprises k sub-layers that use polynomial activation functions adjusted by a factorial coefficient, followed by a Concatenate layer and a layer employing a linear activation function. Leveraging the Stone-Weierstrass approximation theorem, we demonstrate that utilizing a diverse set of polynomial activation functions allows neural networks to retain universal approximation capabilities. The SWAG algorithm’s architecture is then presented, where data normalization is emphasized, and a new optimized version of SWAG is proposed, which reduces the computational challenge of managing higher degrees of input. This optimization harnesses the Taylor series method by utilizing lower-degree terms to compute higher-degree terms efficiently. This paper thus contributes an innovative neural network architecture that optimizes polynomial activation functions, promising more efficient and robust deep learning applications.

Single-inductor 3x multilevel boost (SI-3xMLB) converters, recognized for their potential to supply multiple output voltage levels, are gaining much attention in microgrids due to their capability to diverse sub-systems with changing voltage requirements. However, some loads, such as constant power loads (CPLs), threaten the stability of SI-3xMLB converters due to their negative impedance properties. Thus, advanced regulation mechanisms are necessitated to reduce the destabilization effect of such loads in the microgrids. In response to the issue, in this paper, an adaptive data-driven controller with an observer is designed to mitigate the effect of non-ideal CPLs in dc/dc SI-3xMLB converters. The suggested data-driven controller can stabilize the system outputs using the input-output (I/O). In particular, the multi-agent fuzzy approximate Q-learning (MAFQ-learning) is employed to adjust the coefficients embedded in the data-driven controller dynamically. To do this, a reward function is defined according to the voltage control requirements of the dc/dc SI-3xMLB converter. By interacting the agent of MAFQ-learning with the environment, tunable control coefficients are adaptively adjusted. This process finds optimal policy, allowing the controller to dynamically respond to the effects of CPLs. The real-time responses based on the Arduino platform revealed the superior performance of the suggested controller to stabilize the dc/dc SI-3xMLB boost converter feeding CPL than the discrete time-averaged model predictive control (DTA-MPC), the optimal PI control based on state-action-reward-state-action (SARSA) technique.

In many real classification problems where a limited number of training samples is available, the linear classifiers based on discriminant analysis are unable to deliver accurate results. Moreover, the testing and/or the training data can be erroneous due to noise contamination which further degrades their performance. Regularization techniques become imperative to deal with these problems. However, the existing regularization techniques, to some extent, mainly focus on data scarcity issues but completely ignore the noisy nature of the testing and/or the training data. We propose a novel regularized quadratic discriminant analysis (R-QDA) classifier which addresses both issues simultaneously. The procedure involves a reformulation of the discriminant function of the conventional QDA classifier into least square problems and then solving them by using regularized least squares (Reg-LS) based on $\ell _{2}$ -norm. In contrast to existing R-QDA techniques, the proposed R-QDA classifier employs two regularization parameters pertaining to each class, which can be independently selected by various robust techniques. Numerical results demonstrate the effectiveness of the proposed method over classical R-QDA methods, especially in high noise regimes.

Air pollution is a critical global concern, demanding precise air quality forecasting to mitigate its severe consequences. Our study introduces Future Air Quality Prediction using Long Short-Term Memory based on Hyper Heuristic Multi-Chain Model (H2MCM) to project future air quality, considering various meteorological factors (MFs) and pollution-related variables like atmospheric pressure, temperature, humidity, and wind patterns. Leveraging 12 units of Long Short-Term Memory neural networks (LSTMs), H2MCM accurately predicts forthcoming air pollutants (APs) concentrations such as particulate matter with diameter $2.5 \; \mu $ m (PM2.5), carbon monoxide (CO), and nitrogen dioxide (NO2). Additionally, it accounts for spatiotemporal correlations between these APs and MFs, which significantly influence the air quality prediction for the next immediate time interval. H2MCM utilizes a multi-chain mechanism, employing 1-hour prediction models to forecast air quality hourly, enabling approximations for the next 12 hours. Also, for an efficient model selection, Akaike Information Criterion (AIC), Schwarz Bayesian Information Criterion (SBIC), Hannan-Quinn Information Criterion (HQIC), and corrected AIC (AICc) tools are used based on their ability to balance model fit and complexity. Furthermore, it demonstrates the ability to enhance the performance of any predictor. Experimental results substantiate H2MCM’s superiority over various models, including the Support Vector Regressor (SVR), Multi-Layer Perceptron (MLP), Recurrent Air Quality Predictor (RAQP), and Valchogianni models. H2MCM achieves impressive up to 75% better accuracy and consistency compared to SVR, 60% better than MLP, 38% better than RAQP, and 70% better than Valchogianni models.

In the paradigm of Device-to-Device (D2D) and System-on-Chip (SoC) communications, the optimization of high-speed data transfer while minimizing hardware resource utilization imposing paramount importance. Traditional methodologies have often resulted in undesirable outcomes, such as inflated area, latency, and power consumption. Consequently, this study adopts a focused approach by introducing an innovative Route-Controlling Network-on-Chip (RC-NoC) architecture, integrating critical components like FIFO-Buffer, crossbar switching, route control, and arbiter modules. The operational sequence begins with the FIFO-Buffer logic, which strategically manages data storage from diverse devices, allocating and organizing data flow based on distinct IP addresses. Subsequently, the route controller module orchestrates the operation of heterogeneous routers within the crossbar switching fabric, implementing a prioritized scheduling scheme. The arbiter plays a crucial role in overseeing and directing data flow from source to destination, leveraging request-level prioritization for efficient data transmission. Empirical evidence from simulations unequivocally demonstrates the superiority of the proposed RC-NoC architecture. Comparative analysis against essential state-of-the-art NoC architectures reveals significant enhancements in key metrics, notably area efficiency, latency reduction, and power optimization, with over 45% improvement observed across these metrics compared to GALS-NoC, F-NoC, and CE-NoC in the Orion framework. This robust validation substantiates the efficacy and technical prowess of the RC-NoC design paradigm in addressing the challenges of modern D2D and SoC communication paradigms. Furthermore, the optimization process notably decreased routing costs for different packets, with Router2 exhibiting a 40% reduction in Packet1’s routing cost and Router3 displaying a 44% decrease in Packet2’s cost. Moreover, the attained router capacities demonstrated an average increase of 17%, significantly enhancing network efficiency.

