Photovoltaic (PV) systems have converted solar irradiance into electrical energy through PV cells that exhibit nonlinear voltage-current (V-I) characteristics. A key feature of these characteristics has been the maximum power point (MPP), where the product of voltage (V) and current (I) has reached its maximum, enabling optimal power extraction. The efficient operation of a PV system has required rapid and accurate Maximum Power Point Tracking (MPPT), especially in dynamically changing environmental conditions. This paper has presented a hybrid MPPT approach that has combined a Fuzzy Logic Controller (FLC) with the conventional Perturb and Observe (P&O) algorithm. The proposed two-stage control scheme has continuously identified and adjusted the MPP in response to variations in irradiance and temperature, including partial shading (PS) scenarios. In the first stage, the FLC has provided an intelligent initial estimate of the MPP region to improve the convergence speed of the P&O algorithm. In the second stage, another FLC has dynamically adjusted the step size of the P&O algorithm to improve response and reduce oscillations. The hybrid FLC-P&O algorithm has been validated through simulations under various conditions, including steady irradiance, sudden changes in sunlight, and partial shading of varying severity. The results have demonstrated that the proposed controller has achieved high tracking efficiency and has effectively overcome the limitations of the conventional P&O algorithm, particularly under non-uniform and rapidly changing environmental conditions.

The growing deployment of renewable energy sources and the need for robust low-carbon power systems have accelerated the use of decentralized energy management approaches. Networked Microgrids (NMGs) are an interesting solution to enhance distributed energy coordination and reliability. Nevertheless, the efficient management of energy under uncertainty, with the least possible operating cost and emissions, continues to be a significant challenge. This paper presents a Bi-Level Optimization Framework for NMGs in an Edge–Fog–Cloud infrastructure for supporting real-time, uncertainty-aware energy management. Deep learning models are trained at the cloud level for photovoltaic (PV) and load forecasting and deployed at edge. Forecast uncertainty is measured in terms of Prediction Interval Coverage Probability (PICP) and Prediction Interval Normalized Average Width (PINAW) over a number of different confidence levels. Experimental outcomes prove that Gated Recurrent Units (GRU) and Long Short-Term Memory (LSTM) models exhibit better forecasting performance, with GRU achieving lower PINAW and higher PICP under most conditions. The edge layer utilizes a local Mixed-Integer Nonlinear Programming (MINLP) optimizer for energy dispatch within each microgrid, while the fog layer manages inter-microgrid energy exchange and utility interaction via a global MINLP optimizer. Simulation results show that the robust optimization strategy—based on worst-case uncertainty—yields near-optimal results compared to the base case using actual data. On the validation dataset, the base case achieved a maximum profit of 43.75%, and the robust optimization achieved up to 37.40%. On the test dataset, they achieved 19.96% and 13.25%, respectively, at specific time stamps. A scalability study has also been conducted to demonstrate the scalable nature of the proposed framework, ensuring timely emission-conscious and uncertainty-resilient energy management, thereby supporting the requirements of future smart grid environments.

Autonomous vehicles (AVs) have shown significant potential in recent years, with increasing interest from the public, industry, and academia in their adoption on roads. To achieve reliable autonomous driving, AVs must be able to operate safely under adverse weather conditions such as rain-induced wet roads which pose serious safety risks to road users. Therefore, AVs require a robust Lane Change Decision (LCD) model as a crucial part of their decision-making system. Although existing studies have explored the use of deep reinforcement learning (DRL) for LCD models, most of these works do not consider rainy conditions in the development of the LCD model. To address this gap, this paper proposes a rain-aware LCD model using Deep Q-Networks (DQN). The proposed framework was implemented using the open-source Highway-env simulation environment, modified to incorporate rain-dependent human-driven vehicle behavior and wet-road friction effects on the ego vehicle. In addition, a safety-augmented DQN model was introduced to penalize unsafe close-following behavior before an actual collision occurs. The proposed models were evaluated in terms of safety, efficiency, comfort, and global traffic impact against rule-based (MOBIL) and learning-based benchmarks. The proposed Rain-aware DQN model improved the overall lane-changing performance by 11.2

The lithium-ion batteries (LIBs) are a crucial component for driving the electric power to the electrical components in electric vehicles (EVs); the demand for EVs is increasing day by day. The detailed thermal behavior of the LIBs analysis is essential to avoiding overheating and thermal runaway issues in EVs. The different EV application-oriented thermal modeling of battery packs (BP) was studied, and a 24-cylindrical-cell BP was considered for simulation setup and building the hardware setup, which is within the safety zone operation. A detailed 3D transient thermal model was developed using OpenFOAM, an open-source CFD software, and the practical 24-cell BP designed for EV applications with the help of thermal sensors that monitored the cells’ temperatures. The spacing between adjacent cells is 5 mm with dual open ends to enhance airflow and promote effective heat dissipation. Simulations predicted a peak cell temperature of approximately 38 °C and a temperature difference of less than 5 °C across all cells under maximum load condition, maintaining operation well within safe thermal limits. The experimental results were very similar to the simulations, with differences of less than 3

