Indonesian Journal of Electrical Engineering and Informatics (IJEEI)
Not a member yet
776 research outputs found
Sort by
An Adjacent Gray Code Pairing Approach for Fault Identification in Reversible Circuits
Full text linkThe classical computing works on the concept of irreversibility increasing the heat dissipation in computing machinery. The remarkable ability of reversible computing is to reduce this dissipation of heat and to generate lossless information. The reversible circuit is the way to implement the reversible function. Therefore, the perfection of the functional behavior of the circuit plays a vital part within the realm of testing. Occurrence of faults in the reversible circuits creates a dysfunctional behavior in the circuit. Here in this study, a fault detection approach has been devised for reversible circuits that effectively identifies all categories of Missing Gate Faults (MGFs), such as single, multiple, partial, and repeated gate faults, through the use of the Adjacent Gray Code Pairing (AGCP) technique. The approach includes a process of conversion of binary to gray codes and pairing of non duplicative consecutive adjacent codes by which test vectors can be achieved for finding the respective faults with a 50\% reduction in test set size. An experimental study and comparative analysis with existing methods have been carried out using a variety of standard reversible circuits
Rasa-Powered Conversational AI Framework for Intelligent Electric Vehicle Trip Planning and Energy Management
Full text linkThe rapid growth of Electric Vehicles (EVs) calls for intelligent solutions to optimize grid stability and enhance user experience. This paper proposes a geo-optimized, user-centric EV management system integrating open-source geospatial tools with Rasa-based Natural Language Understanding (NLU). Through an interactive conversational interface, EV owners receive real-time trip planning recommendations based on parameters such as State of Charge (SoC), charging point availability, and route efficiency. The system utilizes APIs for real-time geolocation, charging station data, and Battery Management System (BMS) insights to determine optimal charging locations, durations, and trip costs. Extensive testing demonstrates improved energy management and route planning efficiency, highlighting the system’s potential for smart and sustainable EV infrastructure development
Enhancing Privacy in eGovernment: A Scoping Review of Data Minimization Techniques
Full text linkThe protection of personal data collected by e-government services plays an important role in balancing privacy and personalization establishing user trust, operational efficiency, and regulatory compliance. This scoping review investigates data minimization techniques used in personalized e-government services, identifying available techniques, and challenges. A key strategy for enhancing privacy involves limiting data collection and processing to what is only necessary for service delivery, particularly in e-government services. The scoping review, following the PRISMA ScR approach, addresses research questions on the current data minimization techniques in e-government services, their impact on personalization, challenges and barriers to implementation, and the perceived benefits from different stakeholders’ perspectives. From the formulated research questions covering the objectives of this scoping review it identified 2408 documents using relevant search query statements from available academic databases, after conducting screening and eligibility checks, only 20 documents are included in this review. From the documents, only proportional logic and game theory data minimization technique is used in e-governance systems. The impacts of data minimization techniques to personalization, the barriers and challenges in the implementation of data minimization, and the perceived benefits from the major stakeholders of the e-government systems were identified from the covered documents. This review has provided insights as to the extent of studies which include aspects of data minimization application in various egovernment systems. Findings provide direction to future research, policy formulation, and practice, emphasizing gaps and guiding future studies to a more comprehensive understanding of balancing privacy and personalization through data minimization in e-government services
Optimizing Data Survivability in Unattended Wireless Sensor Networks: A Machine Learning Approach to Cluster Head Selection and Hybrid Homomorphic Encryption