Deep neural networks (DNNs), while powerful, often suffer from a lack of interpretability and vulnerability to adversarial attacks. Concept bottleneck models (CBMs), which incorporate intermediate high-level concepts into the model architecture, promise enhanced interpretability. This study delves into the robustness of Concept Bottleneck Models (CBMs) against adversarial attacks, comparing their original and adversarial performance with standard Convolutional Neural Networks (CNNs). The premise is that CBMs prioritize conceptual integrity and data compression, enabling them to maintain high performance under adversarial conditions by filtering out non-essential variations in input data. Our extensive evaluations across different datasets and adversarial attacks confirm that CBMs not only maintain higher accuracy but also show improved defense capabilities against a range of adversarial attacks compared to traditional models. Our findings indicate that CBMs, particularly those trained sequentially, inherently exhibit higher robustness against adversarial attacks than their standard CNN counterparts. Additionally, we explore the effects of increasing conceptual complexity and the application of adversarial training techniques. While adversarial training generally boosts robustness, the increment varies between CBMs and CNNs, highlighting the role of training strategies in achieving adversarial resilience.

For information technology (IT) projects to be successful, it is imperative to identify the critical success factors (CSFs) that are essential to the IT project management process. This study aims to illuminate these CSFs by incorporating insights from professionals involved in the management of IT projects. The major contribution of this study lies in classifying these factors into different categories and quantifying their impact while taking IT project management experts’ viewpoints into account. To accomplish this, a thorough assessment of the literature was carried out, which resulted in the identification of 20 success factors that fall into five categories: project team capability, top management support, project quality, project management, and organizational process assets (OPAs). A survey covering these variables was then sent to professionals in the IT industry, and responses were obtained from five experts. The collected data was analyzed using the fuzzy analytic hierarchy process (FAHP) method, which mitigates imprecision and avoids bias in the data. The most important factor was determined to be project team capability, whereas organizational process assets were deemed to be the least significant factor. It is anticipated that the insight garnered from this study will help IT project management professionals in supporting, evaluating, and improving project success.

Software-Defined Networking (SDN) is a groundbreaking technology that has transformed network management significantly. By integrating data and control, SDN offers unparalleled flexibility and responsiveness, thereby overcoming the limitations of conventional network architectures. However, a centralized controller, which is a hallmark of SDN, is a double-edged security sword that offers easy control. This also becomes a dangerous point of failure for the entire network. To the best of our knowledge, this is the first comprehensive study to explore traditional-based, artificial intelligence (AI)-based, and moving target defense (MTD) approaches to securing SDN. The study begins with a survey of traditional security solutions for SDN, encompassing authentication, authorization, encryption, security protocols, firewalls, and flow verification, by addressing security threats and vulnerabilities in both data and control planes. The study then investigates the application of AI-based security solutions in an SDN environment, focusing on how Machine Learning (ML) and Deep Learning (DL) techniques are leveraged to address advanced security threats. Additionally, the survey examines MTD mechanisms within data and control plane security. Several in-depth techniques, including the randomization of Internet Protocol (IP) and Media Access Control (MAC) addresses, port numbers, and flow tables, and delving into the relationship between security threats, MTD strategies, and the specific controllers employed in experimental implementations. We utilized the widely recognized STRIDE cybersecurity framework to systematically identify and evaluate the potential threats to SDN security. Our analysis resulted in a comprehensive list of security challenges, and we propose future research directions aimed at addressing emerging threats in both the data and control planes.

Increased traffic density, combined with global population development, has resulted in increasingly congested roads, increased air pollution, and increased accidents. Globally, the overall number of automobiles has expanded dramatically during the last decade. Traffic monitoring in this environment is undoubtedly a significant difficulty in various developing countries. This work introduced a novel vehicle detection and classification system for smart traffic monitoring that uses a convolutional neural network (CNN) to segment aerial imagery. These segmented images are examined to further detect the vehicles by incorporating novel customized pyramid pooling. Then, these detected vehicles are classified into various subcategories. Finally, these vehicles are tracked via Kalman filter (KF) and kernelized filter-based techniques to cope with and manage massive traffic flows with minimal human intervention. During the experimental evaluation, our proposed system illustrated a remarkable vehicle detection rate of 95.78% over the Vehicle Aerial Imagery from a Drone (VAID), 95.18% over the Vehicle Detection in Aerial Imagery (VEDAI), and 93.13% over the German Aerospace Center (DLR) DLR3K datasets, respectively. The proposed system has a variety of applications, including identifying vehicles in traffic, sensing traffic congestion on a road, traffic density at intersections, detecting various types of vehicles, and providing a path for pedestrians.

Within the distribution system operator (DSO) framework, there could be alternative arrangements to enable market participation by all entities, including the privately owned distributed energy resources (DERs). In many cases, the retailers and similar aggregating agencies facilitate this. Even before finalizing the modalities of the market at the distribution level, it is essential to know about the possible market outcome and its impact on the operational decisions of DSO and vice-versa. This paper presents a simulation framework and associated case studies, where the market outcome and DSO’s operational activities are carried out in tandem to establish a realistic and feasible outcome. The market model is represented by a non-cooperative game where an efficient and fair solution is obtained using Nash equilibrium. For this, multi-player power transaction problem (MPTP) is solved using complementarity modeling. The proposed MPTP is further compounded with network reconfiguration and loss considerations which are the part of DSO’s operations. The inclusion of reconfiguration and loss allocation to entities makes the outcome realistic and feasible one. The proposed simulation framework is implemented on various test cases with a heterogeneous set of participants to establish its effectiveness.

The stability related issues may occur in a power system due to disturbances in generating or loading conditions, especially in the presence of distributed generation (DG) based on renewable energy resources (RERs). This paper proposes a novel strategy for the stability enhancement of a wind power generation system (WPGS) by using a combination of three devices, namely, a power system stabilizer (PSS), resistive superconductor fault current limiter (R-SFCL) and static synchronous compensator (STATCOM). The small signal (SS) stability of the test system is enhanced by selecting the best PSS type from the different types of PSS. An R-SFCL is used for improving the rotor angle and the frequency stability of the test system. Two indices, namely, transient stability index and sum of maximum deviations (SMD) index are introduced for determining the optimal locations of different sized R-SFCLs for increasing the rotor angle stability. The sensitivity index (SI) based on the power change between areas is applied for determining the optimal locations of different sized R-SFCLs for enhancing the frequency stability. Along with rotor angle and frequency stability, LVRT capability improvement of the wind farm using STATCOM is also considered. Finally, the combined effect of R-SFCL and STATCOM on the rotor angle and the frequency stability, for different fault locations, is also investigated for determining the optimal location of an R-SFCL in the presence of STATCOM. The results presented in the paper show that STATCOM affects both the number of feasible locations and the optimal locations that can be selected for different sized R-SFCLs for augmenting the rotor angle and the frequency stability of the system during faults. Moreover, it is pointed out that an optimal combination between the different sizes and the locations of R-SFCLs and STATCOM exists to enhance the overall stability of the test system under fault conditions.

