International Journal of Innovations in Science & Technology
Not a member yet
813 research outputs found
Sort by
A Comparative Study of the Energy Efficiency of Traditional Network Topology and Software-Defined Networking (SDN) Topology
Full text linkThis study is intended to be a comparative study of the energy efficiency of the traditional network topology and the software-defined networking (SDN) topology. Energy efficiency has been a priority aspect in network design due to environmental concerns and cost optimization of operational costs. Traditional networking is based on the existing configuration of the hardware devices and decentralized control, which in turn results in ineffective usage of the resources. However, in contrast, SDN centralizes network control and thus facilitates energy-efficient resource allocation and optimization. In this comparison, the energy consumption profiles, the utilization patterns of the resources, and the operating strategies of both methods are evaluated. The goal of this study is to present information on SDN energy efficiency over the conventional networking method and to demonstrate how SDN can offer benefits to the environment and economy by using SDN technologies in network infrastructures. However, the selection of the specific network infrastructure may vary depending on specific user requirements
Comparative Study of Food Image Classification Performance Using the Xception Architecture
Full text linkFood allergies remain a critical issue that needs more research. To identify and manage food allergies, the integration of complex computational approaches is becoming more and more important, opening the door to more individualized and efficient food safety solutions. Which aims to end hunger, achieve food security, improve nutrition, and promote sustainable agriculture. This research investigates the application of image classification techniques for allergen detection in food images. Specifically, we compare two models Model 1 serves as the baseline, trained on 11 classes. Two variations were explored: Model 2 focuses on Pakistani dishes, to investigate the impact of learning rate on the balance between adaptation speed and model precision. The objective is to determine the most effective model for classifying food images therefore Model 2 achieves the highest accuracy of 94%. These findings suggest that Model 2 is a promising candidate for real-world allergen detection applications. Future research will focus on creating a comprehensive new dataset of food images encompassing a wider variety of food items, as well as exploring the integration of a model similar to model 2 into mobile applications for consumer use
A Comparative Analysis of SaaS and PaaS Cloud-Based E-Learning Platforms in Terms of Cost-Effectiveness and Scalability
Full text linkEducational institutions must evaluate Software as a Service against Platform as a Service for their e-learning cloud deployment because their cost models differ while trade-offs in scalability and customization needs exist. The research examines SaaS and PaaS cloud platforms to help educational institutions select better options through assessments of their value for money and flexibility and user satisfaction measures. A mixed quantitative and qualitative research design involved collecting data from twenty institutions which were equally distributed between SaaS and PaaS subscribers. These results were supplemented by interviews with IT administration personnel. The research used subscription fees and infrastructure costs as quantitative data alongside user capacity and response times under load. At the same time it collected qualitative information about usability alongside customization flexibility and security perceptions of the solutions.
Institutions working with restricted budgets should choose SaaS platforms because they provide cost-effective subscription payments at 4,380.72 annually) and specialized technical knowledge. User satisfaction surveys highlighted SaaS’s ease of adoption (75% satisfaction) versus PaaS’s customization advantages (40% extensive customization satisfaction), though both models achieved comparable security satisfaction (70–75%).
The study concludes that SaaS is optimal for institutions prioritizing affordability and simplicity, while PaaS suits those requiring long-term scalability and tailored solutions. Recommendations include hybrid cloud models to balance cost-efficiency and flexibility. These insights aim to empower educational stakeholders in aligning cloud adoption strategies with institutional goals and resource constraints.
Institutions working with restricted budgets should choose SaaS platforms because they provide cost-effective subscription payments at 4,380.72 annually) and specialized technical knowledge. User satisfaction surveys highlighted SaaS’s ease of adoption (75% satisfaction) versus PaaS’s customization advantages (40% extensive customization satisfaction), though both models achieved comparable security satisfaction (70–75%).
