EAI Endorsed Transactions on Internet of Things

Advancing IoT Security with an Innovative Machine Learning Paradigm for Botnet Attack Detection

Punitha P — 2025-02-03

INTRODUCTION: In contemporary society, everyday operations are greatly improved by the Internet of Things (IoT), which connects physical devices to provide digital services. IoT technology offers unified services and streamlines activities across various domains, ranging from remote monitoring to sophisticated welfare systems. However, the growing number of IoT devices presents a security concern. Many of these devices are susceptible to exploitation, leading to diverse vulnerabilities.

OBJECTIVES: Resource-constrained IoT devices become prime targets for botnet attacks, manifesting in various forms and penetration methods. Despite numerous research efforts introducing multiple approaches for detecting botnet attacks in IoT, existing methods often fail to achieve satisfactory detection rates.

METHODS: Additionally, these approaches struggle to comprehensively analyze the diverse communication networks within the expansive realm of IoT devices. This study proposes an innovative machine-learning framework for detecting IoT botnet threats to address these limitations.

RESULTS: This conceptual framework exhibits a remarkable capability to identify a spectrum of botnet attacks, showcasing a detection accuracy of 99.5 per cent, significantly surpassing the performance of other prevalent machine-learning approaches.

CONCLUSION: Through this research, we aim to enhance the security paradigm of IoT networks, ensuring robust protection against evolving botnet threats in the dynamic landscape of interconnected devices.

Exploring the Educational Transformations: A Systematic Literature Review on the Influence of the Internet of Things in Higher Education

Vasileios Paliktzoglou — 2024-11-11

The rapid growth of the Internet of Things innovations stimulates higher educational institutions to invest in and adopt these technologies to support and enhance their learning and teaching strategies. This study aims to investigate the influence of the Internet of Things on teaching and learning in higher education, performing a systematic literature review. Therefore, this research focuses on the following research questions: R-Q1: What are the benefits of the Internet of Things for teaching and learning in higher education? R-Q2: What are the limitations of the Internet of Things for teaching and learning in higher education? The systematic literature review, including the search strategy, the inclusion and exclusion criteria, and the review of the titles, keywords, and abstracts, identified a total of 31 results, mainly journal and conference articles. The findings from the extracted articles in this review were grouped into eleven themes: adoption, personalized learning, learning efficiency, intelligent teaching, collaboration and connectivity, creativity, health and safety monitoring, latency time, security and privacy, quality and ethics, and financing issues. The findings suggest that the Internet of Things can enhance the learning quality, improve the gained knowledge, and reduce costs in higher education. Therefore, adopting a consistent Internet of Things implementation strategy is essential to address identified limitations in higher education.

Navigating the Future of Intelligent Transportation: Challenges and Solutions in 6G V2X and V2V Networks

Spandana Mande — 2025-02-06

Recent advancements in Information and Communication Technologies (ICT) have transformed vehicular communication, facilitating Vehicle-to-Vehicle (V2V) and Vehicle-to-Everything (V2X) connectivity. These technologies enhance travel safety and efficiency while also reducing pollution and accident rates through optimised resource utilisation. V2X applications encounter significant challenges concerning traffic safety, data security, and scalability, necessitating thorough evaluation prior to extensive implementation. The rising prevalence of connected vehicles presents challenges, including network instability, heterogeneity, and extensive data management, which require novel solutions for effective communication. This study gives a new view on how to combine 6G technologies with V2X and V2V communication systems. It addresses important issues like security, resource allocation, and making sure that communication is very reliable and has low latency. Our study is different from others because it looks at all the technologies that make this possible, including Proportional Fairness Algorithms, machine learning, blockchain, and Reconfigurable Intelligent Surfaces (RIS). We emphasise the significance of collective perception (CP) in improving traffic safety and resource efficiency through the sharing of sensor data among V2X-enabled entities, including vehicles, roadside units (RSUs), and vulnerable road users.This research provides a thorough analysis of advanced methodologies and their limitations, making it a valuable resource for scholars, academics, and industry professionals. It offers a framework for developing future intelligent transportation systems that utilise 6G capabilities to achieve high reliability, minimal latency, and optimal spectrum efficiency in vehicular networks.