A digital twin is a powerful tool that can help monitor and optimize physical assets in real-time. Simply put, it is a virtual representation of a physical asset, enabled through data and simulators, that can be used for a variety of purposes such as prediction, monitoring, and decision-making. However, the concept of digital twin can be vague and difficult to understand, which is why a new concept called “capability level” has been introduced. This concept categorizes digital twins based on their capability and defines a scale from zero to five, with each level indicating an increasing level of functionality. These levels are standalone, descriptive, diagnostic, predictive, prescriptive, and autonomous. By understanding the capability level of a digital twin, we can better understand its potential and limitations. To demonstrate the concepts, we use a modern house as an example. The house is equipped with a range of sensors that collect data about its internal state, which can then be used to create digital twins of different capability levels. These digital twins can be visualized in virtual reality, allowing users to interact with and manipulate the virtual environment. The current work not only presents a blueprint for developing digital twins but also suggests future research directions to enhance this technology. Digital twins have the potential to transform the way we monitor and optimize physical assets, and by understanding their capabilities, we can unlock their full potential.

Many face-recognition (FR) methods have been proposed thus far. Although FR has achieved wisdom in square pixel-based image processing (SIP) due to many studies, this wisdom has not been transferred to Hexagonal pixel-based image processing (HIP) until now. This study presents HIP versions of the most basic texture extraction studies in SIP, namely Gray-Level-Co-occurrence-Matrices (GLCM), Local Binary Pattern (LBP), and our recent work, local-holistic graph-based descriptor (LHGPD). The images are first transformed from the SIP domain to the HIP domain. The HIP domain equivalences (HexGLCM, HexLBP, and HexLHGPD) of the SIP domain GLCM, LBP, and LHGPD are then established. Finally, the facial recognition performances of the SIP and HIP domain versions of GLCM, LBP, and LHGPD are evaluated and compared on the primary data sets. The results of the experiments reveal that HIP domain GLCM, LBP, and LHGPD show a par performance, surpassing them in places when compared to their counterparts in the SIP domain regarding face recognition accuracy.

The Internet of Things (IoT) is becoming increasingly popular in areas like wearable communication devices, biomedical devices, and home automation systems. IoT-compatible processors or devices need larger integrated memory circuits, like static random access memory (SRAM). The design of such a larger memory with fast access times and low leakage is a challenge. This article proposes a 7T SRAM cell using an InGaAs-dual pocket-dual gate-tunnel FET (InGaAs-DP-DG-TFET) device. We have compared the key metrics of SRAM, such as read/write delay, stability, and power consumption, with 7T SRAM based on existing TFET devices. The 7T cell designed in this work significantly improved read and write access times. The proposed SRAM shows a decrement of read delay (write delay) by approximately 10x (210x) and 2.5x (43x) times compared to the 7T SRAM cell implemented using Si-DP-DG-TFET and InGaAs/Si-DP-DG-TFET at $V_{DD}$ = 700 mV. The proposed cell has the highest write margin (WM) and the least leakage power compared to the other cells examined here. The WM of the proposed 7T SRAM cell is 5x and 1.6x higher at $V_{DD}$ = 500 mV, compared to Si-DP-DG-TFET and InGaAs/Si-DP-DG-TFET counterparts, respectively. We computed the leakage power of the 7T SRAM cell since it is an essential factor in the design of SRAM. To explore the performance of TFET-based SRAM operation, we have carried out circuit simulations at various supply voltages and for different cell sizes. These simulations also offered an assessment of the cell’s dynamic performance. We employed a Verilog-A-based lookup table technique and ran simulations in the commercial HSPICE circuit simulator. The low leakage power and fast memory access of the proposed cell make it a promising candidate for IoT applications.

A large number of sensors are deployed for performing various tasks in the smart cities. The sensors are connected with each other through the Internet that leads to the emergence of Internet of Things (IoT). As the time passes, the number of deployed sensors is exponentially increasing. Not only this, the enhancement of sensors has also laid the base of automation. However, the increased number of sensors make the IoT networks more complex and scaled. Due to the increasing size and complexity, IoT networks of scale-free nature are found highly prone to attacks. In order to maintain the functionality of crucial applications, it is mandatory to increase the robustness of IoT networks. Additionally, it has been found that scale-free networks are resistant to random attacks. However, they are highly vulnerable to intentional, malicious, deliberate, targeted and cyber attacks where nodes are destroyed based on preference. Moreover, sensors of IoT network have limited communication, processing and energy resources. Hence, they cannot bear the load of computationally extensive robustness algorithms. A communication model is proposed in this paper to save the sensors from computational overhead of robustness algorithms by migrating the computational load to back-end high power processing clusters. Elephant Herding Robustness Evolution (EHRE) algorithm is proposed based on an enhanced communication model. In the proposed work, 6 phases of operations are used: initialization, sorting, clan updating, clan separating,selection and formation, and filtration. These process collectively increase the robustness of the scale-free IoT networks. EHRE is compared with well-known previous algorithms and is proven to be robust with a remarkable lead in performance. Moreover, EHRE is capable to achieve global optimum results in less number of iterations. EHRE achieves 95% efficiency after 60 iterations and 99% efficiency after 70 iterations. Moreover, EHRE performs 58.77% better than Enhanced Differential Evolution (EDE) algorithm, 65.22% better than Genetic Algorithm (GA), 86.35% better than Simulating Annealing (SA) and 94.77% better than Hill climbing Algorithm (HA).

Behavior Driven Development (BDD) is a widely adopted agile methodology for software development that emphasizes the behavior of an application as a series of test cases, using the keywords, which include “Given,” “When,” and “Then.” It involves writing requirements in a structured and testable format that can be evaluated to ensure compliance with the expected behavior. Although a significant amount of research has been conducted to examine the impact of using BDD on software development process yet rare work is observed to synthesize these studies and identify areas for future exploration. This study presents a review of the state-of-the-art BDD by synthesizing the recent advancements in its uses and applications. It aims to systematically investigate the impact of BDD on software development process as well as on product quality by aiding to bridge the communication gap between the stakeholders. The results reveal that BDD is an effective technique to clarify requirements during the software development process as it helps minimizing the intrinsic ambiguities. This work proposes a taxonomy based on the role and applications of BDD in various contexts. It suggests a framework for applying BDD in software development and defines a workflow for its application in software development. Finally, this work highlights some pertinent future directions for the use of BDD in software development.

