Smart grids have attracted more and more attention due to the characteristic of itself that provide much more interactive solution comparing from the conventional grid, such as enabling demand side management (DSM) and demand response and so on. In smart grid, Wide Area Networks (WAN) is a part of the largest communication network that provides communication to the energy generation domains such as solar panels. Internet-of-Thing (IoT) can be adopted to provide communication in WAN. However, IoT may suffer in term of quality of services as well as energy efficiency because of how far the data must transmit since WAN cover the largest area. Therefore IoT-Edge Computing can be adopted to process the data nearer to the sources of data generation before transmitting into the cloud services. Energy efficiency in IoT is critical especially when smart grid is considered. Specifically, when a fault occurs in a grid, the sensors and the technologies in the areas affected by the fault will be isolated and have the shortage of power supply. In this paper, we proposed power control method to curb the problem state before. To find the best action, deep reinforcement learning (DRL) is adopted. DRL has the ability to learn make decision in a dynamic environment through a combination of deep learning and reinforcement learning techniques. Simulation results show that with the proposed method, the energy efficiency manage to elevate.
Managing operating theatres (OT) is critical in achieving an effective health care delivery system. This includes the Master Surgery Scheduling Problem (MSSP), which assigns surgery groups such as medical specialties to OT time slots. Constructive heuristics for the MSSP have been incorporating greedy approaches to generate the solution. However, greedy methods tend to ignore hidden factors and could lead to solution infeasibility. Hence, motivated by other timetabling problems, this study incorporated the Saturation Degree ordering into a ranking-based constructive heuristic for solving the MSSP to improve solution feasibility. The proposed algorithm was benchmarked against the greedy and random approaches, as well as a regret-based approach from the literature and found that it has reduced repair mechanism usage by 57.43%. Furthermore, the ranking-based algorithm avoided infeasible solutions in all runs, whereas the greedy and regret-based approaches failed to do the same when using a greedy repair mechanism. Overall, this work has proved the efficacy of the Saturation Degree ordering in enhancing solution feasibility.
Multi-stage classification of Alzheimer's disease (AD) refers to classifying the disease into its multiple stages. Aside from a binary classification task that classifies between the normal control (NC) and AD stages only, an additional prodromal stage known as Mild Cognitive Impairment (MCI) is also being classified. MCI is the stage between the healthy subjects known as the NC class, and the patients with heavy symptoms, in the AD class. In other words, MCI subjects only have slight or mild symptoms of Alzheimer's, thus leading it to be a challenge for detection. Classification models usually perform well in binary classification tasks, but not in multistage classification tasks, due to the faint difference in their features. Thus, this research proposes the incorporation of an attention mechanism into the classification model to increase its multi-stage classification performance. The attention mechanism facilitates the classification task by identifying the important features in MRI images so that the model can better differentiate the multiple classes. The MRI data used in this study is obtained from the Open Access Series of Imaging Studies (OASIS) database. The experimental results show that the attention-incorporated model has achieved an improved classification performance as compared to the normal model without attention. The generalizability of the enhanced model is also improved as observed from the training-classification gap results. Hence, the exceptional performance of the attention mechanism positions it as a solution to boost and enhance multi-stage AD classification.
Sun Dried coconut meat, also known as Copra is used mainly on coconut oil extraction and its by-products for livestock use. The grade of copra is set by the Philippine National Standard for Copra and one of the criteria for the grading/classification is dryness of the copra meat as indicated by the meat color. This study investigates the use of Faster Region-Based Convolutional Neural Network with Inception v2 model in detecting and classifying the level of dryness of copra images as either Optimally Dry, Under-Dry or Over-Dried levels of copra meat. The researchers trained the model and tested it using a validation set which yielded an accuracy rate of 90% for Under Dry, 92% for Optimally Dry copra and 81% for Over Dry, with an overall accuracy rate of 87.67%.
With the development of modern industry 4.0, intelligent path planning is an essential research direction of schedule systems for an automated guided vehicle (AGV), which has been widely used in logistics distribution centers of enterprises. This work implements the improved Q-learning algorithm to solve the typical obstacle avoidance problems in path planning. Specifically, the conventional Q-learning algorithm has shortcomings including low operational efficiency and slow learning speed. The improved Q-learning algorithm is successively proposed by adding a learning process based on the original Q-learning algorithm, which enables AGV to find obstacles and target locations within the shortest time, thus improving the efficiency of path planning. Finally, the simulation experiments are carried out in the grid environment with MATLAB. In comparison to the conventional Q-learning algorithm, the improved Q-learning algorithm has faster convergence and higher learning efficiency, improved by 20%.