Digital Literacy: Comparative Review on Machine Learning Based Performance Assessment of Students

K. Shwetha — 2024-11-15

The E-learning system paved an opportunity to make drastic changes in the educational system all over the world. Several institutions began to implement online learning to offer internet based courses contrary to the traditional classroom teaching. These online courses tends to provide several potential benefits such as flexibility and opportunities, to discover knowledge of the students. It also offers innovations in learning strategies of the students and resolve several complexities by accessing information from internet. Though e-learning based systems produces certain advantages, they also possess limitations of co-operative learning, active learning and performance mitigations. To address these issues, the present study focused on the different AI based techniques used in the prediction of student’s academic performance. The main objective of the study is to analyze the primary factors that affects the learning through online and analyze the performance using different intelligent approaches. A comparative study of the AI based techniques is performed to analyze the different methods involved in the assessment of academic performance. Further, the present issues and future works of the studies is deliberated to produce optimized analysis systems. This tends to support several researchers to overcome the disputes and provide effective e-learning assessment systems.

Major vulnerabilities in Ethereum smart contracts: Investigation and statistical analysis

Mohammad Pishdar — 2024-12-18

The general public is becoming increasingly familiar with blockchain technology. Numerous new applications are made possible by this technology's unique features, which include transparency, strong security via cryptography, and distribution. These applications need certain programming tools and interfaces to be implemented. This is made feasible by smart contracts. If the prerequisites are satisfied, smart contracts are carried out automatically. Any mistake in smart contract coding, particularly security-related ones, might have an impact on the project as a whole, available funds, and important data. The current paper discusses the flaws of the Ethereum smart contract in this respect. By examining publically accessible scientific sources, this work aims to present thorough information about vulnerabilities, examples, and current security solutions. Additionally, a substantial collection of current Ethereum (ETH) smart contracts has undergone a static code examination to conduct the vulnerability-finding procedure. The output has undergone assessments and statistical analysis. The study's conclusions demonstrate that smart contracts have several distinct flaws, including arithmetic flaws, that developers should be more aware of. These vulnerabilities and the solutions that can be used to address them are also included.

Revitalizing Image Retrieval: AI Enhancement and Metaheuristic Algorithm Adaptation

Kumaravel Pichaimani — 2025-01-17

INTRODUCTION: In recent years, development of digital technology has led to number of images, which can be stored in digital format. However, searching and retrieval of images in large image DB (Database) is a mammoth task. Therefore, different image retrieval techniques have been used for retrieving the suitable images, which includes retrieval of images using keywords or annotations, however, these methods are considered to be time consuming and leads to imprecise outcome.

OBJECTIVES: Therefore Effective and precise retrieval of suitable images from huge DB can thrived by utilizing CBIR (Content Based Image Retrieval) system. However, incorporation of CBIR in most existing studies resulted in low accuracy for IR. So, proposed model incorporates Modified ResNet50 (M-ResNet50) and VGG 16 model for feature extraction in order to extract the best features as M-ResNet50 utilizes extra dense layers which aids in better feature extraction process.

METHODS: After feature extraction, the features are fused using PCA and fed to Modified PSO (M-PSO) model for obtaining optimized features since M-PSO is fast and aids in selecting optimal features after processing from insignificant features that primarily set the preferred number of necessary features.

RESULTS: Moreover, M-PSO require less parameters to tune instead of a huge number of parameters by incorporating K parameters of KNN algorithm in order to find the nearest images to Query Images (QI), thereby making the model appropriate for IR process with better similarity score.

CONCLUSION: The proposed model utilizes 8 different sun images at different intervals for IR process. Finally, the proposed model is evaluated by using several metrics such as accuracy, precision, recall and F1 score, besides the proposed model is compared with various existing models in order to evaluate the efficiency of the proposed model.

Semantic Image Synthesis from Text: Current Trends and Future Horizons in Text-to-Image Generation

Lakshmanan Sudha — 2024-11-29

Text-to-image generation, a captivating intersection of natural language processing and computer vision, has undergone a remarkable evolution in recent years. This research paper provides a comprehensive review of the state-of-the-art in text-to-image generation techniques, highlighting key advancements and emerging trends. We begin by surveying the foundational models, with a focus on Generative Adversarial Networks (GANs) and their pivotal role in generating realistic and diverse images from textual descriptions. We delve into the intricacies of training data, model architectures, and evaluation metrics, offering insights into the challenges and opportunities in this field. Furthermore, this paper explores the synergistic relationship between natural language processing and computer vision, showcasing multimodal models like DALL-E and CLIP. These models not only generate images from text but also understand the contextual relationships between textual descriptions and images, opening avenues for content recommendation, search engines, and visual storytelling. The paper discusses applications spanning art, design, e-commerce, healthcare, and education, where text-to-image generation has made significant inroads. We highlight the potential of this technology in automating content creation, aiding in diagnostics, and transforming the fashion and e-commerce industries. However, the journey of text-to-image generation is not without its challenges. We address ethical considerations, emphasizing responsible AI and the mitigation of biases in generated content. We also explore interpretability and model transparency, critical for ensuring trust and accountability.