This research work proposes an Improved Supervised Learning (ISL)-based Deep Neural Network (DNN) for accurately forecasting the load demand of Electric Vehicles (EVs). This work incorporates Gated Recurrent Unit (GRU), Long Short Term Memory (LSTM), Recurrent Neural Network (RNN), Fully Connected (FC), and Convolutional Neural Network (CNN) architectures. The proposed ISL technique enhances prediction performance by refining the training process with additional features and information. Using a real-world EV charging dataset from Boulder City, USA, the simulations demonstrate consistent improvements in the GRU, LSTM, RNN, FC, and CNN models with the proposed ISL technique. Further, the proposed technique reduces the Normalised Root Mean Square Error (NRMSE) and Normalised Mean Absolute Error (NMAE) values. The accurate load demand predictions facilitated by the proposed models with ISL have significant implications for the planning and management of EV charging stations. This enables stakeholders to optimize resource allocation, effectively plan infrastructure capacity, and ensure the sustainable and reliable operation of grids in the face of increasing EV adoption. By leveraging deep learning architectures and incorporating the ISL technique, this research contributes to advancing load forecasting models for EVs, providing practical solutions for efficient management and planning in the evolving electric mobility landscape.

Access to reliable electricity is crucial for rural development and improving the quality of life in remote areas. Standalone photovoltaic PV systems and hybrid power systems HPS are promising solutions for rural electrification. However, frequency stabilization is a critical challenge in such systems, and conventional control techniques are inadequate. This work proposes a novel approach to optimize the PID controller gains for the standalone PV-based isolated HPS (IHPS). The student psychology-based optimization algorithm (SPBOA) and quasi-oppositional-based whale optimization algorithm (QOWOA) are employed to enhance the performance of the IHPS models separately. The objective function considered is the integral of time absolute error (ITAE), and the frequency response profile is studied in the presence of the proposed SPBOA and QOWOA-based PID controllers. The results show that the proposed SPBOA outperforms the QOWOA in terms of reducing the ITAE by nearly 5% and minimizing the peak and settling time of the frequency response by 5%-7% in different scenarios. The transient responses of the systems to different input conditions are also analyzed, indicating that both the IHPS-I and IHPS-II models are feasible for remote rural electrification. Overall, the proposed approach offers an efficient solution to optimize the PID controller gains for frequency stabilization in standalone PV and HPS. The results demonstrate the benefits of using the SPBOA and QOWOA algorithms to enhance the performance of the systems, making them suitable for remote rural electrification.

Nowadays, data is ubiquitous and has a significant influence on our day-to-day activities due to the emergence of high speed internet and widespread use of sensor-enabled Internet of Things (IoT) devices. Owing to major improvement in sampling rate of sensors in recent years, a low variation of the sensed physical parameters is witnessed during a small time interval of observation, which in turn shows high correlation in time domain. Data-critical applications like personal healthcare monitoring, video surveillance, and other applications where data dropping creates significant barriers are attracted by high correlation data. However, due to their power-constrained nature, such applications do not benefit much from the transmission of redundant data. The ideal solution to this problem could possibly be achieved by adopting typical lossless source coding strategy with low-complex design. This paper presents a novel encoding scheme termed as Dual Run Distribution based Encoding (DRDE) scheme by exploiting high correlation of sensor data to suitably encode them using symbol run statistics, leading to a reduced length of data with a very large percentage of 0s. Employing silent symbol based communication, the transmitter can be kept in silent state during periods of the most dominant symbol ’0’ in the encoded messages and using a hybrid $FSK-ASK$ modulation/demodulation technique for communication with a non-coherent receiver results in a significant reduction in transmitter energy. We simulate the proposed sensor data encoding technique on real-life data with low-cost, low data-rate transceivers like CC2420. Simulation results show about 88% (theoretical) and 79-82% (practical) savings in transmitter and 12% (theoretical) 23.5% (practical) savings in receiver energy over conventional BFSK with real-life sensor dataset. Furthermore, our proposed method outperforms in terms of overall energy savings and reduction of $CO_{2}$ footprint, generating 1.48 – 0.041 mg/day, which is 78% lesser than conventional BFSK modulation scheme, making our proposed scheme suitable for sustainable IoT applications in WSN compared to existing schemes. Furthermore, we investigate the influence of various data compression algorithms on computation time, CPU power consumption, and transmission cost on an LPC2148 microcontroller built upon a 16-bit/32-bit ARM7TDMI chipset.

Semantic segmentation is a significant task in the field of Remote Sensing and Computer Vision. DeepLabV3+ is a convolutional neural network architecture that excels in the task of semantic segmentation, which involves assigning a class label to each pixel in an input image. This Paper proposes an Adaptive Deeplabv3+ model for semantic segmentation of Aerial Images, which combines Deeplabv3+ with the Improved Golden Eagle Optimization Algorithm (IGEO), to solve imprecise target segmentation and poor border segmentation accuracy. To enhance the quality of segmentation, Adaptive DeepLabV3+ employs atrous spatial pyramid pooling (ASPP) with multiple dilation rates in the encoder and allows the model to capture multi-scale context information efficiently, enabling it to distinguish between objects with varying scales. The proposed model effectively segmented the aerial images by optimizing the hyper-parameters such as hidden neuron count and learning rate. The suggested model achieved 98.46% accuracy, 96.32% correlation, 96.48% precision and 98.36% dice coefficient within the computation time of 136.8912 and 147.2684 seconds for dataset 1 and 2 respectively. Therefore the evolutional outcomes of the proposed model show significantly improved than the state-of-the-art techniques.

This article presents a comprehensive overview of the digital twin technology and its capability levels, with a specific focus on its applications in the wind energy industry. It consolidates the definitions of digital twin and its capability levels on a scale from 0-5; 0-standalone, 1-descriptive, 2-diagnostic, 3-predictive, 4-prescriptive, 5-autonomous. It then, from an industrial perspective, identifies the current state of the art and research needs in the wind energy sector. It is concluded that the main challenges hindering the realization of highly capable digital twins fall into one of the four categories; standards-related, data-related, model-related, and industrial acceptance related. The article proposes approaches to the identified challenges from the perspective of research institutes and offers a set of recommendations for various stakeholders to facilitate the acceptance of the technology. The contribution of this article lies in its synthesis of the current state of knowledge and its identification of future research needs and challenges from an industry perspective, ultimately providing a roadmap for future research and development in the field of digital twin and its applications in the wind energy industry.

