Indonesian Journal of Electrical Engineering and Computer Science
Not a member yet
9109 research outputs found
Sort by
Detection and severity classification of ataxia using gait features and a hybrid model
Get PDFAtaxia, a neurological disorder characterized by impaired coordination and unsteady movements, presents significant challenges for accurate diagnosis and classification. traditional machine-learning (ML) and deep-learning (DL) models often struggle to achieve high accuracy in predicting and classifying this complex condition. This study addresses these limitations by introducing a novel hybrid model, XGBoost-multi-layer-perceptron (XGB-MLP), specifically designed to enhance the accuracy of ataxia prediction and classification. The objective of this research is to develop a more reliable and precise diagnostic tool that outperforms existing ML and DL approaches. The methodology involved integrating the strengths of XGBoost, known for its powerful gradient boosting, with the multi-layer perceptron (MLP) neural network, creating a robust hybrid model. The proposed XGB-MLP model was rigorously tested against conventional models like random forest (RF), logistic regression (LR), support vector machine (SVM), MLP, and standalone XGBoost. The findings reveal that the XGB-MLP model achieves outstanding accuracy rates of 98.91% for ataxia prediction and 97.91% for classification, significantly surpassing the performance of the traditional models
Field-level sugarcane yield estimation utilizing Sentinel-2 time-series and machine learning
Get PDFThis work focused on developing a methodology for using machine learning (ML) approaches to establish a pre-harvest yield prediction model for sugarcane at field level by integrating time-series remote sensing imagery data with ML techniques. Ground truth agro data and thirty-one spectral vegetation indices were extracted from Sentinel-2 imagery and were considered for yield modeling. A two-level feature selection technique was used to determine the most significant variables that best correlated with sugarcane yield to predict yield in advance. Seven ML algorithms, including those based on regularization, decision trees, and ensemble methods like boosting, were used to predict yield. The approach achieved the highest R2 score of 0.73 and the lowest root mean squared error (RMSE) of 13.45 t/ha with random forest (RF) among the seven ML models tested. Furthermore, all feature selection procedures identified normalized difference red edge (NDRE), red edge chlorophyll index (RECI), and ratio vegetation index (RVI) as major yield-driving variables. The experiments during feature selection demonstrated the potential of red edge spectral bands in development of a reliable sugarcane-yield prediction approach. The RF model obtained using the proposed methodology outperforms the two baseline models developed using NDVI and GNDVI indices, with an improved RMSE of 16-18%
Classification of weather conditions based on automatic weather station data using a multi-layer perceptron neural network
Get PDFWeather is one of the important elements that greatly determines human activities, especially those related to economic factors. Therefore, understanding weather conditions using weather parameters as a reference is important for human life, so a method is needed to classify weather according to its category so that the information produced can be used for various needs. Determining weather conditions in an area will not run well without a reliable method that can analyze existing weather parameters. Therefore, in this study, the weather condition classification process was carried out using the multilayer perceptron algorithm, a type of neural network (NN) algorithm. All data analyzed were weather parameter data collected by mini weather stations placed on land. The weather parameters used were temperature, humidity, air pressure, wind speed, dew point, wind chill, daily rainfall, solar radiation, and UV index. This study was conducted in Palu city, Central Sulawesi Province, Indonesia. The classification process carried out by the multilayer perceptron algorithm was carried out on the Altair AI Studio application and produced an accuracy value of 93.87%, recall of 92.33%, and precision of 91.29%
Building knowledge graph for relevant degree recommendations using semantic similarity search and named entity recognition
Get PDFCareer guidance is a critical and often daunting process, particularly during the transition from high school to higher education within the Moroccan education system. Faced with a vast array of university programs and career options, students frequently struggle to make informed decisions that align with their aspirations and skills. To address this challenge, our research introduces an innovative system that combines semantic similarity search with knowledge graph (KG) construction to enhance the precision and personalization of academic recommendations. By utilizing Sentence-BERT (SBERT) for semantic similarity, we generate embedding vectors that capture nuanced relationships between student profiles and degree descriptions. Subsequently, named entity recognition (NER) is applied to extract essential information such as skills, fields of study, and career opportunities from these profiles and descriptions. The extracted entities and their interrelationships are then structured into a coherent KG, stored in a Neo4j database, enabling efficient querying and visualization of complex data connections. This approach provides a transparent and explainable framework, ultimately delivering tailored advice that aligns with students’ individual needs and educational goals
LMD-based fault detection scheme for TCSC compensated wind integrated transmission lines
Get PDFIn this paper, a fast fault detection scheme is presented to detect the faults in thyristor-controlled series capacitor (TCSC) compensated transmission line connected with the large wind farms to export the electrical power to grid. The proposed logic utilizes the current information at the relay location and processes through the local mean decomposition technique to extract the magnitude features of the current. Cumulative sum of these features are computed for each phase currents to detect the faults in the transmission lines and further to classify the faulty phase in the system. The residual component of the current is used to detect the ground involvement in the faulty phase. The proposed method is tested during variety of faults by changing the nature of the fault using the fault parameters. Furthermore, the impact of the TCSC is also investigated along with the dynamic changes of the WF and their influence on the protection scheme. All the simulations are performed in MALTAB-Simulink software
IDCCD: evaluation of deep learning for early detection caries based on ICDAS
