JAREE (Journal on Advanced Research in Electrical Engineering)

Classification of Diabetic Retinopathy Using ResNet50

La Ode Ansyarullah S. Sagala — 2025-08-01

Deep learning has been proposed as an automated solution for classifying the severity levels of Diabetic Retinopathy (DR). In this study, we utilized ResNet50 architecture to classify DR using the APTOS2019 dataset. As an initial step, we initialized the model with pre-trained weights from ResNet50 on ImageNet and implemented augmentation and resampling during training. We adopted an ensemble approach combined with classifiers such as SVM, Random Forest, and Logistic Regression, resulting in a ResNet50-Ensemble (SVM+RF+LR), with outputs obtained using a Soft Voting Classifier. The model achieved an accuracy of 85%, with a precision of 0.72, recall of 0.71, and F1-score of 0.71. The AUC values for the normal, mild, moderate, severe, and proliferative classes were 1.00, 0.96, 0.95, 0.95, and 0.91, respectively, with a Macro-average AUC of 0.96. These findings indicate that the appropriate use of ensemble methods can significantly enhance DR classification performance with suitable optimization strategies.

A Stress Level Monitoring System for Rescue Teams During Search and Rescue Operations Based on Electroencephalography

dedy hariyadi — 2025-08-01

Search and Rescue (SAR) officers work in high-risk conditions that require physical and mental resilience. Prolonged stress can affect the performance and success of SAR operations. This study evaluates the effectiveness of Electroencephalography (EEG) coherence analysis as a method for monitoring stress in SAR personnel. Using the OpenBCI EEG device and electrodes in the F3 and F4 areas, the brain activity of SAR personnel was recorded in two conditions, office activity (baseline) and rescue operations (SAR condition). The data collection for this research involved the same participants in both baseline and SAR operation conditions, resulting in 30 raw EEG data for further analysis. Data collection on operational conditions was carried out while the rescue officers conducted a search and rescue operation for a capsized boat in the Bengawan Solo River, Ngadirejo Village, Tuban Regency. Data analyzed based on coherence values obtained through the Power Spectral Density (PSD) features of alpha, beta, and gamma sub-band to detect changes related to stress levels. The results showed an increase in coherence in the alpha sub-band by 85.5%, beta sub-band by 92.9%, and gamma sub-band by up to 94.9% during moderate stress conditions, reflecting increased attention, alertness, and intensive information processing required in emergency situations. These findings indicate that EEG coherence analysis can be an effective tool for monitoring stress in SAR personnel in real-time.

The Selection of Energy Storage in the Southern Sulawesi Electricity System with AHP-TOPSIS Method

Peri Indriyanto — 2025-08-01

In the fourth quarter of 2023, a power deficit of up to 600 MW was observed in the Southern Sulawesi (Sulbagsel) electricity system due to the prolonged El Nino that affected hydropower plants. Based on the 2021-2030 RUPTL, the Sulbagsel system does not allow the addition of renewable energy (RES) plants, except with a battery firming scheme while the construction of fossil-based power plants is also limited by the government. One potential mitigation strategy for PLN (Perusahaan Listrik Negara) is to evaluate the potential application of Energy Storage Systems (ESS) in the Sulbagsel system. The purpose of this research is to evaluate the selection of ESS in the South Sulawesi electricity system through a combination of AHP (Analytical Hierarchy Process) and TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution). The AHP-TOPSIS method is an optimal choice for this evaluation due to its capacity to address complex evaluation scenarios by considering a multitude of comprehensive criteria and conducting weighting analyses to ascertain accuracy and identify optimal solutions. The evaluation criteria encompass technical, economic, environmental, and social considerations, with a total of 15 subcriteria. The alternatives to be evaluated are PHES (Pumped Hydro Energy Storage), BESS (Battery Energy Storage System), and HESS (Hydrogen Energy Storage System). The results of the evaluation demonstrate that PHES is the optimal alternative, with a preference value of 0.68. BESS is ranked second with a preference value of 0.33, while HESS is ranked third with a preference value of 0.20.

A Novel 350 MHz Capacitive Soil Moisture Sensor for Precision Agriculture

Pablo Petrashin — 2025-08-01

This paper presents a novel soil moisture sensor system based on a Colpitts oscillator operating at 350 MHz. The sensor utilizes the variation in capacitance of a sensing capacitor formed by two electrodes inserted into the soil. As soil moisture changes, the dielectric constant of the soil-water mixture also changes, directly affecting the capacitance and thus the oscillation frequency of the Colpitts circuit. This frequency range (150-500 MHz) was specifically chosen to minimize the influence of soil salinity on measurements, as supported by previous research.