The field of steganography has witnessed considerable advancements in square-pixel-based image processing (SIP). However, the application of steganography in Hexel (Hexagonal pixel)-based Image Processing (HIP) is still underexplored. This study introduces a pioneering spatial steganography method called the Reversible Logic-Based Hexel Value Differencing (RLBHVD) method in the HIP domain. Our approach draws inspiration from Pixel-Value-Differencing (PVD), a SIP fundamental spatial-domain (S-D) steganography method. Initially, the image is transformed into the HIP domain using the custom software infrastructure developed for this project. Due to the absence of commercial equipment capable of producing HIP-domain images, traditional digital imaging systems are employed with their sensor components, analog-to-digital conversion units, and square-pixel-based displays. Once the image is converted, it is partitioned into standardized heptads, each comprising seven hexels. Simultaneously, the secret message is segmented for embedding into the hexels within each heptad. Unlike SIP-domain PVD, which embeds segments into independent pixel pairs, our method performs iterative embedding within each heptad. Additionally, we leverage Feynman gates, a core element of reversible logic, to achieve retrieval of both the cover image and the secret message. Unlike PVD in SIP, our approach enables reversibility in the recovery process. Experimental results demonstrate that our proposed method, RLBHVD, outperforms its SIP counterpart, PVD, by achieving a low Mean Squared Error (MSE), high Peak Signal-to-Noise Ratio (PSNR), and significant similarity between the stego-image and cover image histograms. These findings highlight the efficacy and superiority of our HIP-based steganography approach in comparison to existing SIP methods.

In today’s world, the importance of the Green Internet of Things (GIoT) in the transformed sustainable smart cities cannot be overstated. For a variety of applications, the GIoT may make use of advanced machine learning (ML) methodologies. However, owing to high processing costs and privacy issues, centralized ML-based models are not a feasible option for the large data kept at a single cloud server and created by multiple devices. In such circumstances, edge-based computing may be used to increase the privacy of GIoT networks by bringing them closer to users and decentralizing them without requiring a central authority. Nonetheless, enormous amounts of data are stored in a distribution mechanism, and managing them for application purposes remains a difficulty. Hence, federated learning (FL) is one of the most promising solutions for bringing learning to end devices through edge computing without sharing private data with a central server. Therefore, the paper proposes a federated learning-enabled edge-based GIoT system, which seeks to improve the communication strategy while lowering liability in terms of energy management and data security for data transmission. The proposed model uses FL to produce feature values for data routing, which could aid in sensor training for identifying the best routes to edge servers. Furthermore, combining FL-enabled edge-based techniques simplifies security solutions while also allowing for a more efficient computing system. The experimental results show an improved performance against existing models in terms of network overhead, route interruption, energy consumption, and end-to-end delay, route interruption.

Recent reviews of the literature indicate the need for empirical studies on cross-project defect prediction (CPDP) that would allow aggregation of the evidence and improve predictive performance. Most empirical studies predict defects at granularity levels of method, class, file, and module/package during the coding phase, and thereby avoid external failure costs. The main goal of this study is to perform an empirical study on early defect prediction at the beginning of a project at the product level of granularity for using it as input in planning quality activities of the project. Hence, both internal and external failure costs could be avoided as much as possible through proper planning of quality. We first made a systematic mapping study (SMS) on secondary studies (literature reviews) on defect prediction to identify the most used datasets, the project attributes and metrics utilized as estimators, and the supervised learning methods employed for training the data. Then, we made an empirical study on defect density prediction using cross-project data. We collected 760 project data from the International Software Benchmarking (ISBSG) dataset version 11, which reported both defects and functional size attributes. We trained the prediction models using: i) the complete set of project attributes, ii) the individual attributes, and iii) multiple subsets of attributes. We employed classification and regression approaches of machine learning. The machine learning models are trained using original values of the dataset, and z-score and logged transformations of original values to explore the effects of data normalization on prediction. Most machine learning models trained on the z-score transformation of the dataset performed best for classifying defects. The Multilayer-Perceptron (Neural Network) model trained on the z-score transformation of complete dataset predicted defects with the highest F1-score of 0.89 using binary classification. The logged transformation and feature selection methods improved the results for multivariable regression. The multivariable regression predicted defects with the highest Root Mean Squared Error (RMSE) and R2 (r-squared) values of 0.4 and 0.9, respectively, with a subset of 11 features using logged transformation. The results of classification and regression approaches indicate that defects can be predicted with reasonable accuracy at the software product level using cross-project data.

Sparsity-promoting techniques show promising results in improving the generalization of neural networks. However, the literature contains limited information on how different sparsity techniques affect generalization when using neural networks to model non-linear dynamical systems. This study examines the use of sparsity-enhancing techniques to improve accuracy and reduce the divergence rate of neural networks used to simulate such systems. A range of sparsity methods, including hard and soft thresholding, pruning and regrowing, and L1-regularization, were applied to neural networks and evaluated in a complex nonlinear aluminum extraction process by electrolysis. The results showed that the most effective technique was L1 regularization, which enhanced the important connections in the network and improved the model performance. In contrast, many of the more advanced sparsity techniques resulted in significantly worse performance and higher divergence rates. Additionally, the application of Stochastic Weight Averaging during training increased performance and reduced the number of diverging simulations. These findings suggest that carefully selecting the right sparsity techniques and model structures can improve the performance of neural network-based simulations of dynamical systems.

Presents corrections to the paper, Security Control and Data Planes of SDN: A Comprehensive Review of Traditional, AI, and MTD Approaches to Security Solutions.

Distributed Agile Software Development (DASD) is the most important approach for the modern software industry that allows geographically independent software development. In the past, different tools and frameworks were proposed to solve communication and collaboration issues in DASD but they lacked transparency, trust, traceability, and security. These shortcomings resulted in project failure or overdue, customer dissatisfaction, project deal cancellations, and payment clashes between the customers and development teams. This paper addresses and overcomes the major issues of transparency, trust, security, traceability, coordination, and communication in DASD by embedding blockchain technology. We have proposed a novel blockchain-based framework named as AgilePlus which executes the smart contracts on a private ethereum blockchain for acceptance testing, secure payment, verification of developer’s payment requirements, and automatic payment distribution into the digital wallets of development teams. The execution of these smart contracts automatically assign penalties to the customers for late or non-payments and also to the developers for overdue tasks. Moreover, we have also solved the blockchain’s scalability challenge in AgilePlus by utilizing Interplanetary File System (IPFS) as off-chain storage. Lastly, experimental results prove that the proposed framework enhances transparency, communication, coordination, traceability, security and solves trust issues of both customers and developers in DASD.