Get PDFDental caries is a common oral disease in children, influenced by environmental, psychological, behavioral, and biological factors. The American academy of pediatric dentistry recommends screening from the time the first tooth erupts or at one year of age to prevent caries, which mostly affects children from racial and ethnic minorities. In Indonesia, the 2023 health survey reported a caries prevalence of 84.8% in children aged 5-9 years. This research introduces early caries detection using three deep learning models: faster-RCNN, you only look once (YOLO) V8, and detection transformer (DETR), using Indonesian dental caries characteristic datasets (IDCCD) focused on Indonesian data with international caries detection and assessment system (ICDAS) classification D0 to D6. The results showed that YOLO V8-s and DETR gave good results, with mean average precision (mAP) of 41.8% and 41.3% for intersection over union (IoU) 50, and 24.3% and 26.2% for IoU 50:90. Precision-recall (PR) curves show that both models have high precision at low recall (0 to 0.2), but precision decreases sharply as recall increases. YOLO V8-s showed a slower and more regular decrease in precision, indicating a more stable performance compared to DETR
Four quadrant operation of bidirectional DC-DC converter for light electric vehicles
Get PDFThis paper discusses the closed-loop control of a bidirectional full bridge DC-DC converter which aids in the four-quadrant operation of an electric vehicle (EV). Several topologies of bidirectional converters have been recently investigated for optimizing vehicle performance. The bidirectional converters with buck and boost modes of operation aid the four-quadrant operation of drives. The proposed bidirectional converter aids buck and boost modes of operation in both forward and reverse directions of the drive. The buck/boost operation in the forward direction is suitable to operate the traction drive in motoring mode. Also, the buck/boost operation in the reverse direction aids the drive to operate in charging mode. The performance analysis of the bi-directional converter-fed EV drive is done using MATLAB/Simulink software. The different modes of operation of the converter which is utilized for the four-quadrant operation of the drive are validated using a 12-60V hardware prototype. DSP TMS2837D controller is used to control the bi-directional converter and the code generation for the controller is done in MATLAB-DSP integrated platform. The hardware results validate theoretical analysis and simulation studies
Flexible hybrid graphene-based NFC tag antenna for temperature monitoring application
Get PDFA hybrid graphene-based material, composed of reduced graphene oxide (rGO) and silver nanoparticle (AgNP), has been proposed for a near field communication (NFC) tag antenna with an integrated, flexible temperature monitoring circuit. The limited availability of high-conductivity graphene-based materials in the market has restricted the use of graphene in NFC tag applications. Therefore, this paper proposes a hybrid graphene-based composition featuring a high conductivity of 3.95×106 S/m. The feasibility of this material for NFC tags had not been validated previously, which is the main motivation for this research. The synthesis of the materials, along with the design, fabrication, and characterization of the NFC tag, is also presented. Results show that the inkjet-printed tag achieves a good reading range of up to 3 cm and demonstrates robustness against bending from 60⁰ to 190⁰, maintaining a maximum reading range of 1.3 cm. Performance on various materials, such as plastic, paper, and carton, also shows minimal impact on frequency shifting. Additionally, the graphene-based NFC tag integrates well with the temperature circuit, effectively monitoring temperatures in the 20-60 ⁰C range in real-time. This makes the developed tag suitable for applications such as food safety monitoring systems through NFC-integrated packaging
Analysis of cryptographic methods for ensuring security in the field of internet of things
Get PDFThe number of internet of things (IoT) devices continues to grow, and so do the associated concerns regarding their security and privacy. Evaluating the efficacy of cryptographic solutions within IoT systems emerges as a crucial endeavor to uphold the integrity and reliability of these systems. Amidst the rapid evolution of IoT technology, safeguarding the confidentiality, integrity, and availability of data emerges as a top priority. This article underscores the significance of deploying robust cryptographic algorithms to fortify IoT devices against a myriad of potential threats. Effective evaluation of cryptographic solutions within IoT systems entails a comprehensive analysis and comparison of diverse algorithms, coupled with an assessment of their performance, resilience against attacks, and resource utilization. Central to evaluating the effectiveness of cryptographic solutions within IoT systems is a consideration of various factors including computational complexity, power consumption of devices, ease of implementation, and compatibility with existing infrastructures. This article reviews a number of cryptographic solutions including Rivest–Shamir–Adleman (RSA), El-Gamal, Paillier. These algorithms are implemented on the ATmega2560 microcontroller, which allows for a comprehensive assessment of key parameters such as efficiency in terms of encryption and decryption time, power consumption, and memory usage of IoT devices
SRCNN-based image transmission for autonomous vehicles in limited network areas
Get PDFHigh-quality images are crucial for navigation, obstacle detection, and environmental understanding, but transmitting high-resolution images over constrained networks presents significant challenges. This study introduces an image transmission system using super-resolution convolutional neural networks (SRCNN) to enhance image quality without increasing bandwidth requirements by transmitting low-resolution images and upscaling them with SRCNN. The first phase of the research involved data collection, in which information was acquired directly from an appropriate locus to produce training, validation, and testing datasets. The second, three SRCNN models (915, 935, and 955) were trained using such a training dataset. The last was an evaluation, in which model 915 showed quick learning and stable performance with initial high loss, while model 935 had rapid convergence but potential overfitting. Model 955 achieved high initial performance. Three SRCNN model configurations were tailored to the specific needs of autonomous electric vehicles operating in limited areas, such as the locus. Input image resolution ranged from 128×128 pixels to 256×256 pixels, while output resolution varied from 256×256 pixels to 512×512 pixels. These resolutions can be acceptable for efficient image transmission over IEEE 802.11ac, but on the long range (LoRa) network, it still produces some delay