Full text linkThe research relies on machine learning-based Cluster Head (CH) selection and optimised Attribute-Based Encryption (ABE) with Homomorphic Encryption to improve data survivability in Unattended Wireless Sensor Networks (UWSNs). Integrating blockchain technology would enable tamper-proof data storage and provenance. The suggested method uses machine learning techniques like Deep Q-Networks (DQNs) or other models for intelligent and adaptive CH selection in UWSNs. Dynamically selecting CHs takes into account energy efficiency, network coverage, communication dependability, and node characteristics. The second part protects data using optimised Attribute-Based Encryption (ABE) and Homomorphic Encryption. ABE offers fine-grained attribute-based access control to restrict data access to authorised entities. Secure processing of encrypted data using homomorphic encryption protects privacy and integrity. These encryption algorithms are optimised to balance security and computational performance for efficient data processing and transmission while guaranteeing data privacy and integrity. Blockchain technology is suggested for tamper-proof data storage and provenance. To optimise the suggested solution's performance, the study uses the Seagull Optimisation Algorithm (SOA) and the Whale Optimisation Algorithm (WOA). These algorithms fine-tune system parameters, optimise CH selection, and boost UWSN performance. This holistic strategy uses machine learning-based CH selection, optimised ABE with Homomorphic Encryption, and blockchain technology for tamperproof data storage and provenance to improve UWSN data survival. Optimisation algorithms boost the solution's efficacy and efficiency, protecting UWSN data, latency, and energy usage
A Hybrid Deep Learning Framework for Accurate Polycystic Ovary Syndrome Detection Using Ultrasound Imaging
Full text linkPolycystic Ovarian Syndrome (PCOS) is a hormone-related health condition in women, commonly classified as an endocrine disorder. It is most prevalent during the childbearing years, typically between the ages of 15 and 44. PCOS leads to hormonal imbalances that cause irregular menstrual cycles, hair loss, and other symptoms, and it is associated with long-term health risks such as heart disease and diabetes. Recent advances in deep learning have shown promising results in accurately recognizing and differentiating ovarian cysts from other ovarian tumours. This study proposes a novel technique for PCOS symptom detection by analysing ovarian images through feature extraction, classification, and metaheuristic-based optimization. Ovarian images are first pre-processed for noise removal and smoothing, followed by feature extraction and classification using a Convolutional Wavelet Attention Neural Network with a Naïve Bayes Fuzzy Autoencoder (CWANN–NBFA). Optimization is then performed using the Metaheuristic Multilevel Hawks Algae Optimization (MMHAO) algorithm. Experimental evaluations were conducted on multiple ovarian image datasets. The proposed technique achieved an accuracy of over 98% across the PCOSUSG, KFHU, and MMOTU datasets, demonstrating its robustness and effectiveness in addressing the challenges of PCOS detection
Optimizing IoT Protocol Coexistence and Security using Software Defined Network and Intelligent Machine Learning Detection
Full text linkThe rapid growth of heterogeneous IoT environments has made seamless communication across protocols like MQTT and CoAP increasingly difficult, leading to interoperability gaps, latency issues, and security vulnerabilities. This paper proposes a Software-Defined Networking (SDN)-based architecture that integrates MQTT and CoAP through a bidirectional translation layer, while embedding machine learning (ML) intelligence for real-time flag status monitoring and Denial-of-Service (DoS) attack detection. The system leverages classifiers such as SVM, DT, NB, RF, and KNN within the SDN controller to dynamically predict operational states and mitigate malicious traffic. To evaluate performance, a Mininet-based IoT testbed with 50 heterogeneous nodes was deployed. Simulation results demonstrate that the proposed system achieves up to 95% message delivery success, reduces average latency by 18% compared to baseline translation methods, and saves 12–15% residual energy when using SVM-based classification. While the system improves interoperability and security, it also introduces computational overheads at the SDN controller due to ML inference, which may impact CPU and memory utilization in resourceconstrained environments. The proposed solution is highly relevant for smart city, industrial IoT, and healthcare applications, where interoperability and real-time resilience against attacks are critical. By unifying heterogeneous devices and enhancing security, this approach provides a scalable and practical pathway for next-generation IoT networks
BCDNN: Enhancing CNN Model for Automatic Detection of Breast Cancer Using Histopathology Images
Full text linkThe United Nations has identified health and well-being for all as one of its sustainable development goals. Research efforts in the healthcare domain worldwide are aligned with this goal. According to the World Health Organization (WHO), there has been an increasing incidence of breast cancer globally. The emergence of Artificial Intelligence (AI) has enabled learning-based approaches for diagnosing various ailments in the healthcare domain. Numerous efforts have been designed to efficiently diagnose breast cancer using deep learning algorithms, with the Convolutional Neural Network (CNN) being the widely used model due to its efficiency in processing medical images. However, CNN-based models may experience deteriorated performance without empirical studies to improve the underlying