The Internet of Things (IoT) is an emerging paradigm that is transforming real-world things (objects) into smarter devices. IoT is applicable to a variety of application domains including healthcare, smart grid, and agriculture. This domain has started revolutionizing the agriculture industry by providing smart solutions for precision farming, greenhouse management, and livestock monitoring. This article aims to present a comprehensive survey on the role of IoT in the Livestock field by categorizing and synthesizing existing research work in this area. To this end, a detailed discussion has been provided on IoT network infrastructure, topologies and platforms employed for livestock management. In addition, a list of communication protocols and connections of IoT-based livestock systems with relevant technologies have also been explored. Furthermore, numerous IoT-based livestock monitoring, controlling, and tracking applications have been discussed. Apart from this, it also analyses distinct security issues in IoT-based livestock field and developed a collaborative security model to detect and minimize the security risk. Lastly, pertinent open research challenges in the domain of IoT-based livestock management have been presented with future research directions.

With the rapid increase in the world’s population, the global electricity demand has increased drastically. Therefore, it is required to adopt efficient energy management mechanisms. Since the energy consumption trends are rather dynamic. Therefore, precise energy demand estimation and short and/or long-term forecasting results with higher accuracy are required to develop the optimization and control mechanism. Consequently, the machine learning (ML) techniques along with distributed demand response programs are being adopted to predict the future energy demand requirement with satisfactory results. In this paper, different state-of-the-art ML algorithms such as logistic regression (LR), support vector machines (SVM), naive Bayes (NB), decision tree classifier (DTC), K-nearest neighbor (KNN), and neural networks (NNs), have been implemented to analyze their performance. The main objective of this paper is to present a comparative analysis of ML algorithms for short-term load forecasting (STLF) regarding accuracy and forecast error. Based on the implementation and analysis, we have identified that, among other algorithms, the DTC provides comparatively better results. Therefore, we devised the enhanced DTC (EDTC) by integrating fitting function, loss function, and gradient boosting in DTC mathematical model for fine-tuning the control variables. The implementation results show that the proposed EDTC algorithm provides better forecast results (i.e., 99.9 % recall, 100% F1, 100% precision, 99.21 % training accuracy, and 99.70% testing accuracy.)

The use of digital twins to represent a product or process digitally is trending in many engineering disciplines. This term has also been recently introduced in the medical field. In arthroscopic surgery education, the paradigm shift from apprenticeship to simulation training has driven the need for better simulators, and the current focus is on improving simulators with respect to computational efficiency and system accuracy. However, expanding surgical simulations towards digital twins has not yet been explored. This paper introduces the digital twin concept for arthroscopic surgery, and explores its potential in light of the existing scientific literature. Thus, a systematic review was conducted to summarize and analyze the literature with respect to fast and robust design of an arthroscopic digital twin using patient-specific information, and methods for interactive surgical soft tissue simulation. The review was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol with three reliable scientific search engines: IEEE Xplore, ScienceDirect and PubMed. Eighty papers were included in the review, and the extracted data included modeling methods, tissue types, constitutive behavior, computational efficiency or accuracy, hardware configuration, haptic device description, software tools, and system architectures. Considering the review, a novel macro-level conceptual arthroscopic digital twin system is presented, and the applicability of the review findings for the identified subsystems are discussed. The proposed system integrates patient-specific images, diagnostic data, intraoperative sensor data, and surgical practice as inputs, and conceptually enables surgical skills training, preoperative planning, and a database of virtual surgeries.

Short-term load forecasting plays an essential role in the efficient management of electrical systems. Building an optimization model that will enhance forecasting accuracy is a challenging task and a concern for electrical load prediction. Due to Artificial Neural Networks (ANNs), the final result depends on initial random weights and thresholds that affect the stability of the forecast. Although much devotion is being given to improving the forecast accuracy, convergence, complexity, and resilience need to be considered for stable predictive models. To overcome this limitation, this work has jointly considered the Wavelet Neural Network (WNN) and Self-Adaptive Momentum Factor (SAMF) to achieve fast convergence, stability, and high accuracy. The proposed hybrid model is developed by combining the Feature Engineering (FE) and SAMF with the WNN model. The FE removes the irrelevant data and shallow features to ensure high computational performance. In contrast, the SAMF combines the wavelet transform’s time and frequency domain properties and adjusts the WNN model’s corresponding parameters. This ensures the global optimum solution while returning accurate predictive results. Finally, the SAMF is used to tune the control parameters of WNN by initializing the random weights and thresholds to accelerate the convergence rate and improve the accuracy compared to the Back-Propagation (BP) method. The proposed hybrid model is tested on the real-time datasets taken from the Australian states of (New South Wales (NSW), and Victoria (VIC)). Experimental results show that the developed model outperforms other benchmark models such as WNN-IGA, BPNN, WNN-AMBA, and Enhanced WNN in terms of instability, rate of convergence, and accuracy.

Modular multilevel converters (MMCs) have emerged as a viable choice in future DC grid architectures due to their scalability to meet voltage level requirements. However, MMC-based DC distribution systems are at risk of short-term outages during the faults in either the DC or AC networks feeding the MMC, so it remains a challenge to accomplish AC and DC fault ride-through (FRT) capability in such applications. To ensure stable operations of the DC terminals, FRT strategies are required for the faults on both the AC and DC sides of the converter. This paper proposes a FRT strategy for the AC and DC side of the converter to ensure stable and economically viable operation of the DC distribution network. The asymmetrical faults in the upstream AC grids are managed by using the integrated energy of the MMC. Whereas, the DC FRT capability of the MMC is accomplished by changing the redundant submodules of the MMC to full-bridge submodules (FBSMs), thus allowing a DC FRT to be achieved by using DC circuit breakers that are low cost and reduced in size. Applying the proposed DC FRT strategy, which makes possible the use of low-cost and reduced in size DC circuit breakers in DC distribution, results in a reduction in the overall initial investments.

Water quality prediction (WQP) plays an essential role in water quality management for aquaculture to make aquaculture production profitable and sustainable. In this work, we propose hybrid deep learning (DL) models, convolutional neural network (CNN) with the long short-term memory (LSTM) and gated recurrent unit (GRU) for aquaculture WQP. CNN can effectively fetch the aquaculture water quality characteristics, whereas GRU and LSTM can learn long-term dependencies in the time series data. We conduct experiments using the two different water quality datasets and present an extensive study on the impact of hyperparameters on the performance of the proposed hybrid DL models. Furthermore, the performance of hybrid CNN-LSTM and CNN-GRU models are compared with different baseline LSTM, GRU and CNN DL models and also with attention-based LSTM and attention-based GRU DL models. The results show that the hybrid CNN-LSTM outperformed all other models in terms of prediction accuracy and computation time.