The sensor design is simple, consisting of readily available and low-cost components such as capacitors, inductors, and only one RF transistor. This simplicity makes the sensor suitable for mass production using standard PCB fabrication techniques. Laboratory tests were conducted using a GW INSTEK GSP-827 spectrum analyzer and a Digital Electronics L/C Meter IIB to calibrate the sensor and validate its performance. The tests demonstrated a strong correlation between oscillation frequency, capacitance, and soil moisture, as evidenced by the data presented.

Key advantages of the system include its simplicity, low cost, low energy consumption, and robustness against soil salinity, surpassing the performance of traditional resistive sensors in conductive soils. The sensor offers potential applications in automated irrigation systems and precision agriculture, enabling optimized water usage and improved crop management. Future research directions include linearizing the sensor's response to enhance measurement accuracy, particularly in soils with high conductivity, and developing biodegradable electrodes using materials like beeswax and soy mixtures, balsa wood, or polylactic acid (PLA) to enhance the sensor's sustainability and minimize its environmental impact

Markerless Facial Reconstruction Motion Capture Using Triangulation Method

Muhammad Alwali — 2025-08-01

Motion capture is a popular research topic, with one of its main applications being human face reconstruction. The demand for converting 2D images into 3D reconstructions continues to increase, especially in facial reconstruction, where progress is made in improving the accuracy of facial position prediction. However, there is still a significant gap in developing facial reconstruction technologies that can consistently convert 2D to 3D data with high accuracy, especially in scenarios involving dynamic facial expressions, diverse facial angles, and complex environmental conditions. Therefore, an approach using the triangulation method for 3D face reconstruction in the real world was developed. In the experiments, two cameras were used to obtain two face landmark coordinates so that the triangulation method can be implemented for 3D face reconstruction. This research aims to develop a motion capture approach that is able to accurately and efficiently transform 2D data into 3D face models without the need for complex hardware. The main contribution of this research is the development of a machine learning-based markerless motion capture technique designed to improve the accuracy of face position prediction in 3D face reconstruction from 2D data in realistic environments. This method seeks to bridge the current technology gap by providing a more flexible and reliable solution, expanding the potential applications of motion capture in various fields without dependence on specialized hardware. The results of face reconstruction research using markerless motion capture and triangulation method show RMSE values of 3.560839 for eyes, 1.644749 for nose, and 4.054638 for lips.

Optimal Variable Speed Control of BLDC Diesel Generator to Enhance Fuel Efficiency

Bakhtiar Sudibyo — 2025-08-01

The growing adoption of renewable energy technologies still faces challenges such as instability, intermittent, and limited energy storage capacity. Diesel engine generators, known for their stability and reliability, remain essential as primary or backup power sources, especially in remote areas. However, conventional diesel generators operating at constant speed are inefficient in fuel consumption and produce high emissions. This study investigates the implementation of a variable-speed diesel generator system using a BLDC (Brushless Direct Current) generator controlled by a fuzzy logic-based controller (FLC). The proposed system adjusts engine speed and the duty cycle of the converter to optimize fuel efficiency while maintaining voltage and frequency stability. Simulation results demonstrate that the system reduces fuel consumption by up to 7.6% (0.86 liters/hour) for a 100 kW generator. Additionally, the FLC effectively stabilizes voltage and frequency during load changes and finally enhancing overall system performance.

Diagnostic of generator partial discharge and acoustic imaging scheme for monitoring winding health condition and repair

Riyo Purnomo — 2025-08-01

Partial discharge in a generator is a phenomenon of electrical discharge that occurs in the generator winding insulation. This phenomenon, if it continues, will cause the generator winding insulation failure. Therefore, partial discharge needs to be monitored, both on-line and off-line. Currently, partial discharge monitoring equipment has been installed at the Cirata Hydroelectric Power Plant (PLTA Cirata), and routine off-line partial discharge testing has been carried out. With this data, it can be used as a reference for the necessary generator maintenance. If the partial discharge trending value increases 2 times in a year, maintenance is required. During maintenance, acoustic imager testing is carried out to map the location of the partial discharge. By combining the three types of testing, a good generator monitoring system will be obtained, and the right type of maintenance will be determined according to the partial discharge mapping results. The analysis begins by conducting on-line and off-line partial discharge trending. If the trending has exceeded the limit, mapping is then carried out with an acoustic imager during maintenance, and repairs are made according to the location and type of damage. Analysis of maintenance results is carried out by comparing the magnitude of the partial discharge value before and after maintenance. The maintenance results based on the test results of the three methods proved effective, this was indicated by a decrease in the off-line PD value from 751.3 nC to 3.792 nC and on-line PD from 23.92 nC to 1.059 nC.