architecture. Motivated by this fact, our paper proposes a deep learning-based system for breast cancer diagnostic automation by enhancing a CNN model called the Breast Cancer Detection Neural Network (BCDNN). We also introduce an algorithm called Enhanced Deep Learning for Breast Cancer Detection (EDL-BCD), which leverages the enhanced deep learning model for better disease diagnosis performance. Our evaluation with a benchmark dataset comprising breast histopathology images shows that our suggested framework significantly outperforms state-of-the-art models, achieving an impressive accuracy of 97.99%. Therefore, the proposed system can be integrated with healthcare applications to assist in automatic screening by utilizing histopathology pictures to visualize breast cancer
Wn-Based Skin Cancer Lesion Segmentation of Melanoma Using Deep Learning Methods
Full text linkThe incidence rate of skin cancer, particularly malignant melanoma, has risen to high levels during the last decades. The biopsy method used for cancer treatment was found to be a painful and time-consuming one. Also, laboratory sampling of skin cancer leads to the spread of lesions to other body parts. Due to the different colours and shapes of the skin, segmentation and classification of melanoma are more challenging to analyze. An automatic method of dermoscopic skin lesion detection will be introduced. Recognizing the skin lesions at an early stage is essential for effective treatment. Proposed an efficient skin cancer image segmentation method using Fixed-Grid Wavelet Network (FGWN) and developed a novel classification method using deep learning techniques. FGWNs constitute R, G and B values of three inputs, a hidden layer and an output. Input skin cancer image is segmented, and the exact boundary is determined accordingly. The features of the segmented images were extracted using the Orthogonal Least Squares (OLS) algorithm. The AlexNet model was first used to classify pictures of melanoma cancer. Next, ResNet-50 and Ordinary Convolutional Neural Networks (CNN) was deployed. Wavelet Network (WN)-Based segmentation achieved an accuracy of 99.78% in detecting skin cancer lesion boundaries. Ordinary CNN shows an accuracy of 93.37% for 100 epochs. ResNet-50 models show 88.37% accuracy for melanoma classification. The number of training epochs and the volume of training data both impact accuracy. Deep learning algorithms can significantly improve categorization efficiency
Strengthening Cybersecurity: DDoS Attack Detection with Deep Learning and Innovative Hybrid Methods
Full text linkDistributed Denial-of-Service (DDoS) attacks continue to disrupt the availability of online services, motivating the development of robust and scalable detection mechanisms. This work proposes a hybrid CNN–LSTM detection framework evaluated in a controlled, sandboxed testbed for traffic generation and monitoring. The framework is implemented under a supervised learning setting and is positioned to incorporate semi-supervised and transfer learning strategies to address label scarcity and potential distribution shift in future extensions. Using a dataset of 6,000 labeled traffic logs and an 80/10/10 train/validation/test split, the proposed model achieves 98.67% accuracy, 98.01% precision, 96.73% recall, and 97.37% F1-score, outperforming Random Forest (96.42%) and a standalone LSTM (97.10%). Overall, the hybrid design supports improved detection robustness and can serve as a practical component within layered DDoS defense strategies (e.g., filtering and elastic scaling) in operational environments
Waste Classification Using NasNet-Mobile: A Multi-Stage Deep Learning Approach for Environmental Sustainability
Full text linkImproper waste management remains a significant global challenge, resulting in severe environmental and health impacts. Existing classification systems were designed and studied on large deep learning models, which are computationally expensive and not well-suited for embedded systems. To overcome this challenge, this study introduces a lightweight NasNet Mobile architecture that was trained using a three-stage learning framework. The framework employs transfer learning, fine-tuning, and hyperparameter optimisation to improve the model’s performance and generalisation capabilities progressively. To validate the proposed approach, experimental evaluations were conducted on TrashNet and Garbage Classification datasets. The model achieved an accuracy of 91.25% on the TrashNet dataset and 94.85% on the Garbage Classification dataset using the optimal hyperparameter set obtained through the random search technique. These results indicate that the proposed strategy effectively adapts to varying data distributions and outperforms popular Convolutional Neural Network (CNN) architectures, such as VGG-16, ResNet, AlexNet, etc. Therefore, the proposed model provides a reliable foundation for developing scalable and efficient waste classification systems for environmental applications. This study contributes to a practical deep learning approach that improves classification performance while maintaining low resource requirements for sustainable waste management