Automated recognition of handwritten characters and digits is a challenging task. Although a significant amount of literature exists for automatic recognition of handwritten characters of English and other major languages in the world, there exists a wide research gap due to lack of research for recognition of Urdu language. The variations in writing style, shape and size of individual characters and similarities with other characters add to the complexity for accurate classification of handwritten characters. Deep neural networks have emerged as a powerful technology for automated classification of character patters and object images. Although deep networks are known to provide remarkable results on large-scale datasets with millions of images, however the use of deep networks for small image datasets is still challenging. The purpose of this research is to present a classification framework for automatic recognition of handwritten Urdu character and digits with higher recognition accuracy by utilizing theory of transfer learning and pre-trained Convolution Neural Networks (CNN). The performance of transfer learning is evaluated in different ways: by using pre-trained AlexNet CNN model with Support Vector Machine (SVM) classifier, and fine-tuned AlexNet for extracting features and classification. We have fine-tuned AlexNet hyper-parameters to achieve higher accuracy and data augmentation is performed to avoid over-fitting. Experimental results and the quantitative comparisons demonstrate the effectiveness of the proposed research for recognition of handwritten characters and digits using fine-tuned AlexNet. The proposed research based on fine-tuned AlexNet outperforms the related state-of-the-art research thereby achieving a classification accuracy of 97.08%, 98.21%, 94.92% for urdu characters, digits and hybrid datasets respectively. The presented methods can be applied for research on Urdu characters and in diverse domains such as handwritten text image retrieval, reading postal addresses, bank’s cheque processing, preserving and digitization of manuscripts from old ages.

Machine learning and deep learning algorithms are widely used in computer science domains. These algorithms are mostly used for classification and regression problems in almost every field of life. Semantic segmentation is an instantly growing research topic in the last few decades that refers to the association of each pixel in the image to the class it belongs. This paper illustrates the systematic survey of advanced research in the field of semantic segmentation till date. This study provides the brief knowledge about the latest proposed methods in the domain of semantic segmentation. The proposed study comprehends the concepts, techniques, tool, and results of different research frameworks proposed in the context of semantic segmentation. This study discusses the latest research papers in which machine learning and deep learning techniques are exploited for semantic segmentation and published between 2016 and 2021. The systematic literature review collected from seven different article libraries including ACM digital Library, Google Scholar, IEEE Xplore, Science Direct, Google Books, Refseek and Worldwide Science. For assuring the quality of the paper those papers are selected which have several citations on standardized platforms. Most of the studies used COCO, PASCAL, Cityscapes and CamVid dataset for training and validation of the machine learning and deep learning models. The results of the selected research articles are collected in the form of accuracy, mIoU value, F1 score, precision, and recall. In this study, we also conclude that most of the semantic segmentation studies use ResNet as the backbone of the architecture and none of the researchers used ensemble learning methods for semantic segmentation that is the loophole of the selected studies.

Social media communication serves as an integral part of the crisis response following a mass emergency (disaster) event. Regardless of the kind of disaster event, whether it is a hurricane, a flood, an earthquake or a man-made disaster event like a riot or a terrorist attack, social media platforms like Facebook, Twitter etc. have proven to be a powerful facilitator of communication and coordination between disaster victims and other communities. Consequently, several research articles have been published on social media utilization for disaster response. Many of those recent research articles discuss automated machine learning approaches to extract disaster indicating posts, useful for coordination from various social media posts. Despite this, there is a scarcity of comprehensive review of all the major research works pertaining to the utilization of machine learning approaches for disaster response using social media posts. Thus, this study reviews academic research articles in the domain and classifies them across three disaster phase dimensions – early warning and event detection, post-disaster coordination and response, damage assessment. This review would help researchers in choosing further research topics pertaining to automated approaches for actionable information classification and disaster coordination and would help the emergency teams to make well-informed decisions in disaster situations.

E-learning in higher education is exponentially increased during the past decade due to its inevitable benefits in critical situations like natural disasters (e.g. COVID-19 pandemic etc.) and war circumstances. The reliable, fair, and seamless execution of online exams in E-learning is highly significant. Particularly, online exams are conducted on E-learning platforms without the physical presence of students and instructors at the same place. This poses several issues like integrity and security during online exams. To address such issues, researchers frequently proposed different techniques and tools. However, a study summarizing and analyzing latest developments, particularly in the area of online examination, is hard to find in the literature. In this article, a Systematic Literature Review (SLR) of online examination is performed to select and analyze 53 studies published during the last five years (i.e. Jan 2016 to July 2020). Subsequently, five leading online exams features targeted in the selected studies are identified. Moreover, underlying development approaches for the implementation of online exams solutions are explored. Furthermore, 16 important techniques / algorithms and 11 datasets are presented. In addition to this, 21 online exams tools proposed in the selected studies are identified. Additionally, 25 leading existing tools used in the selected studies are also presented. Finally, the participation of countries in online exam research is investigated. Key factors for the global adoption of online exams are identified and compared with major online exams features. This facilitates the selection of right online exam system for a particular country on the basis of existing E-learning infrastructure and overall cost. To conclude, the findings of this article provide a solid platform for the researchers and practitioners of the domain to select appropriate features along with underlying development approaches, tools, and techniques for the implementation of a particular online exams solution as per given requirements.

The implementation of blended learning in higher educational institutions is increasing due to its perceived effectiveness in affording the benefits of both face-to-face traditional mode and the fully online mode of instructions. However, the leading challenge associated with the online component of blended learning is students’ inability to properly self-regulate their learning activities as self-regulation has been an integral competence for the success in online environments. Research on supporting students’ self-regulation in the online component of blended learning has been on the rise, but these approaches have primarily focused on designing systems and techniques to support students self-regulated learning and have not provided solutions to self-regulation strategies, more importantly online peer-learning self-regulation strategy. The lack of online peer learning self-regulation strategy has been the hallmark of the challenges in blended learning instruction. This study proposed an approach for scaffolding students peer-learning self-regulation strategy in the online component of blended learning. The study involves forming online peer-learning groups based on learning potential and affinity structures; leveraging the design of learning system features to instigate group dynamic elements for facilitating and motivating small group peer-learning, as well as promoting pro-social behavior by offering incentives for combating the inherent problems of reluctance to students’ peer-learning participation. An experiment involving 120 university Students shows that effective online peer-learning self-regulation strategy scaffold involves preparing and inspiring students so as to prevent social loafing and refusal to participate; and secondly to facilitate and fully engage students in the actual online peer-learning discussion. The experimental results show significant improvement of students’ academic performance by using this approach. In addition, our study proposed recommendations for future research.