Transmission congestion has become a major challenge in deregulated power systems (DPS). Although flexible AC transmission system (FACTS) devices are effective for congestion management, their high installation cost and complex control requirements limit widespread deployment. This paper proposes a battery energy storage system (BESS)-based congestion mitigation approach for DPS. The optimal locations of BESS units are identified using bus sensitivity factors (BSFs), while a weighted-sum scalarized single-objective optimization problem involving active power loss, voltage deviation, and system security margin is solved using salp swarm optimization (SSO). The main contribution of this work is the integrated BSF–SSO framework, which combines sensitivity-based placement with an efficient metaheuristic optimization technique. The proposed methodology is validated on the IEEE 30-bus test system. Results show that the optimally placed BESS units reduce active power losses by 31.7% relative to the base case, while it is reduced by 39.1% when using the post contingency of 25 MW as the baseline, voltage deviation by approximately 30.26%, and the congestion security index by 25.14%, thereby significantly alleviating transmission congestion. In addition, reactive power flow distribution and 24 h state-of-charge (SoC) dynamics are analyzed to provide a more comprehensive assessment of system performance. Compared with FACTS-based solutions, the proposed approach offers superior technical and economic benefits. A parametric sensitivity analysis further shows that SSO attains near-optimal convergence with as few as 15 agents, offering a favorable accuracy-versus-runtime trade-off for real-time DPS operation; the main results reported in this study, however, were generated using the more conservative configuration of n = 30 agents to maximize solution robustness.

The growing penetration of renewable energy sources in modern power systems has significantly complicated the challenge of load frequency control (LFC). The intermittent nature of solar and wind generation, coupled with increasingly unpredictable residential consumption patterns, undermines the effectiveness of traditional control mechanisms. Conventional proportional-integral-derivative (PID) controllers — including fractional-order, tilted, and filtered variants — are commonly employed but often fail to maintain frequency stability under such dynamic and nonlinear conditions. To overcome these limitations, this study proposes a novel Tilted Proportional Integral Filtered-Derivative Filtered-Accelerative (TPIDnAn) controller specifically designed to enhance LFC performance in renewable-dominated environments. The controller parameters are optimized using the Snake Optimizer (SO) algorithm, which offers robust tuning capabilities and improved adaptability. The proposed control strategy is rigorously tested on a two-area power system integrating solar photovoltaic (PV) and wind energy sources. Its performance is compared against other advanced controllers across multiple scenarios, including step load changes, random disturbances, generation rate constraint (GRC) non-linearity, and hybrid renewable integration. Evaluation metrics such as overshoot/undershoot (p.u.), transient time (s), peak time (s), and the integral of time-weighted absolute error (ITAE) confirm the superior effectiveness and resilience of the TPIDnAn controller in maintaining frequency stability and meeting LFC standards under diverse operating conditions.

Human Activity Recognition (HAR) in consumer healthcare and Internet of Health Things (IoHT) devices demands models that are accurate, energy-efficient, and explainable for on-device deployment. We propose MSA-TCN, an ultra-compact Adaptive Temporal Convolutional Network that combines multi-scale temporal convolutions, causal dilated encoding, and hybrid channel-temporal attention to model both short- and long-range motion dependencies. A subject-aware augmentation strategy enhances generalization across users and sensor configurations without additional computation. The quantized INT8 variant achieves an average accuracy of 98.7% across five benchmark datasets, with a compact size of 0.08 MB and an inference latency of 1.8 ms per 256-sample window on a mid-range smartphone (Snapdragon 845). Energy consumption is reduced by over 98% (from 32.5 mJ to 0.36 mJ) compared to FP32 inference. Post-hoc SHAP analysis provides clinically meaningful attributions, improving transparency and user trust in healthcare settings. By enabling on-device prediction, MSA-TCN minimizes wireless transmission and ensures reliability under intermittent connectivity, offering an interpretable, energy-efficient solution for next-generation IoHT applications.