In recent years, Japan has been facing a super- aging society. The problems of tight emergency medical care and an increase in the number of emergency medical transports have become ¶uite serious. In response to this situation, our research group has been conducting joint research with the Kobe City Fire Department. This study aims to propose and establish a method for medium to long-term prediction of the number of emergency medical transports and to provide an indicator for the strategic deployment of Emergency Medical Services (EMS) and the expansion and contraction of the scale of medical services in the field. This medium to long-term prediction of the number of ambulance transports is achieved by analyzing big data, population data, and future population estimates in each region without machine learning. The proposed method was evaluated in Kobe City, Japan. The results showed that the predicted number of transports is reasonable based on the actual number of transports in the past.
This paper presents an evaluation of IEEE 802.15.4 signal strength in indoor environments for positioning applications. The study involves experimental signal strength measurements, comparing different numbers and placements of access points (APs). The results reveal that signal strength variation is influenced by the quantity and strategic positioning of APs, the presence of obstacles, and the effects of multi path propagation. The theoretical path loss model is computed and compared with the empirical path loss model of the indoor setting. The study demonstrates the reliability of the calculated theoretical model in estimating IEEE 802.15.4 Signal Strength in indoor environments with an error rate of less than 5.5 %. As a future research direction, the application of machine learning techniques is proposed to derive a path loss model that aligns better with real path loss data at specific points in the indoor environment. Such an approach promises improved adjustments to measured signal strength values, particularly in indoor settings.
The increasing variety of radio standards utilized by modern mobile devices necessitates that these devices handle numerous frequency bands. As the frequency range increases, the support for multiple bands creates design issues for frequency synthesizers. The increased frequency range provides the benefit of broader bandwidth and smaller antennas that require less space. An existing frequency synthesizer's performance and range can be enhanced by including frequency dividers. Injection-locked frequency dividers (ILFD) receive the most attention among other frequency dividers due to their better phase noise performance and low-power at high frequencies. This paper presents a divide-by-2 Complementary Metal-Oxide-Semiconductor (CMOS) LC-ILFD with a wide locking range percentage and low power. The proposed CMOS LC-ILFD uses sinusoidal tail current-shaping and cross-coupled topology to reduce phase noise and power consumption. At an injection power of 0 dBm and an injection voltage of 0.6 V, the proposed ILFD has an input frequency locking range of 3.8 GHz to 6.5 GHz (52.43%). Aside from that, the proposed ILFD had a phase noise of -125.97 dBc/Hz at a noise frequency of 1 MHz. The proposed ILFD was realized in 180 nm CMOS process with ADS software and had a power dissipation of 1 mW at a 1V DC supply.
The Malaysian Government has been developing smart cities progressively since the 1990s. However, the urbanization development in Malaysia has been delayed since the announcement of Vision 2020. It is undeniable that the construction industry plays an important role in smart city urbanization development. As such, the aim of this study is to investigate the perspective of the construction industry involvement in Malaysia in developing smart cities urbanization. Quantitative study and Partial Least Square-Structural Equation Modelling (PLS-SEM) are used for data collection and data analysis. The results revealed that the construction industry player's readiness is highly perceived compared to awareness and preparedness. In conclusion, the construction players are ready to be involved in smart city urbanization development. However, in order to expedite the overall progress of smart city urbanization development in Malaysia it is crucial for the government to implement rules and regulations so that construction industry players familiarize the procedures.
For the daily life application of BCI, any assumptions could be violated during the practical test, especially with the small number of channels. Multiple electroencephalogram (EEG) channels could have high accuracy but are not practical for daily life usage compared to a single-channel EEG. However, the single-channel EEG signal has limited sources to extract informative features. Therefore, the aim of the study is to analyze the significant impact of different feature extraction techniques on single EEG channel classification performance. The BCI competition III (IVa) dataset is used for this purpose. Time-frequency domain (E1), and time-domain feature extraction techniques (E2) as well as the combination of both techniques (E3) have been considered to extract the relevant features set from the EEG signals for MI tasks classification. The classification accuracy is obtained with two classifiers: Support Vector Machine (SVM) and Logistic Regression (LR). The result shows that there is an enhancement in classification accuracy on certain channels and participants, with E3 being higher than E1 and E2. The C3 channel in E3 has the highest accuracy (87.1 %) when SVM is the classifier. The proposed approach improves the classification accuracy of AF4 by up to 70% and 70.4% by using SVM and LR, respectively, which is considered acceptable for usage in a BCI system. This implies that MI tasks can be detected with a single-channel EEG by using the proposed method even if the channel is located on the forehead.