Performance Evaluation of Various Path Planning Methods for Robotics and Computational Geometry

Gokuldas Vedant Sarvesh Raikar — 2024-11-12

INTRODUCTION: In integrating Spiral Coverage into Cellular Decomposition, which combines structured grid-based techniques with flexible, quick spiral traversal, time efficiency is increased.

OBJECTIVES: In the field of robotics and computational geometry, the study proposes a comparative exploration of two prominent path planning methodologies—Boustrophedon Cellular Decomposition and the innovative Spiral Coverage. Boustrophedon coverage has limitations in time efficiency due to its back-and-forth motion pattern, which can lead to lengthier coverage periods, especially in congested areas. Nevertheless, it is useful in some situations. It is critical to address these time-related issues to make Boustrophedon algorithms more useful in practical settings.

METHODS: The research centres on achieving comprehensive cell coverage, addressing the complexities arising from confined spaces and intricate geometries. While conventional methods emphasise route optimization between points, the coverage path planning approach seeks optimal paths that maximize coverage and minimize associated costs. This study delves into the theory, practical implementation, and application of Spiral Coverage integrated with established cellular decomposition techniques.

RESULTS: Through comparative analysis, it illustrates the advantages of spiral coverage over boustrophedon coverage in diverse robotics and computational applications. The research highlights Spiral Coverage's superiority in terms of path optimization, computational efficiency, and adaptability, proposing a novel perspective into cell decomposition. The methodology integrates the Spiral Coverage concept, transcending traditional techniques reliant on grids or Voronoi diagrams. Rigorous evaluation validates its potential to enhance path planning, exemplifying a substantial advancement in robotics and computational geometry.

CONCLUSION: Our findings show that spiral coverage is on an average 45% more efficient than conventional Boustrophedon coverage. This paper set the basis for the future work on how different algorithms can traverse different shapes more efficiently.

GTBTL-IoT: An Approach of Curtailing Task Offloading Time for Improved Responsiveness in IoT-MEC Model

Eram Fatima Siddiqui — 2024-11-08

INTRODUCTION: The Internet of Things (IoT) has transformed daily life by interconnecting digital devices via integrated sensors, software, and connectivity. Although IoT devices excel at real-time data collection and decision-making, their performance on complex tasks is hindered by limited power, resources, and time. To address this, IoT is often combined with cloud computing (CC) to meet time-sensitive demands. However, the distance between IoT devices and cloud servers can result in latency issues.

OBJECTIVES: To mitigate latency challenges, Mobile Edge Computing (MEC) is integrated with IoT. MEC offers cloud-like services through servers located near network edges and IoT devices, enhancing device responsiveness by reducing transmission and processing latency. This study aims to develop a solution to optimize task offloading in IoT-MEC environments, addressing challenges like latency, uneven workloads, and network congestion.

METHODS: This research introduces the Game Theory-Based Task Latency (GTBTL-IoT) algorithm, a two-way task offloading approach employing Game Matching Theory and Data Partitioning Theory. Initially, the algorithm matches IoT devices with the nearest MEC server using game-matching theory. Subsequently, it splits the entire task into two halves and allocates them to both local and MEC servers for parallel computation, optimizing resource usage and workload balance.

RESULTS: GTBTL-IoT outperforms existing algorithms, such as the Delay-Aware Online Workload Allocation (DAOWA) Algorithm, Fuzzy Algorithm (FA), and Dynamic Task Scheduling (DTS), by an average of 143.75 ms with a 5.5 s system deadline. Additionally, it significantly reduces task transmission, computation latency, and overall job offloading time by 59%. Evaluated in an ENIGMA-based simulation environment, GTBTL-IoT demonstrates its ability to compute requests in real-time with optimal resource usage, ensuring efficient and balanced task execution in the IoT-MEC paradigm.

CONCLUSION: The Game Theory-Based Task Latency (GTBTL-IoT) algorithm presents a novel approach to optimize task offloading in IoT-MEC environments. By leveraging Game Matching Theory and Data Partitioning Theory, GTBTL-IoT effectively reduces latency, balances workloads, and optimizes resource usage. The algorithm's superior performance compared to existing methods underscores its potential to enhance the responsiveness and efficiency of IoT devices in real-world applications, ensuring seamless task execution in IoT-MEC systems.