The Comparison of GAN and CNN Models in the Innovation of Coloring Madura and Bali Batik

Yohan Permana — 2025-08-01

This study aims to innovate automatic coloring of batik patterns using deep learning models. Specifically, it compares the performance of Generative Adversarial Network (GAN) with pretrained Caffe-based Convolutional Neural Networks (CNN) in coloring images of Madura and Bali batik. The dataset consists of 388 Madura batik images for training, 97 for validation, and 20 distinct images of both Bali and Madura batik for testing. This dataset was obtained through web scraping from batik posts on social media platforms like Instagram, Bing Image Search using specific keywords, and Kaggle, followed by a manual combination and cleaning process. The GAN model was trained with varying epochs (40, 80, 150), while the CNN utilized pretrained Caffe weights. Evaluation was conducted using Peak Signal-to-Noise Ratio (PSNR), Fréchet Inception Distance (FID), Mean Squared Error (MSE), and Structural Similarity Index (SSIM). The results indicate that the GAN model with 150 epochs outperformed the CNN, achieving a PSNR of 29.702, an FID of 84.016, an MSE of 511.8812, and an SSIM of 0.9925, demonstrating superior color creation and artistic detail in batik. Conversely, the CNN model exhibited lower performance, with a PSNR of 28.218, an FID of 200.271, and an SSIM of 0.7925, indicating its limitations in preserving the intricate patterns and colors of batik. This research demonstrates the applicability of GAN in automatic batik coloring, potentially providing innovative solutions for the batik industry while maintaining the cultural and artistic integrity of traditional designs.

Modified an Automatic-Cleaning Strainer for Seawater Cooling System in Combined Cycle Power Plant

Okwaldu Purba — 2025-08-01

The power plant industries on the coast use seawater as a cooling water system. Seawater conditions in densely populated areas, such as Jakarta, have high levels of pollutants which can affect cooling water systems. To filter seawater from pollutants, a strainer is used due to reliability must be maintained. This research is located at a power plant in North Jakarta. This research describes the latest proposed design of an automatic-cleaning strainer that poses less risk to the system and is faster in the work process. The results of this design were implemented at the generating company. This design can maintain strainer reliability with a faster strainer cleaning process, which reduces the TTR (Time to Repair) in the generating unit from 3 hours to 5 minutes, with no risk of damage to the system due to foreign material in the cooling water system. In addition, the results of this research help improve plant operation and maintenance performance, by increasing EAF 1.11%, increasing 1 level in Maintenance Mix Cost and Manhour from Level 4 to Level 5, increasing Reliability Management by 0.1 level, and also reducing EFOR by 0.98%. Work processes that are carried out automatically via smartphones also contribute to the digitization of the power generation industry to make it more effective and efficient.

Integration of Liquid Organic Fertilizer Fermentor with Automated Hydroponic Fertilization Based on IoT

Muhammad Fakhrudin Zukhri — 2025-08-01

As urbanization continues to accelerate, particularly in cities like Surabaya, the availability of agricultural land has been steadily decreasing, making food security a growing concern. In response to these challenges, urban farming, particularly through hydroponic systems, has emerged as a promising solution to ensure sustainable food production in limited spaces. However, issues such as the high cost and limited availability of high-quality fertilizers, as well as the difficulty in maintaining a consistent farming schedule, have posed significant barriers. This study aims to address these challenges by integrating IoT-based systems for temperature and pH monitoring, aiming to enhance farming efficiency. The validation results for both the DS18B20 Temperature Sensor and the pH Sensor 4502-C demonstrate their high accuracy and reliability for environmental monitoring. The DS18B20 sensor showed minimal error, with 0.89% for increasing temperatures and 1.34% for decreasing temperatures, achieving 99.11% and 98.66% accuracy, respectively. These results confirm the sensor’s effectiveness in real-time temperature control applications, such as those used in hydroponics and fermentation systems. Similarly, the pH Sensor 4502-C exhibited remarkable performance, with 99% accuracy in the acidic buffer, 98.99% in the neutral buffer, and 99% in the basic buffer. The error rates were extremely low, at 0.002% for acidic and basic buffers, and 0.01% for the neutral buffer, reinforcing the sensor’s reliability for pH monitoring in controlled environments.