Multi-stage classification of Alzheimer's disease (AD) refers to classifying the disease into its multiple stages. Aside from a binary classification task that classifies between the normal control (NC) and AD stages only, an additional prodromal stage known as Mild Cognitive Impairment (MCI) is also being classified. MCI is the stage between the healthy subjects known as the NC class, and the patients with heavy symptoms, in the AD class. In other words, MCI subjects only have slight or mild symptoms of Alzheimer's, thus leading it to be a challenge for detection. Classification models usually perform well in binary classification tasks, but not in multistage classification tasks, due to the faint difference in their features. Thus, this research proposes the incorporation of an attention mechanism into the classification model to increase its multi-stage classification performance. The attention mechanism facilitates the classification task by identifying the important features in MRI images so that the model can better differentiate the multiple classes. The MRI data used in this study is obtained from the Open Access Series of Imaging Studies (OASIS) database. The experimental results show that the attention-incorporated model has achieved an improved classification performance as compared to the normal model without attention. The generalizability of the enhanced model is also improved as observed from the training-classification gap results. Hence, the exceptional performance of the attention mechanism positions it as a solution to boost and enhance multi-stage AD classification.
Load Frequency Control is indeed important in any country that operates an interconnected power grid for regulating system frequency by adjusting the generation to match the load demand. To maintain the power system stability and improve efficiency in scheduling, this paper proposes Two-area Load Frequency Control (LFC) Optimization with Particle Swarm Optimization. The design process begins with understanding the fundamental concepts, starting from Automotive Generation Control in a single-area power system design and progressing to LFC design in a two-area power system. The two-area power system consists of Tie-line Bias Control (TLBC) for maintaining zero tie-line power and Area Control Error (ACE) for stabilizing system frequency. To achieve fitness results in simulations, this paper presents three modeling designs: conventional design, controller design, and controller design with PSO. The controller system with PSO leads an evaluation of 53.4 % time schedule in system frequency performance and an evaluation of 76.9% time schedule in generation power deviation with respect to the conventional model. In addition, the controller system lags an evaluation of 70.0% time schedule in system frequency performance and an evaluation of a 78.7 % time schedule in generation power deviation with respect to the conventional model.
Control and optimization of exothermic batch reactions are crucial tasks in the field of chemical engineering, with extensive applications in various industries such as healthcare, monitoring, and production. This review comprehensively analyzes the state-of-the-art evolutionary algorithms used for controlling and optimizing exothermic batch reactions. The paper explores and compares the performance of popular algorithms such as GA, PSO, and other hybrid evolutionary algorithms. The review focuses on the strengths and limitations of each algorithm, analyzing their capabilities in handling different exothermic batch reaction processes, including their applicability, robustness, and complexity. Additionally, the paper investigates the commonly used optimisation techniques based on evolutionary algorithms for controlling exothermic batch reactions in practical case studies. This review also reveals the latest advancements and emerging trends in this field, such as multi-objective optimization, multi-mode optimization, and the integration of deep learning methods.
The freshness of fruits is important in making sure good quality of fruits can be provided to consumers. Manually classifying fruits based on the characteristics of fruits by hand and eyes can be time consuming and unreliable. Convolutional Neural Network (CNN) is the most common deep learning method that has been exploited to classify the freshness of fruits through processing the images of fruits. The classification effectiveness of the CNN is highly depending on its hyperparameter. However, the hyperparameter setting of CNN method has not yet been tuned under diverse values in recognizing the fresh and rotten fruits. In this work, the hyperparameter of CNN is tuned to identify the best setting that can effectively recognize the fresh and rotten fruits particularly orange, apple, and banana. These including the number of epoch, batch size, learning rate, and optimizer with respect to ReLU and Sigmoid activation function. The experiment results show that the hyperparameter tuned CNN model achieved the highest classification accuracy of 99.04%, on real world dataset. Hence, this indicates that the hyperparameter tuned CNN model is capable to recognize the fresh and rotten fruits well.
Hyperspectral imaging provides abundant spectral information with hundreds of bands, but selecting an optimal number of bands is crucial for efficient and accurate data analysis. The k-means clustering method is widely used for band selection, but the quality of clustering and efficiency of band selection depends on the similarity measure used. In this paper, we propose an improved version of k-means clustering using spectral similarity measures (SSM) such as Spectral Angle Mapper (SAM), Spectral Information Divergence (SID), and the hybrid measure of SID and SAM (SIDSAM) for optimal band selection. Empirically it is proved that the proposed k-means with spectral similarity measures will identify the best bands and thereby improve the classification accuracy.