Assessment of CatBoost for Diabetes Prevention in Comparison to XGBoost: AI model capable of predicting the onset of diabetes

Jagadeesh Selvaraj — 2025-02-10

The metabolic disease known as diabetes is defined by consistently elevated blood sugar levels. An increase in hunger, thirst, and urine production are symptoms of high blood sugar. Untreated diabetes may lead to a variety of complications. Acute complications of diabetes include hyperosmolarity, hyperglycemia, diabetic ketoacidosis, and perhaps death. The most devastating long-term effects are cardiovascular disease, cerebrovascular accident, chronic kidney disease, foot ulcers, and vision loss. The World Diabetes Organization estimates that 463 million people were diagnosed with diabetes in 2019. This population will increase by 578 million by 2030 and by 700 million by 2045, if forecasts pan out. The ability to quickly and accurately diagnose sickness is one of its current medical uses. Therefore, we might potentially reduce death rates via the use of machine learning by creating an AI model that can anticipate when diabetes will start. We will compare the CatBoost and XGBoost algorithms to find the one that is most suited for this purpose. Finally, using a number of health markers from the dataset, the study contrasted XGBoost and CatBoost, two models that may predict diabetes. We train and build our recommended system using Python on a real-world dataset taken from Kaggle. We evaluate our algorithms using precision, recall, F1score, and accuracy metrics, among other performance evaluation parameters. While XGBoost achieved an F1 Score of 91.75, an accuracy rate of 93.33%, a precision of 90.38%, and a recall of 90.63%. The accuracy, precision, recall, and F1 score for CatBoost are 96.09%, 93.38%, 91.38% and 92.13%, respectively. It's the most effective ensemble method, according to CatBoost.

Wireless 5G Network in Edge Computing Based On MIMO with Federated Learning and Clustering Integrated Reinforcement Learning

Manikandan Parasuraman — 2025-02-03

Edge Computing (EC) is a revolutionary architecture that brings Cloud Computing (CC) services closer to data sources than ever before. This research proposed novel technique in edge computing network based on wireless 5G technology using MIMO_federated learning integrated with Reinforcement neural network. Here the aim is to enhance the resource allocation by Decentralized Federated learning in multiple user based MIMO (De_Fed_L- MIMO) networks. Then the energy efficiency and channel optimization of the network is carried out using K-means clustering integrated with Reinforcement learning (K-means_RL). Here the experimental analysis is carried out in terms of number of users of network as well as number of edge server by DoF of 92%, Spectral efficiency of 92%, Energy efficiency of 96%, Signal to noise ratio (SNR) of 85%, Coverage area of 92%, RL training accuracy of 95%, FL training accuracy of 98%.

Heart Disease Diagnosis and Diet Recommendation System Using Ayurvedic Dosha Analysis

Kuldeep Vayadande — 2024-11-08

The current healthcare system often fails to account for individual health needs, leading to ineffective preventive measures and dietary guidance. Ayurvedic principles, which focus on the Dosha, offer a profound understanding of an individual's constitution, influencing their health, vulnerability to specific diseases, and ideal dietary choices. This paper explores the evolving intersection of ancient Ayurvedic wisdom and modern technology in the realm of disease diagnosis. Ayurveda, with its emphasis on personalized well-being, has long been a source of holistic health practices. In this context, the study delves into the intricate system of Ayurvedic Dosha analysis and its potential applications in contemporary healthcare. The research introduces an innovative way that seamlessly integrates traditional Ayurvedic pulse examination with state-of-the-art technology. By employing pulse sensors and advanced algorithms, the system not only identifies specific ailments but also classifies patients into Ayurvedic Prakriti types. Going beyond conventional diagnosis, this holistic approach extends to personalized recommendations, encompassing diet, lifestyle, Ayurvedic treatments, exercise, and daily routines. While addressing the challenges of harmonizing ancient principles with modern technology, the paper also presents the performance metrics of the model. The accuracy rates are as follows: Logistic Regression (LR) - 85.94%, Random Forest - 89.21%, Decision Tree - 99.70%, and k-Nearest Neighbors (KNN) - 86.43%. These metrics underscore the robustness of the system. In addition to outlining core concepts, methodologies, and model accuracies, the study explores current trends and recent developments in the field, offering readers a comprehensive understanding of Ayurvedic Dosha-based disease diagnosis. The research contributes to the broader discourse on healthcare by paving the way for early detection and individualized, holistic well-being for patients.

A Comparative Study on Machine Learning Classifiers for Cervical Cancer Prediction: A Predictive Analytic Approach

Khandaker Mohammad Mohi Uddin — 2024-11-19

INTRODUCTION: Cervical cancer is a significant global health concern, particularly in underdeveloped nations where preventive healthcare measures are limited. Early identification of the risks associated with cervical cancer is essential for both prevention and treatment.