The increasing demand for sustainable energy in residential buildings and public concerns on greenhouse gas (GHG) emissions has driven the integration of smart homes with hybrid renewable energy systems (HRESs). This research proposes an optimal scheduling strategy for home energy consumption in a grid-connected HRES that comprises a grid, wind turbines, photovoltaics and battery storage systems. The objective of the study is to reduce the net energy cost, scheduling inconvenience cost (SIC), GHG cost and battery degradation cost. An ant colony optimization algorithm is utilized in the MATLAB environment, with load profiles and meteorological data of Upington, South Africa, obtained from NASA and a residential consumption dataset to accomplish the objectives of the study. The outcomes of the study show that case study 3 is the most feasible configuration based on a net energy revenue cost of $9.8382, GHG cost of $0.0627, battery degradation cost of $0.461 and SIC of $0.66. Simulation results demonstrate that energy purchased from the grid has been reduced by 98% and 48% relative to case studies 1 and 2. The results of the study can assist households to improve the sustainability and resilience of the power system in residential environments where the grid supply is unstable and electricity costs are high.

The growing energy demand, acceleration of urbanization and expansion of industrial operations have raised the demand for intelligent energy management strategies for achieving energy efficiency and cost savings while minimizing carbon emissions. Most traditional energy management systems are based on rule based control structures and statistical methods that are not able to adjust to dynamic occupancy patterns, varying environmental conditions, and complex industrial processes. In recent years, a new technique, namely Machine Learning (ML), has emerged as a viable solution to predict, optimize, and automatically control energy use in smart buildings and industrial systems. The paper examines the various ML techniques for energy optimization in detail, and categorizes them into five areas: energy optimization for HVAC systems, energy optimization for lighting control, integration of renewable energy systems, energy optimization for industrial processes, and predictive maintenance. The proposed framework involves the combination of IoT sensors, real-time data collection, data preprocessing, feature engineering, development of ML models, and optimization algorithms, all aimed at realizing intelligent energy management. A set of supervised, unsupervised, deep learning, and reinforcement learning algorithms is examined, such as Random Forest, Support Vector Machine, XGBoost, Artificial Neural Network, Long Short-Term Memory network and Deep Reinforcement Learning with regard to their prediction accuracy, computational efficiency and energy saving requirement. As shown in the comparative analysis, advanced ML models consistently outperform traditional methods in predicting energy consumption, detecting consumption trends and reducing equipment idle time, as well as optimizing operational schedules. The results show that an energy optimization approach based on ML can greatly contribute to energy efficiency, operational cost savings, comfort of the occupants, and sustainable production in industry. The paper also presents the challenges and opportunities that are currently being faced, such as data quality, model interpretability, scalability, cybersecurity, and real-time deployment, and suggests future research directions, including the application of explainable AI, edge computing, digital twins, federated learning and autonomous energy management systems. In this study, the researchers give an overview of emerging ML techniques and practical lessons to the researchers and practitioners to develop intelligent, scalable and sustainable solutions for energy optimization in next-generation smart buildings and industrial facilities.

With the increase in the number of cases related to respiratory diseases, there is an urgent need to detect them early and diagnose them accurately. Convolutional neural networks have given promising results when used for diagnosing diseases using imaging tests. In this study, we investigate the potential of applying deep learning algorithms such as VGG16, VGG19, and ResNet50 for classification of lung ailments based on X-ray images. A detailed analysis of the aforementioned models' performances was conducted to assess how well they can classify various types of lung ailments, including pneumonia, tuberculosis, lung cancer, and normal lungs. In order to do that, these deep learning models were trained on a vast amount of X-ray images. The results of our study show that while all three models provide good results, ResNet-50 performs best in comparison with other models due to its efficiency and high level of accuracy. We believe that these deep learning models can be successfully implemented in the practice of diagnosing pulmonary diseases in the future. It helps with early disease detection and improves patient outcomes.

Over the last few decades, the world of exoskeletons has witnessed significant advancements, particularly in the actuation and control systems. These advancements have resulted in more sophisticated and lightweight designs that make them better suited for medical rehabilitation and mobility assistance. Recent developments in them have centred on improving flexibility and comfort through better sensing and enhanced control systems. Control strategies such as proportional–integral–derivative (PID) controllers were previously commonly used in control systems. But these methods relied on static parameters, which prevented them from adapting to a person’s gait variations. To overcome these problems, a new control approach known as data-driven control emerged. They involved models that can be trained using advanced learning algorithms and can synthesise large amounts of data to analyse intrinsic gait patterns and make more accurate predictions about movement trajectories. As these models are data-driven, it’s necessary to improve the extraction of key features from human gait data. Markerless pose estimation techniques are currently used to estimate joint coordinates accurately from videos of people walking. Modern markerless pose estimation techniques, such as the Mediapipe algorithm, are now used to estimate joint coordinates from videos of people walking without intrusive markers. This helps make the data recoding process more comfortable for participants. In this study, we propose a new approach to extract personalised gait profiles as a first step toward human motion-capture-based control of lower-limb exoskeletons, enabling them to better adapt to individual users’ distinct gait patterns and needs. This is a step towards developing exoskeletons that enhance the mobility and quality of life of their wearers.