The study concludes that SaaS is optimal for institutions prioritizing affordability and simplicity, while PaaS suits those requiring long-term scalability and tailored solutions. Recommendations include hybrid cloud models to balance cost-efficiency and flexibility. These insights aim to empower educational stakeholders in aligning cloud adoption strategies with institutional goals and resource constraints
A Modified Incremental Conductance-Based Active Power Curtailment for Voltage Regulation in Highly PV-Fed Distribution Grids
No full textIncreasing the penetration of photovoltaic (PV) systems in the distribution grid poses challenges, including voltage rise and potential violations during low load conditions or peak PV generation. This paper introduces a modified incremental conductance (IC) method to limit the active power injection from PV systems, mitigating voltage violations. A voltage-sensitive control loop is integrated into the IC method to improve its performance. This algorithm allows for the dynamic adjustment of active power output based on real-time voltage measurements from the grid, enabling precise voltage regulation while maintaining maximum power point tracking (MPPT) under normal conditions. When necessary, it shifts the operating point away from the maximum power point (MPP) to limit active power injection, ensuring voltage stays within safe limits. The proposed algorithm is less complex, cost-effective, and can be implemented in existing inverters using the IC method. The algorithm’s effectiveness is validated through Simulink/MATLAB simulations, using a setup consisting of a distribution grid with two solar PV-based distributed generators (DGs), each with a capacity of 100 kW
Optimal Coordination of Directional Overcurrent Relays in Interconnected Networks Using an Improved Multi-Strategy Coati Optimization Algorithm
Full text linkThe proper coordination of directional overcurrent relays (DOCRs) in interconnected power systems is essential for selective and time-efficient protection. This study introduces a Tuned Non-inertial T-distribution based Weighted Coati Optimization Algorithm (TNTWCOA) to solve the challenging, highly constrained DOCR coordination problem. The proposed method optimizes time dial settings (TDS) and plug settings (PS) to minimize relay operating times while maintaining selectivity. Coordination is critical for both primary and backup protection devices to prevent fault currents from rising to dangerously high levels too quickly. TNTWCOA uses a chaotic sequence mechanism for better population initialization, a nonlinear inertia weight to balance exploration and exploitation, an adaptive T-distribution strategy to increase diversity and avoid local optima, and an alert update mechanism to improve search adaptability. The algorithm was tested on IEEE 3-bus and 15-bus networks, considering both mid and near-end faults with a normal inverse relay characteristic. A comparative analysis with other advanced metaheuristics shows that TNTWCOA outperforms classical and recent optimization methods by reducing total relay operation time. The results confirm that TNTWCOA helps prevent premature convergence and boosts search efficiency, making it a highly effective solution for DOCR coordination in modern power systems
Disease Detection Using Wrist Pulse Analysis
Full text linkEarly detection of diseases is crucial for effective treatment and management. Traditionally, disease detection involves invasive and costly medical procedures. However, recent advancements in non-invasive methods have proven highly successful in identifying various illnesses. The wrist pulse has long been an important tool for detecting diseases, with traditional Chinese medicine making extensive use of this method. It shows great promise in diagnosing a wide range of conditions. This study provides a detailed analysis of research on wrist pulse analysis and its applications in disease detection. It examines the physiological basis of wrist pulse analysis, focusing on the relationship between underlying medical conditions and the characteristics of wrist pulses. Additionally, the study explores how wearable pulse detectors and machine learning algorithms can improve the accuracy and effectiveness of wrist pulse analysis. In this research, we use a dataset of 300 samples from various diseases, analyzing it with MATLAB and applying ensemble learning algorithms. We have achieved accuracies above 80% for nearly all algorithms, and accuracy can be further improved by expanding the dataset with more samples and extracting additional features
Energy Harvesting Implementation in WBAN Routing Protocols with Multi-Relay Co-Operation