OBJECTIVES: In recent years, machine-learning algorithms have gained popularity as potential techniques for determining a person's risk of developing cancer based on demographic and medical information. This study uses a dataset that contains patient demographics, clinical history, and results from diagnostic tests to examine how machine learning-based algorithms can be used to predict the risks of cervical cancer.

METHODS: Various machine learning approaches are used to create predictive systems, including Support Vector Machine (SVM), Naïve Bayes (NB), Decision Tree (DT), K-Nearest Neighbors (KNN), Random Forest (RF), Logistic Regression (LR), Gradient Boosting (GB), Nearest Centroid (NC), Multilayer Perceptron(MP), and AdaBoost (AB).

RESULTS: The prediction capability of these models is assessed using performance metrics such as accuracy, sensitivity, specificity, f-measure, precision, and area under the receiver operating characteristic curve (AUC-ROC). Our results show that the decision tree has the highest accuracy, precision, and f1-score (98.91%, 97.81%, and 0.9889). Additionally, model performance was optimized by the use of hyperparameter tuning. After hyperparameter adjustment, the Support Vector Machine (SVM) showed superior accuracy of 99.64%, precision of 99.26%, and an F1-score of 0.9963, thereby indicating its potential in cervical cancer probability prediction. We also created a web application that uses a machine-learning model to estimate the risk of cervical cancer.

CONCLUSION: The findings of this study highlight the significance of SVM and demonstrate the potential and capabilities of machine learning techniques to enhance accurate prediction and patient outcomes for cervical cancer screening.

A New Approach for Solving Fractional Differential Equations Incorporating Ramadan Group Transform and Machine Learning

Prabakaran Raghavendran — 2025-03-04

This paper examines various types of fractional differential equations using fractional calculus methods. It extends the classical Frobenius method and introduces key theorems that apply the Ramadan Group transform and other techniques. Additionally, the research incorporates machine learning, specifically neural networks, to solve these equations. The paper demonstrates that machine learning can enhance the solution process through data generation, model design, and optimization. Examples provided illustrate how combining traditional methods with machine learning can effectively solve fractional differential equations.

Comparative Study on Anomaly based Intrusion Detection using Deep Learning Techniques

Sabeena S — 2024-11-27

With an array of applications, Wireless Sensor Networks (WSNs) have the potential to transform the world into a smart planet. WSNs consist of a collection of resource-constrained sensors that gather data, which is then utilized for decision-making and analysis, leading to improvements in quality of service, management, and efficiency. However, the open nature of WSNs exposes them to numerous vulnerabilities and threats. Operating in potentially hostile and unattended environments makes these networks attractive targets for adversaries. Therefore, it is essential to detect the presence of malicious attacks within the networks and implement robust security systems to address these challenges. While traditional security mechanisms such as authentication and cryptographic methods are commonly employed, they often fall short in effectively countering the dynamic nature of modern attacks. Hence, IDS (Intrusion Detection System) tends to continuously monitor the network and detect potential threats in real-time scenarios. This method possess the ability of identifying, responding promptly, preventing and thus ensures resilience of the network. Therefore, the present study reviews the various intrusion detection techniques and data collection methods. The main aim of the study is to investigate the design challenges of deploying IDS in a WSN environment. So, the study analysed the AI (Artificial Intelligence) based techniques involved in intrusion detection and how these techniques could be adopted in WSN. In addition, the comparative analysis of several ML (Machine Learning) and DL (Deep Learning) algorithms are also deliberated to portray the different deployment technique with corresponding outcomes. Further, the main challenges faced by each studies with their limitations are specified for supporting future researchers in developing new trends in intrusion detection for WSN.

Design Of Intelligent Road Eye Using AI And Machine Learning For Automobiles

Chandrasekharan Nataraj — 2025-01-17

The project aims to design an intelligent road eye by using AI and a machine learning approach to detect speedbumps and potholes on the road. The system design utilizes a YOLOv5 custom-trained model and COCO dataset in detecting the objects on the road. The system is integrated with lane detection algorithms to achieve active steering feedback and pothole avoidance. Based on the detection results, feedback will be given in the form of visual, audio, and steering angles, allowing the driver to have sufficient response time to perform braking or steering adjustments where applicable. The trained model can achieve a mean average precision value (mAP) of up to 0.995 for all classes, and a maximum detection range of 5.77m and 34.8m for potholes and speedbump respectively. The future works of the project include integrating the algorithm into the vehicle to achieve autonomous braking and active pothole avoidance with the help of sensors and cameras on the vehicle, as well as adopting augmented reality (AR) to project the visual feedback on the vehicle windscreen.