This study presents a method that combines Monte Carlo Simulation (MCS) with Modified Honey Badger Algorithm (MHBA) for the first time to investigate the impact of photovoltaic (PV) system uncertainties on power losses under different solar radiation conditions. Moreover, this research proposes the chance-constrained probabilistic planning by generating probability distribution functions (pdf) of bus voltage and line current and taking into account probability constraints for encouraging more efficient and secure utilization of electricity. This study contributes to the information base concerning the optimal PV system placement by considering chance-constrained method. The simulation results reveal that the MHBA approach demonstrates more advantageous computational performance as compared to Differential Evolution (DE) in the IEEE 118 bus distribution system. By running both algorithms several times, MHBA method produces the best optimal power loss of 531.8649 kW in 1475.6075 s, whereas DE approach gives that of 540.9138 kW in 3373.4966 s for high radiation scenario, respectively. It has been revealed that the power loss may be further reduced by the increasing solar radiation conditions. In order to provide evidence that the proposed methodology is effective, the results of optimization are validated under different scenarios through the application of MCS. In addition, the constraints of network are investigated in order to make a determination regarding the probabilities of exceeding their limits. The validation of optimization results has concluded that the confidence levels of 0.7, 0.8, and 0.9 have been maintained for the low, medium, and high solar radiation scenarios, respectively.

This study introduces an innovative air-cooling strategy by integrating the flow spoilers for cylindrical lithiumion battery systems across eight different configurations (BTMS-I to BTMS-VIII). The coupled 1D electrochemical and 2D thermal-fluid model was employed with realistic heat generation for the turbulent air inflow of 3-5 m/s and cell discharge rates of 1C-5C. The integration of spoilers enhances turbulence kinetic energy, significantly improving cooling efficiency. At 3C discharge rate and 3.5 m/s inlet velocity, the configurations BTMS-I (Z-type), II (U-type), III, VI, VII, and VIII achieved notable temperature reductions of 4.54 K, 1.0 K, 0.55 K, 2.35 K, 0.15 K, and 2.78 K, respectively, compared to baseline designs without spoilers, with only marginal increases in the pressure drop. Conversely, BTMS-IV and BTMS-V exhibited degraded thermal performance with spoilers. A comprehensive parametric analysis was performed on BTMS-I and II, optimizing spoiler dimensions (length, height, position, inclination angle) and fillet radii at inlet/outlet corners to enhance thermal uniformity, safety, and energy efficiency. In BTMS-I, an optimal fillet radius of 10 mm reduced peak temperatures while maintaining lower pressure drop. Additionally, BTMS-II with a 3 mm fillet radius demonstrated improved thermal uniformity, albeit with a modest rise in pressure drop. Finally, a performance metric was developed to balance cooling effectiveness against the power consumption, providing a systematic approach for evaluating BTMS designs. In essence, this work advances air-cooled BTMS via aerodynamic and geometric optimizations, offering practical insights for electric vehicle applications.

Accurate forecasting of load, photovoltaic (PV), and wind power is essential for microgrid operation under renewable uncertainty. The renewable energy output varies in a stochastic manner, which makes accurate prediction difficult. The study presents a hybrid deep learning framework that combines wavelet decomposition (WD) and ensemble empirical mode decomposition (EEMD) to preprocess the time series, followed by convolutional neural network–long short-term memory (CNN–LSTM) processing enhanced with attention and Transformer components to capture local patterns and long-range temporal dependencies. The model is tested using data from two geographically and climatically different sites with seasonal variation. Forecasting performance is evaluated using mean absolute error (MAE), mean squared error (MSE), root mean square error (RMSE), and coefficient of determination (R2). The normalized RMSE values obtained are 1.68% for load, 1.81% for PV, and 1.75% for wind, which are lower than those of eleven models. Model robustness is further examined through twenty independent simulation runs and computational efficiency analysis. To examine the role of forecasting accuracy in operational decision-making, the forecasting outputs are integrated into a microgrid energy management problem. A microgrid case study shows that the use of the forecasting model reduces the dispatch cost to $6.91E+03, which is 0.01% less than the actual cost. The results show that the proposed framework improves prediction accuracy and supports cost-effective microgrid operation in uncertain conditions.