The new emerging networks such as smart grids, smart homes and Internet of Things have enabled user accessibility across the globe and employ non-orthogonal multiple access (NOMA) scheme to accommodate huge number of connected devices. These devices which include smart meters, sensors and actuators etc. suffer from impulse noise (IN) while operating with power systems. Furthermore, NOMA scheme provides power domain multiple access (PDMA) which is found to be susceptible to IN. Based on the aforementioned IN intervention and its degrading effect on communication applications, novel mechanisms are desired to mitigate and classify the IN induced in the received signal. In this research work, novel IN mitigation and classification techniques are presented using deep learning methods for NOMA-based communication systems. The IN detection is performed by first identifying the IN occurrences using a deep neural network (DNN) which learns statistical traits of noisy samples followed by removal of harmful effect of IN in the detected occurrences. Using the proposed DNN, higher bit error rates (BER) were achieved when compared with the existing IN detection methods. The proposed method was further validated for high and low IN, and weak and strong IN occurrence probabilities. Moreover, another deep learning network is proposed in this research work to effectively distinguish between high IN and low IN in the noise contaminated NOMA symbols which can help improve the performance of IN detection models. Both of the deep learning methods proposed in this study show strong potential to address IN problem faced by the NOMA scheme.

This paper proposes a decision support framework for defining planning parameters for national crop production. The proposed framework addresses the gaps in policymaking, the role of all stakeholders, and uses historical data of crop production in different sectors of land in optimizing the profits that shall meet the defined constraints including national requirement, and export demand of different crops. There are many agro-economy-based countries where agriculture is the main contributor to their GDP, while there are others who always struggle to meet their national need while optimizing the agro-economy component of their earnings. Crops are not always produced with keeping the demand and production balance. Rather, the crops are mostly produced according to the farmer’s ease and last year’s prices for profit maximization leading to excessive production of one type of crop, lowering of profit margins, and shortage of some key crops for national need. This paper presents a decision aiding tool that can be used for farmer’s awareness for crop production planning that meets the national and export needs while maximizing farmer’s earnings. The model is tested on historical data of different segments of cultivated lands in Pakistan for validation, with the first experiment performed on single farmland with multi-fields while the second performed for multi-locations and multi-fields. The results are promising and provide estimates of net profit, expected production against the demand for each crop, and analysis that aids in crop planning, before and after the application of our model while meeting all necessary constraints.

There has been significant development in the facial recognition technology during past few decades. This technology has been widely used by different organizations and governments for defense, security, and surveillance projects. Furthermore, it has now been incorporated into our daily usages, such as consumer applications, personal data protection, or cyber-security, particularly while using smartphones. Most of these systems work very efficient, however, there are some challenges related to the accuracy of results of facial recognition systems when tested on images of people with dark skin. As a matter of fact, various studies demonstrate higher accuracy when tested on data set with white skin personnel, while exhibit a much lesser accuracy when the same algorithms are tested on dataset of people with dark skin. This article highlights the variation in accuracy of existing facial recognition algorithms when applied to dark-skinned people. Furthermore, as a principal contribution it presents a hybrid algorithm based on Gaussian and Explicit rule model that improves the accuracy for face-detection for dark skinned people. Thorough experimental evaluation has been conducted with a data set of black faces by first identifying skin and non-skin regions and then applying skin segmentation. The results have been compared with existing face detection algorithms with a clear improvement in the accuracy of 89% for dark skin.

Medium Access Control (MAC) is the sub-layer of data link layer in the communication OSI model, whose basic function is to control the hardware that interacts with the transmission medium. In Wireless sensor network, the communication is done without any wired connectivity and the sensor nodes get energy through batteries. MAC layer protocol is classified as contention based and contention free. Over the years, these protocols have served efficiently in controlling the data transmission, yet they have some compromising and challenging issues, and data collision is one of them. Contention based protocols are reported to be more vulnerable to collisions, as they allow sharing of channels among transmitting nodes. Collisions result in message delay, data loss and retransmissions, resultantly demanding more energy. Researchers are trying their best to reduce the collisions during transmission. This research is aimed to evaluate the priority-based collision control in contention-based protocol. We present a new technique Energy efficient Collison mitigation MAC (ECM-MAC) protocol to handle the QoS and energy constraints for variety of applications. ECM-MAC uses different contention window and node priority mechanism to get the overall performance of the network. Nodes with most important data or with least residual energy will be allowed to transmit first, thereby minimizing the chances of collisions and maximizing the lifespan of the network. In case of high level of collision, the data traffic will be intelligently reduced to minimize the collisions. A contention-based MAC protocol is evaluated to compare the impacts of suggested mechanism with the standard protocol. It will be useful for designers to develop a better collision avoiding contention-based protocol. In wireless Sensor Networks (WSNs), a sensor node is typically comprising detection unit, transceiver unit (to send and receive the data packets), processing unit and an energy supply unit for energy provisioning. Enhancement in network lifetime is an important and major challenge in WSNs. Since the transceiver unit has more impact on the energy consumption of the node and its reduction results in significant improvement on network lifetime. In our work, we propose an efficient mechanism that prioritized the nodes to get the high performance in quality of service. MAC protocol is targeted to achieve this, as MAC layer has the authority access to the transmission channel. We prioritize the nodes with lower energy to increase the chances of that node to get access of the transmission channel before their power and battery ran out. Our proposed strategy outperforms other protocols in terms of content retrieval Time, retransmissions and energy efficiency.

Cardiovascular diseases are considered as the most life-threatening syndromes with the highest mortality rate globally. Over a period of time, they have become very common and are now overstretching the healthcare systems of countries. The major factors of cardiovascular diseases are high blood pressure, family history, stress, age, gender, cholesterol, Body Mass Index (BMI), and unhealthy lifestyle. Based on these factors, researchers have proposed various approaches for early diagnosis. However, the accuracy of proposed techniques and approaches needs certain improvements due to the inherent criticality and life threatening risks of cardiovascular diseases. In this article, a MaLCaDD (Machine Learning based Cardiovascular Disease Diagnosis) framework is proposed for the effective prediction of cardiovascular diseases with high precision. Particularly, the framework first deals with the missing values (via mean replacement technique) and data imbalance (via Synthetic Minority Over-sampling Technique - SMOTE). Subsequently, Feature Importance technique is utilized for feature selection. Finally, an ensemble of Logistic Regression and K-Nearest Neighbor (KNN) classifiers is proposed for prediction with higher accuracy. The validation of framework is performed through three benchmark datasets (i.e. Framingham, Heart Disease and Cleveland) and the accuracies of 99.1%, 98.0% and 95.5 % are achieved respectively. Finally, the comparative analysis proves that MaLCaDD predictions are more accurate (with reduced set of features) as compared to the existing state-of-the-art approaches. Therefore, MaLCaDD is highly reliable and can be applied in real environment for the early diagnosis of cardiovascular diseases.