Blockchain Based Cryptographic Algorithm for Data Protection in Electronic Voting System

Vinayachandra — 2025-01-08

The development of digital world created advancement of technology, by which the data services are electronically transacted. Electronic voting is used to improve the election system in several ways. But technical issues associated with protection of data, is considered as a major concern leading to illegal activities and threats. Some of the conventional methods used in securing e-votes are matching finger prints of individual, control unit accessibility, which provides decreased capability of attaining proper authenticity and security. Even several cryptographic methods are used for data security, but produced limited accuracy. So, to improve the reliability and accuracy of data protection in e-voting, combination of AES (Advanced Encryption System) and RSA (Rivest, Shamir, Adleman) combined with blockchain technology is implemented. AES uses larger key size which makes the data privacy robust. But encryption and decryption time produces decreases computational speed. So, AES is combined with RSA, as it uses shorter key length and easy to implement with block-chain algorithm, it is deployed to decrease encryption and decryption time, thus producing highly sensitive data security. The performance of the system is validated by calculating the encryption and decryption time, block size and verification time of original data with output data.

Spellbound Kannada: Harnessing Conditional Generative Adversarial Networks for Transformative Word Suggestion Systems in Kannada Language Processing

Prathibha R J — 2025-04-01

INTRODUCTION: The advancement of a word suggestion system model is driven by the need to enhance user interaction and efficiency in digital communication. Hence, the word suggestion system helps minimize typographical errors and spelling mistakes. Therefore, various traditional methods are used to suggest words to sentences; however, these traditional models are extremely time consuming, prone to errors and tedious. METHODS: Owing to these factors, the present paper focuses on developing a Kannada word suggestion system using cGAN (Conditional Generative Adversarial Networks), as this system is designed to significantly enhance user interaction by offering predictive text suggestions in the Kannada language. RESULTS: The training dataset, which resides on AWS S3, comprises a comprehensive collection of Kannada texts utilized for both training and validation purposes. Furthermore, the implementation of the model leverages the TensorFlow and keras framework, specifically employing long short-term memory (LSTM) networks for effective sequence prediction and generation. LSTMs are particularly advantageous in NLP processing because they can capture long-term dependencies within sequential data. To facilitate user interaction, a web-based interface has been developed using Flask, enabling users to input initial characters and receive dynamically generated Kannada word suggestions. CONCLUSION: This paper not only delves into the application of cGANs within the realm of NLP but also illustrates practical deployment strategies utilizing cloud services and modern web technologies. Overall, the proposed approaches demonstrate the potential of the cGAN in enhancing the user experience through intelligent text prediction systems tailored for the Kannada language.

Innovative Human Interaction System to Predict College Student Emotions Using the Extended MASK-R-CNN Algorithm

Dinesh P — 2025-03-10

There is a rising demand for emerging machines that can be self-decisive and intelligent. Machines can capture the emotions and gestures of college students to mechanise tasks and handle interactions better. Facial expressions based on emotion recognition are practices that play a substantial role in the modern fields of artificial intelligence and computer vision. Numerous manual methods for detecting emotions are focused on few basic emotions. Additionally, significant time is needed for appropriate detection. Nonetheless, these techniques are time-consuming and inefficient for obtaining better results. Therefore, an effective object detection model is needed to address such issues. To overcome these challenges, several studies have focused on object detection systems to provide effective emotion prediction. Conversely, it results in a lack of speed, precision and computational complexity. To improve object detection performance, the proposed model employs deep learning (DL)-based adaptive feature spatial anchor refinement with a mask region-based convolutional neural network (Mask RCNN). It uses the Facial Expression Recognition (FER) 2013 dataset for the evaluation process. Correspondingly, the efficacy of the projected model is calculated via various evaluation metrics, such as the recall, precision and mean average precision (mAP), to estimate the performance of the proposed DL method. It achieves 0.75298 for MAP@50, 0.70252 for precision and 0.66606 for recall. Furthermore, a comparison of existing models reveals the efficiency of the proposed DL method. The present research is intended to contribute to emerging object detection methods for enhancing real-time analysis of student emotions in various environments, such as classrooms and online education.

Variation of Production Line Energy Consumption: Stochastic Process Models for Single and Multiple Machine Systems

Po-Jung Lai — 2025-04-15

The energy consumption of production lines can vary across different machines due to diverse factors such as varying processing rates, machine conditions, product characteristics, and the compatibility between products and machines. Additionally, identical products on the same type of machine may experience temporal variations in energy consumption due to real-time equipment status and operational activities. Accurately understanding these influencing factors can aid in developing more efficient energy management strategies.