The rising demand for renewable energy sources has fueled extensive research in photovoltaic (PV) systems. However, conventional Maximum Power Point Tracking (MPPT) algorithms often encounter challenges when tracking the global maximum power point under non-uniform irradiance conditions. To address this issue, the Clonal Selection Algorithm (CSA) is proposed as an effective approach to enhance MPPT algorithm performance. The CSA dynamically adjusts the voltage perturbation size based on instant ambient irradiance and temperature, leading to improved global maximum power point tracking and enhanced efficiency in PV systems. Experimental results demonstrate the superiority of the proposed CSA over conventional MPPT algorithms, especially in scenarios with varying solar irradiance. The CSA's adaptability allows PV systems to operate closer to their optimal efficiency, maximizing energy harvest from available solar resources. Overall, this research contributes valuable insights into sustainable and efficient energy solutions by leveraging the capabilities of the CSA. Successfully integrating the CSA in PV systems plays a critical role in establishing an eco-friendly and resilient renewable energy infrastructure, for a greener future.
In modern times, mental health issues have become an important concern with far-reaching effects on both individuals and companies. This research paper investigates the impact of geographical location, work environment, gender, and employment status on mental health illness prediction. This study aims to investigate the possible connections between these variables and mental health outcomes, as well as individuals' attitudes towards mental health. This study seeks to identify the strongest factors influencing mental health disorders and attitudes, contributing to our knowledge of the complex mechanisms behind mental well-being. To achieve the objective, various classification models were used, including the Random Forest Classifier, AdaBoost Classifier, Gradient Boost Classifier, XGB Classifier, and a hybrid Stacking Model. The models were trained and evaluated using various metrics to measure their performance. The metrics considered were train and test accuracy, precision, recall, F1- score, ROC curve, and AVC. The metrics mentioned are important indicators of the models' predictive accuracy, ability to distinguish, and overall performance. In this study, we conducted a comparative analysis of various techniques and successfully implemented them. Through our evaluation, we identified the Stacking technique as the most accurate method, achieving a prediction accuracy of 82.27%. This study aims to contribute towards addressing the potential rise of a widespread “mental health epidemic.”
The production of palm oil on a commercial scale is labour intensive with many of its processes handled by humans. In some countries, there can be as many as 500,000 plantation workers in the palm oil sector as the plantations are usually large. However, such dependence on humans for low skills manual work has led to many problems. Unmanned aerial vehicles (UA Vs) have been seen as a possible alternative to support some of the processes that require low skills in the palm oil industry. However, the flying time of the UAVs is finite and hence it is important to maximize the number of palm trees that each UAV can service. This paper proposes the segmentation of large palm oil estates into smaller areas to be modeled as a bin packing problem with the Jaya Algorithm. The resultant segmented areas would have an optimal number of palm trees that can be comfortably serviced by the UAV. Good results were achieved when tested on several datasets especially when compared to those computed by human experts.
Addressing the problem of language identification in code-mixed datasets poses notable challenges due to data scarcity and high confusability in bilingual contexts. These challenges are further amplified by the associated imbalance and noise characteristic of social media data, complicating efforts to optimize performance. This paper introduces an augmentation approach designed to enhance language identification in bilingual code-mixed social media data. By incorporating reverse translation, semantic similarity, and sampling techniques alongside customized reprocessing strategies, our approach offers a comprehensive solution to these complex issues. To evaluate the effectiveness of the proposed approach, experiments were conducted on language identification at both the sentence and word levels. The results demonstrated the potential of the approach in optimizing language identification performance, offering a compelling combination of generation techniques for addressing the challenges of language identification in code-mixed data.
This paper presents an approach to enhance the economic efficiency of power systems through tap settings optimization using Cuckoo Search Algorithm (CSA), while ensuring voltage stability. With the escalating demand for electrical power, power system networks face challenges such as voltage decay, increased current flows, and energy losses. Reactive power planning, including tap settings optimization, plays a crucial role in addressing these challenges and optimizing system performance. Specifically, this study focuses on the role of on-load tap-changing transformers (OLTCs) in regulating voltage levels without interrupting the power supply. CSA is employed as an optimization technique due to its effectiveness in handling complex and nonlinear optimization problems. The objectives of the optimization process include minimizing active power losses and generation costs while maintaining voltage stability. The IEEE 26-Bus Reliability Test System (RTS) is used as a test system to demonstrate the implementation and effectiveness of the proposed approach. It has been found that CSA successfully improved the power system operation in terms of reduced active losses and generation costs compared to the non-optimal solution obtained from the power flow analysis. The results contribute to enhancing the economic efficiency of power systems and provide valuable findings for power system planning and operation.