This study introduces an innovative strategy to enhance the efficiency and profitability of electrical distribution networks (DN) in the context of expanding electric vehicle (EV) usage. This work Introduces a hybrid algorithm-based strategy to optimize the placement, sizing, and operation of EVs and capacitors in DNs, improving power loss reduction, voltage regulation, and system reliability. The research leverages a hybrid optimization technique, HGWPSO, which uniquely combines the strengths of Grey Wolf Optimization (GWO), and Particle Swarm Optimization (PSO). These two methods together enable HGWPSO to efficiently identify ideal locations for EV charging stations and capacitor placement. The primary objective is to maximize the network’s efficiency by minimizing power losses and stabilizing voltage profiles. The proposed approach is validated on IEEE 33-bus, with results demonstrating a significant reduction in energy losses and improvements in voltage stability. Results reveal substantial gains, with energy loss costs dropping by 51.62% and in the IEEE 33-bus setups. Through its unique hybrid design, HGWPSO presents a powerful solution for integrating EV charging and reactive power support, providing an attractive option for utilities seeking to advance both economic and operational metrics in their networks.

One technique to supply power to small, low power devices such as smart grid meters used in the Internet of Things or sensors in Wireless Sensor Networks is operating based on radio frequency energy harvesting. In this paper, we give the design of RF energy harvesting circuits, and consider some of the main components to take into account–circuits performance, rectifiers, impedance matching, converters, and power management. First of all we will discuss what RF energy harvesting is and why it’s important in modern electronics. Recent research and new concepts are used in this paper. It explains how to save power and capture energy by rectifier capcitor circuits and impedance matching. It shows how to store and use energy in the sections on converter design and power management. The paper also provides new research in efficient, low power circuits. In particular, these include adaptive impedance matching and ultra low power rectifiers. Lastly, it points out potential future challenges and opportunities, stressing the importance of innovation that would enable improvement in energy harvesting efficiency and practicability of RF energy harvesting. This will allow researchers and engineers to clearly understand the status of this technology and potential future development.

Molybdenum disulfide (MoS2) has been found to be a promising material for electronic and optoelectronic device applications due to its unique optical and electrical characteristics. However, the large-scale synthesis of MoS2 thin films is limited by challenges in achieving reproducible and uniform device fabrication. In the present study, we utilized a sputtering technique and post-treatment by ion beam irradiation for large-scale fabrication of uniform MoS2 thin films. The effects of the low-energy ion beam on the optical, structural, electrical transport, and morphological characteristics of the MoS2 thin films were studied by Raman spectroscopy, atomic force microscopy (AFM), x-ray photoelectron spectroscopy (XPS), photoluminescence (PL) spectroscopy, and electrical transport analysis. Tuning the electrical and optical characteristics of few- and monolayer MoS2 through regulation of defects provides an excellent approach for fabricating two-dimensional (2D) MoS2 thin films for electronic device applications. Thin film transistors (TFTs) have been widely studied for driving active-matrix displays given their promising electrical characteristics including significant on/off current ratio and mobility. In the present work, we report a back-gate MoS2 TFT fabricated by sputtering. TFTs based on MoS2 thin films were fabricated, and the current–voltage characteristics were studied at room temperature, which confirmed that the transport behavior differed between the pristine and ion-irradiated samples. Pristine MoS2-based TFTs displayed significant Schottky barrier effects, resulting in lower mobility than ion-irradiated samples. Our comprehensive study focuses on the fundamental transport characteristics via the metal–MoS2interface, which represents a substantial step towards achieving highly efficient electronic devices based on 2D semiconductors.

Smart meters play a crucial role in transferring information between customers and utility companies by utilizing bidirectional communication and introducing vulnerabilities to cyber-attacks. These cyber-attacks, such as data replay attacks (DRA) and false data injection into databases, introduce errors in power demand load forecasting, which further results in energy supply scheduling problems. This paper provides a learning framework for smart meter cyber attacks by implementing a supervised machine learning algorithm to detect any kind of data anomalies or manipulation. The machine learning model employed in this paper includes Decision Tree, Random Forest, AdaBoost, and XGBoost. The fraudulent dataset is generated by employing two types of attack scenarios, and the performance of each machine-learning model is evaluated across diverse attack scenarios. In this proposed work, the percentage of attack level have been employed as 50%, 30%, 15%, and 5%. The performance metrics employed in this study to analyze the performance of the machine learning model are recall, precision, accuracy, f1-score, and area under the curve. The simulation results show that the attack level or imbalanced data proportion has a significant impact on the performance metrics of AdaBoost as the recall score has decreased from 0.9940 to 0.2608 if the attack level is decreased from 50% to 5%.