This study first analyzes the energy consumption of single machines processing single products, establishing a probabilistic model for the stochastic process of single-machine energy consumption over time. Mathematical tools, including convolution, are then employed to develop a stochastic process model for energy consumption across multiple machines and products within the entire plant. The overall stochastic process model is optimized, leading to the creation of a "Peak Energy Consumption Alert" and an "Energy Efficiency Alert" system to mitigate the risks of peak energy consumption and overall energy inefficiency.

By utilizing the proposed optimization model for machine component energy consumption and the production scheduling module, the likelihood of exceeding contracted power capacity is significantly reduced. This approach also enhances the energy efficiency of production scheduling, thereby reducing the total energy consumption of the entire plant. Furthermore, while meeting energy planning goals, the model considers non-energy-related production indicators (e.g., completion time, order delivery dates), ultimately improving the operational efficiency of production lines.

Synergistic Integration of Quantum and Classical Machine Learning Models for High-Fidelity Asteroid Hazard Detection

Kuldeep Vayadande — 2025-04-30

This research investigates the use of quantum machine learning (QML) to classify asteroids into non-hazardous and hazardous groups, which yields successful results in detecting the hazard. In addition to the complexity involved in analyzing orbits and physical objects, QML performs better than traditional machine learning in modeling the relationship between data. This method involves data-intensive preprocessing steps, such as feature selection by removing unnecessary features and correlation analysis to find predictors. Quantum circuits are used for specification and classification, and the standard evaluation is based on accuracy, recall, F1 score, and precision. A strong and weak method is provided by cross-validation and hyperparameter tuning. The best classical model here is the decision tree, which is a good model for high-resolution and low-budget social benefit with 0.883 accuracy, 0.955 recovery rate, 0.981 F1 score, and 0.883 sensitivity. However, the quantum model has made great leaps. AMSGRAD QCNN (Adaptive Moment Estimation with Gradient Thresholding Quantum Convolutional Neural Networks) achieves a non-uniform accuracy of 0.997, which is 13% higher than the decision tree with 0.984 accuracy, 0.955 recovery rate, and 0.981 F1 score. The accuracy of SPSA QCNN(Simultaneous Perturbation Stochastic Approximation) is 0.993, the recall is 0.974, and the F1 score is 0.977. This improvement shows the excellent ability of the quantum model to better resolve correlation data and, more importantly, to reduce false negatives in detecting stars. These results demonstrate the ability of quantum computing to analyze complex data and provide the best results. Our future work will focus on identifying and repairing quantum circuits, as well as exploring hybrid quantum classical models to improve model accuracy and interpretation. The findings open the door to new ways to predict the simplest solutions and the most powerful and accurate way to date of estimating the danger zone.

All-in-Net: Scorekeeping in Basketball Training using a Mobile Phone Camera and Image

Yonit Rusho — 2025-04-09

As basketball is one of the crowd's favorite invasion games, methods are being developed to enhance players' performance as technology improves. Recording large parts of training significantly reduces subjective training assessment compared to objective aspects. In this paper, basketball players and coaches use a mobile application based on image processing for achievements, measuring shooting the basket from different positions on the court and displaying feedback. The shooting technique and the percentage of success are measured using an algorithm that identifies the angle of the shot to the basket. The system monitors players' knowledge of results during training using one mobile phone camera. The paper describes the architecture and design of the mobile computing and application: Pre-training (define goals based on past performance and level of training difficulty), during training (data collection and compatible camera), and post-training (analysis and visualization of results). Finally, the paper discusses validation and implications.

Optimizing Healthcare in the Digital Era: Fusion of IoT with other Techniques

Rashmi Singh — 2025-01-10

The Internet of Things (IoT) has helped explore the healthcare industry. The present paper discusses the benefits and challenges associated with IoT in healthcare, highlights notable use cases, and presents the future prospects and considerations for successful implementation. Through a comprehensive examination of the topic, this paper aims to provide insights into the role of IoT in enhancing healthcare delivery, improving patient outcomes, and transforming the healthcare domain. A case study of brain tumor classification is investigated to explore IoT's applicability in healthcare. The VGG 16 model improved more consistently over the epoch, achieving higher validation accuracy than other models. In contrast, the discrepancies in validation accuracy and loss indicate the degree of variability of these models. The concept is augmented with fuzzy logic, nearness monitoring, and IoT in healthcare to understand future applicability, promising a better perspective on their transformational prowess.