Full text linkMostly simulations are used to evaluate the performance of Wireless Body Area Networks (WBANs). The recent researches are focused on channel modelling and energy conservation at the Network/MAC layer. Normally, collaborative learning, path loss, and energy harvesting are ignored in these schemes of studies. In this research, we will try to use an Energy Harvesting (EH) mechanism to recharge the batteries instead of replacing them time and again. In contrast with the existing studies, the proposed scheme considers collaborative learning and energy harvesting. Cost functions are used to identify the most feasible wireless route from a given node to the sink while sharing each other’s distance and residual energy information. The human body temperature (thermal energy) and the pumping of the heart can be used for energy harvesting within the body, while solar energy can be used for energy harvesting of nodes on the human body
Auscultation-Based Pulmonary Disease Detection and Classification Using Deep Neural Networks
Full text linkPulmonary diseases like Pneumonia, Bronchiectasis, and Chronic Obstructive Pulmonary Disease cause a large number of deaths worldwide. For such diseases to be treated and managed effectively, an early and accurate diagnosis is essential. In this work, we propose a deep learning model based on Recurrent Neural Networks (RNN) that can detect three different pulmonary diseases, as well as healthy lung sounds, using only auscultation recordings. The model was trained using the ICBHI dataset, which contains 920 recordings from 126 people and covers more than 6,800 respiratory cycles. To uniform the data, the audios are padded to equal length. To tackle class imbalance in the dataset, augmentation techniques of Gaussian noise injection, time-shifting, and time stretching are used. We employ a simplified version of the Gated Recurrent Unit (GRU)-based RNN architecture to deal with the padded sequences, along with a dropout layer to avoid overfitting. The model is trained using the Adamax optimizer with categorical cross-entropy loss, along with a model checkpoint to ensure learning consistency. Apart from the evaluation of model accuracy, we also evaluated the F1-score, accuracy, and loss graphs to ensure the competitive performance of our approach. Out of the six different experiments, with different data variations and two different model architectures, the outperforming model exhibited an accuracy of 98.53%, a precision of 98.57%, a recall of 98.53%, and an F1-score of 98.52%
Leveraging Machine Learning for Spreading Factor Optimization in Lora WAN Networks
Full text linkThe Internet of Things (IoT) has witnessed exponential growth and widespread integration across diverse sectors such as agriculture, logistics, smart cities, and healthcare. Among various IoT communication paradigms, the Long-Range Wide Area Network has emerged as a prominent and preferred technology, attributed to its extended transmission range, energy efficiency, and cost-effectiveness. Nevertheless, the escalating proliferation of IoT endpoints has amplified the complexity of efficient resource orchestration, particularly in Spreading Factor (SF) optimization within infrastructures. To mitigate this challenge, this study introduces a Machine Learning–driven Adaptive Data Rate (ML-ADR) framework for dynamic SF management. A Long Short-Term Memory (LSTM) neural network was meticulously trained using a dataset synthesized via ns-3 network simulations to achieve optimal SF classification. The pre-trained LSTM model was subsequently deployed on end-device nodes to enable intelligent and adaptive SF allocation using real-time data during simulation. Experimental evaluations reveal significant enhancements in packet delivery ratio and notable reductions in energy consumption, thereby validating the efficacy and scalability of the proposed ML-ADR approach
EEG Based BCI for Intelligent Wheelchair Control System Using Deep Learning
Full text linkThis research study presents the design of an Electroencephalography (EEG) based Brain Computer Interface (BCI) for intelligent wheelchair control to assist patients with mobility disorders. The concept of this research is to enable a direct communication link between the human brain and the machine without physical movement. This study used the BCI Competition IV 2a dataset, which contains EEG recordings of nine subjects performing four motor imagery (MI) tasks that were mapped to wheelchair navigation commands such as turning left, right, moving forward, and stopping. In this study, a deep learning architecture, TCFormer (Temporal Convolutional Transformer), was implemented to learn the spatial and temporal correlations between EEG channels. A lightweight Fusion Head module was added to enhance performance. It consisted of one-dimensional convolution and adaptive pooling operations for improved local temporal feature extraction. The proposed TCFormer-Fusion model achieved an overall classification accuracy of 75%, outperforming the baseline TCFormer model by 72%. Overall, this research study demonstrates that transformer-based models can learn complex EEG signal representations for motor imagery classification. The proposed model contributes toward developing an intelligent wheelchair control system that operates on brain signals, reducing external assistance. This work, with further optimization and real-time implementation, can contribute significantly to the assistive technology and human-computer interaction fields