Interlink Platform for School, Higher and Technical Education in India: Design Platform

Nitesh Ghodichor — 2025-04-13

This research paper aims to fill the knowledge gap in understanding the factors that contribute to academic performance and dropout rates in the Indian education system. The study proposes a “Interlinked platform for school, higher and technical education in India”, a unified digital space for students, educators, administrators and government agencies. The platform is designed to track students' academic performance across different educational levels and visualize dropout rates. The research uses a comprehensive methodology that integrates data analysis techniques and visualization frameworks and uses Python libraries such as NumPy, Pandas, Matplotlib and Scikit-Learn. Student academic outcomes are analyzed using linear regression and K-means clustering, and dropout rates are predicted using logistic regression. The aim of the research is to provide institutions with valuable insights to understand the factors that contribute to dropout rates and to develop targeted interventions to address potential triggers of dropout.

Brackish water parameters monitoring dashboard using Internet of things and industry 4.0

V. Sowmiya — 2024-11-12

INTRODUCTION: Brackish water aquaculture plays a crucial role in meeting the growing global demand for seafood. It offers an opportunity to diversify aquaculture production and reduce pressure on overexploited marine resources.

OBJECTIVES: By harnessing the unique properties of brackish ecosystems, this practice contributes to food security, economic growth, and sustainable resource management, while also promoting the conservation of valuable marine habitats. The development of a cutting-edge Indigenous Water Quality Monitoring Prototype named "Aqua BuoySis" for precision brackish water aquaculture utilizing machine intelligence.

METHODS: The prototype integrates sensors for Dissolved Oxygen (DO), pH, Temperature, Turbidity, and Total Dissolved Solids (TDS). These sensors are calibrated using a dynamic temperature-based machine-learning approach to ensure accuracy in real-time environments. Sensor calibration constants are uploaded to a server for comprehensive data calibration.

RESULTS: The system collects data at 20-second intervals, associating it with specific pond IDs. Data refinement is achieved through Long Short-Term Memory (LSTM) processing. An Android and Web application, available in native languages such as Tamil and Telugu, has been developed to provide live updates to aqua farmers, facilitating informed decision-making.

CONCLUSION: This technology represents a significant step towards enhancing precision in brackish water aquaculture through the fusion of machine intelligence and water quality management.

Integration of Blockchain and IoT for Enhanced Transparency in Diamond Supply Chain

Amit Sharma — 2025-01-02

PURPOSE: With a focus on enhancing transparency, lowering the risk of fraud, and ensuring ethical sourcing practices, this research aims to investigate how blockchain and IoT technologies can be incorporated into the diamond supply chain. This study addresses the complexities and challenges of implementing these technologies in an industry characterized by fragmented information sharing and centralized data storage.

DESIGN/METHODOLOGY/APPROACH: Using both qualitative and quantitative analysis, the research uses a mixed-methods approach. While secondary data was obtained from previously published works, industry reports, and case studies, primary data was gathered through semi-structured interviews with professionals in the field. The implementation of the prototype system was carried out in three phases: Define, Operate, and Test. Ethereum was chosen for its smart contract capabilities, and various IoT sensors were deployed to monitor environmental conditions and track the real-time location of diamonds.

FINDINGS: The integration of blockchain and IoT technologies significantly enhanced transparency within the diamond supply chain. The immutable nature of blockchain ensured tamper-proof records of transactions, while IoT sensors provided continuous real-time data, reinforcing transparency. The study observed a notable reduction in fraud due to the robust mechanisms of the system, which detected and prevented unauthorized alterations to the recorded data. Smart contracts automated compliance checks, ensuring adherence to ethical standards. Quantitative analysis revealed improvements in key metrics such as fraud reduction rates, transparency enhancements, and adherence to ethical sourcing standards.

ORIGINALITY/VALUE: This study bridges a notable gap in existing research by focusing on the diamond supply chain. It provides comprehensive, data-driven insights and practical recommendations for industry stakeholders and policymakers. The results highlight how combining blockchain and IoT technology can improve operational efficiency, transparency, and ethical practices in the diamond business. It is also feasible and scalable. The study's methods and findings add a great deal to the body of information already in existence and provide a framework for further investigation and application in related situations.

A Secure and Efficient Blockchain-Based Framework for Smart Cities Using Physics-Informed Neural Networks

Mohd. Asif Gandhi — 2025-04-14

The massive scale and extensive implementation of the Internet of Things (IoT) makes it difficult to provide secure and private communications over it. Privacy and decentralisation have been made easier using blockchain technology. Unfortunately, these solutions aren't practical for the majority of IoT uses because of how much time and computing power they require. Secure and private IoT that makes efficient use of available resources is proposed in this study. With the use of Physics Informed Neural Networks, the technique takes advantage of the computing power available in IoT settings like smart cities. This solution examines the reliability of the blockchain-based Smart Cities Architecture with respect to accessibility, privacy, and integrity. When weighed against the security and privacy advantages our system offers, our simulation findings reveal that the overheads (distribution, processing time, and energy usage) are negligible.